The fiscal magnitude of pension expenditures at the state level and the significance of pensions ... this respect, the focus of the study is on the following questions: ..... simply different ways of putting a claim on future economic production. ...... conditions for adoption of paradigmatic pension reform, conducted a survey of.
CONTENTS PUBLICATIONS RELATED TO THE THESIS .................................................................... 7 ACKNOWLEDGEMENTS ..................................................................................................... 9 LIST OF ABBREVIATIONS ................................................................................................ 10 INTRODUCTION.................................................................................................................. 11 PART I. CONCEPTUAL AND THEORETICAL APPROACHES...................................... 13 1. CONCEPTUAL AND THEORETICAL FRAMEWORKS ON PENSION SYSTEMS... 13 1.1. Old age as a social risk ................................................................................................ 13 1.2. Parameters of pension schemes and basic policy choices ........................................... 14 1.3. Vulnerability of pension schemes to various risks ...................................................... 18 1.4. Multi-pillar pension systems........................................................................................ 22 2. THEORETICAL EXPLANATIONS OF PENSION REFORM PROCESS ..................... 25 2.1. Contributions from welfare state studies ..................................................................... 25 2.2. Policy changes, pension reforms and transformation .................................................. 26 2.3. Theoretical approaches to explain pension reforms .................................................... 29 2.3.1. Historical institutionalism .................................................................................. 30 2.3.2. Actor-centered institutionalism.......................................................................... 31 2.3.3. Ideational approaches......................................................................................... 36 3. METHODOLOGY AND RESEARCH METHODS OF THE THESIS............................ 41 3.1. Research strategy, design, and methodological issues ................................................ 41 3.2. Data and methods ........................................................................................................ 43 PART II. THE STUDY OF THE ESTONIAN PENSION SYSTEM ................................... 46 4. PENSION REFORM PROCESS IN ESTONIA ................................................................ 46 4.1. Background factors..................................................................................................... 46 4.1.1. Demographic changes ...................................................................................... 46 4.1.2. Economic background...................................................................................... 50 4.1.3. Pension system up to 1990 ............................................................................... 54 4.2. The early reforms 1990–1993..................................................................................... 56 4.3. Pension policy over the transition period 1993–2000 ................................................ 60 4.4. Operation of the state pension system in 1990–2000 ................................................. 62 4.5. Reasons for multi-pillar reform and expected results................................................. 69 4.5.1. Pension reform debate in the mid 1990s .......................................................... 69 4.5.2. The 1997 reform blueprint ............................................................................... 70 4.5.3. Pension reform preparations in 1998–2002...................................................... 74 4.6. Main elements of the multi-pillar reform .................................................................... 78 4.6.1. Changes in the state pension system ................................................................ 78 4.6.2. Implementation of the second pillar................................................................. 87 4.7. International influences ............................................................................................... 99 4.8. Early post-multi-pillar reform experience................................................................. 104 4.8.1. Results of first pillar reforms.......................................................................... 104 4.8.2. Transition to the mixed system ...................................................................... 112 4.8.3. Transition costs .............................................................................................. 118 4.8.4. The third pillar................................................................................................ 120
5
5. CONCEPTUALIZATION AND THEORIZING OF TRANSFORMATION OF THE ESTONIAN PENSION SYSTEM...................................................................122 5.1. Overview of reform stages and policy outcomes ....................................................122 5.2. Analysis of the reform process in estonia against existing theories of pension reform......................................................................................................126 5.2.1. Re-experimentation and re-calibration in the Estonian case ..........................126 5.2.2. Explanations along historical institutionalism................................................127 5.2.3. Explanations along actor-centered institutionalism........................................129 5.2.4. Ideational explanations ...................................................................................132 5.3. CONCLUSIONS .....................................................................................................133 EESTI PENSIONISÜSTEEMI TRANSFORMATSIOON: POLIITIKA VALIKUD JA TULEMID. Kokkuvõte ...................................................................................................137 REFERENCES .....................................................................................................................140 ELULOOKIRJELDUS.........................................................................................................152 CURRICULUM VITAE.......................................................................................................153
6
LIST OF PUBLICATIONS I.
Lauri Leppik 2006. Coordination of Pensions in the European Union: the Case of Mandatory Defined-Contribution Schemes in the Central and Eastern European Countries. – European Journal of Social Security. Vol.8. No.1, 35–55.
II.
Lauri Leppik, Andres Võrk 2006. Pension reform in Estonia. – E. Fultz (Ed.) Pension Reform in the Baltic States: Estonia, Latvia, Lithuania. Budapest: ILO, 17–141.
III.
Lauri Leppik 2005. Impact of the EU social policy on a new Member State – reflections on the Estonian case. – E. Palola, A. Savio (Eds.) Refining the Social Dimension in an Enlarged EU. Helsinki: STAKES, Ministry of Social Affairs and Health, 103–108.
IV.
Lauri Leppik 2005. Pensionipoliitika. – R. Tammist, H. Rumm (koost.) Eesti poliitika eile, täna, homme. Reforme seitsmest valdkonnast. Tallinn: Johannes Mihkelsoni Keskus, 139–165.
V.
Lauri Leppik 2005. The route to Estonian pension reform. – Global Pensions. July 2005, 34.
VI.
Lauri Leppik, Ene-Margit Tiit, Andres Võrk 2004. Eesti pensionisüsteem Euroopa Liidu ühiste pensionieesmärkide valguses. – Riigikogu Toimetised No.9, 81–90.
VII.
Ene-Margit Tiit, Lauri Leppik, Andres Võrk, Reelika Leetmaa 2004. Euroopa Liidu ühiste pensionieesmärkide mõju Eesti pensionisüsteemile. – PRAXISe Toimetised 14/2004. http://www.praxis.ee/data/PRAXISe_Toimetised_Pension_2704048.pdf
VIII.
Lauri Leppik 2004. Pensionisüsteemi adekvaatsuse ja jätkusuutlikkuse tasakaalu poole. – Sotsiaaltöö, No.5, 38–39.
IX.
Lauri Leppik, Ruta Kruuda 2003. Study on the Social Protection Systems in the 13 Applicant Countries: Estonia Country Study. – Gesellshaft für Versicherungsvissenshaft und –gestaltung e.V. Social Protection in the Candidate Countries: Country Studies Estonia, Latvia, Lithuania. Berlin: AKA, 3–140.
X.
Lauri Leppik 2003. Social protection and EU enlargement: the case of Estonia. – V. Pettai, J. Zielonka (Eds.) The road to the European Union, volume 2: Estonia, Latvia and Lithuania. Manchester and New York: Manchester University Press, 140– 162.
XI.
Lauri Leppik 2003. Pensionisüsteemi ja tööturu vastasmõjudest: Poliitikauuringute Keskuse PRAXIS pensioniuurimuse vahetulemustest. – Sotsiaaltöö, No.6, 5–7.
XII.
Lauri Leppik, Georg Männik 2002. Transformation of old-age security in Estonia. – W. Schmähl, S. Horstmann (Eds.) Transformation of pension systems in Central and Eastern Europe. Cheltenham – Northampton: Edward Elgar Publishing, 89–124.
XIII.
Lauri Leppik 2002. Disability protection in Estonia. – E. Fultz, M. Ruck (Eds.) Reforming Worker Protections: Disability Pensions in Transformation. Budapest: ILO-CEET, 93–145.
XIV. Lauri Leppik 2002. La protection contre l'invalidité en Estonie. – E. Fultz, M. Ruck (Eds.) Réformer la protection des travailleurs: les pensions d’invalidité en transformation. Budapest: BIT-EECO, 103–159. XV.
Lauri Leppik 2000. Läti käivitab kogumispensioni. – Eesti Majanduse Teataja 4(107), 19–20.
7
XVI. Lauri Leppik 1999. Social protection system of Estonia facing the EU enlargement. – T. Tulva (Ed.) Estonian Social Welfare at the Threshold of the Millennium. Tallinn: Tallinn University of Educational Sciences, 74–89. XVII. Lauri Leppik 1999. Euroopa Liidu mõju Eesti sotsiaalpoliitikale. – A. Oja, A. Raukas (Eds.) Eesti 21. sajandil: arengustrateegiad, visioonid, valikud. Tallinn: Teaduste Akadeemia Kirjastus, 22–28. XVIII. Lauri Leppik 1998. Eesti pensionireform: kolm sammast. Tallinn: Jaan Tõnissoni Instituut. XIX. Lauri Leppik, Hilja Kõgel 1998. Social protection. – Estonian Statistical Office. Social trends. Tallinn, 61–68. XX.
Frans Pennings (Ed.), Lauri Leppik, Jaak Vitsur, Natalia Karotam, Steven Vansteenkiste, Jan Nelissen 1998. The consistency of Estonian social security law with EU legislation on equal treatment of men and women. Report of the EC Phare Consensus Programme. Tilburg.
XXI. Lauri Leppik 1998. Pensionireform. – M. Malvet, M. Mikkola. Sotsiaalhoolekanne. Helsinki: Karelactio, 135–142.
8
ACKNOWLEDGEMENTS Professional development is a long process influenced by many factors and actors. The thesis is one of the milestones in this process and a good occasion to thank all those persons who have been influential for my professional development. It is through exchange of ideas, information and experience with other people that new knowledge and ideas – including those presented in the current thesis – emerge, and for that I owe many thanks to my colleagues at the Tallinn University, Tartu University, Ministry of Social Affairs, Social Insurance Board and the PRAXIS Center for Policy Studies, as well as to my collaboration partners in several other institutions, including the European Commission, the Council of Europe, the World Bank, the Ministry of Finance and elsewhere. In academic terms, I am particularly grateful to two influential persons: Professor Winfried Schmähl from the Center of Social Policy of the Bremen University and Elaine Fultz from the ILO. Some parts of this dissertation have been published as a chapter “Pension Reform in Estonia” in the book Pension Reform in the Baltic States: Estonia, Latvia, Lithuania edited by Elaine Fultz. These parts are reprinted with a kind permission from the copyright holder, ILO.
9
LIST OF ABBREVIATIONS CEECs CPI DB DC ECDS EEA EEK ENSIB ESO EU EUR FCR FRR FRP FSA GDP ILO IMF MDC NAV NDC OECD PAYG SSRC WB
10
Central and Eastern European Countries Consumer Price Index Defined-benefit Defined-contribution Estonian Central Depository for Securities European Economic Area Estonian Kroon (currency) Estonian National Social Insurance Board Estonian Statistical Office European Union Euro (currency) Fixed contribution rate Fixed replacement rate Fixed relative position Financial Supervision Authority Gross Domestic Product International Labour Organisation International Monetary Fund Macro-level defined-contribution Net Asset Value Notional defined-contribution Organisation of Economic Co-operation and Development Pay-as-you-go Social Security Reform Commission World Bank
INTRODUCTION The current dissertation focuses on developments of the Estonian pension system over the period of 1990–2005. Pension systems in modern societies are the largest mechanisms of income transfers. Pensions are a significant source of income for a large share of population (incl. elderly and disabled persons), while the economically active population is participating in the financing of pension schemes by paying taxes and/or contributions. The fiscal magnitude of pension expenditures at the state level and the significance of pensions and contributions on household revenues and expenses make the pension system a major economic vehicle. Although the economic aspect is important, it is only one of several aspects of pension systems. What makes pension systems one of the most complex social policy mechanisms is the sophisticated interplay of different social, demographic, economic, legal, and political issues. Estonia inherited its social systems, including the pension arrangement from the Soviet Union. Regaining of independence and restoration of statehood were accompanied with deep changes in various aspects of society, including the economy, social, political and legal order. Transition from the paternalistic socialist state with centrally-planned command economy to a democratic society based on free markets, witnessed by Estonia along with other Central and Eastern European countries, was unparalleled. At the outset, there were no theories to guide this transition or predict its outcomes. As noted by Wagener (1999), there was no generally accepted theory of transformation from socialism to capitalism. Against this background, it is particularly interesting to analyze what happened to the largest system of income transfers – the pension system – over the period of societal change. From one side, the context of societal transition opened a ‘policy window’ – an opportunity for alternative solutions (see Kingdon 1984). However, at the same time, implications of societal transition on the economy, demographic and social structures imposed constraints on policy alternatives. A detailed study of evolution of the Estonian pension system helps to shed light into the process of formation of postsocialist social policy in the context of societal transition. In 1990s and early 2000s pension reforms were undertaken in a number of Central and Eastern European countries (including all three Baltic states) that were going through a transition from socialism to capitalism. In spite of broad similarities both in respect of the legacy and in terms of opportunities and constraints created by the transition, the pension reform process, reform design and outcomes have varied across countries in the region. This raised a theoretical quest for factors that are crucial in triggering pension reforms and explanation of diversity and common patterns of reform processes and reform designs in Central and Eastern Europe. First attempts to offer theoretical explanations of pension reforms in Central and Eastern Europe emerged relatively lately, in late 1990s (Müller et al 1999; Müller 1999). These attempts of theory-building suffered from a ‘small-N problem’, simply because in most of the Central and Eastern European countries pension reform has been only a second-phase transformation phenomenon, undertaken several years after major political and economic reforms (Wagener 1999). In this context, transformation of Estonian pension system may be regarded as another example in the ‘family’ of
11
pension reforms in Central and Eastern Europe. A case study of Estonian pension reform may serve as an additional test of validity for the existing theoretical frameworks. Against this background, the current thesis serves two major aims. The primary aim is to provide an analytical description of pension policy changes in Estonia over the period of 1990–2005. This includes identification of causal factors and actors that have impinged policy change, and conceptualization of the transformation process. In this respect, the focus of the study is on the following questions: - how the design of pension policy in Estonia has changed? - which factors and actors triggered reforms? - why reforms were undertaken? - what are the main outcomes of reforms? The secondary aim is to contribute to the broadening of theoretical frameworks and understandings of pension systems and pension reforms. Obviously, a single case study is not sufficient for theory formation and building of a new comprehensive theory of transformation of pension systems goes beyond the scope of the current thesis. Instead, drawing on recent theorizing about pension reform process in Central and Eastern European countries, I will test some of the existing theoretical explanations of pension reforms on the Estonian case. As will be indicated, the Estonian case of pension reform challenges some of the earlier theoretical explanations on transformation of pension systems in Central and Eastern Europe. The study takes a social policy perspective, looking into the pension policy reform process and policy outcomes. The first chapter depicts conceptual and theoretical frameworks for describing and analyzing pension systems, main pension policy choices, vulnerabilities of pension systems to various external risks and the main strategies to cope with these risks. The second chapter presents the key concepts to describe policy changes, in particular as these have been applied to pension policy, and reviews three theoretical frameworks – historical institutionalism, actor-centered institutionalism, and ideational approach – which have been used to explain pension reform processes. The conceptual and theoretical frameworks outlined here are revisited in the final concluding paragraph in the light of findings of the study. In the third chapter the methodological approach and the research strategy and design of the study is explained and main methods and source data are described. The fourth chapter looks into details of pension policy changes over the period 19902005. Initially, the main trends of context indicators on demographic and economic background are presented. This is followed by descriptive analysis of pension system developments and analysis of the factors that have accounted for policy changes. In the final part of this chapter the early post-reform experience is analyzed, looking at reform outcomes both on macro and (to a lesser extent) micro level. In the fifth and final chapter, policy developments and main outcomes are summarized followed by an analysis of the pension reform process in the light of the theoretical approaches that are presented in the second chapter. The thesis integrates a number of earlier works of the author (Leppik 1998, 2002, 2003; Leppik and Männik 2002; Leppik and Kruuda 2003; Leppik et al 2004; Leppik and Võrk 2006). However, it also contains a number of previously unpublished components, which make it more than just a synthesis of earlier publications.
12
PART I. CONCEPTUAL AND THEORETICAL APPROACHES 1. CONCEPTUAL AND THEORETICAL FRAMEWORKS ON PENSION SYSTEMS 1.1. OLD AGE AS A SOCIAL RISK Old age is one of the core social risks addressed by many social security systems. However, it is not self-evident as to what the nature of ‘old age’ contingency is, i.e. what is precisely the risk against which the social security system shall offer protection. Typically, social risks are associated with suspension or reduction of earnings or increased expenses (e.g. Pieters 1993). However, in practice, with many old age pension systems the link to suspension of earnings is often only indirect. Pension systems often assume that persons attaining a certain age (pension age) have reduced options to earn income in the labor market. Whether this assumption actually holds, is a different matter. The 1952 ILO convention No.102 on minimum standards in social security and the 1964 European Code of Social Security stipulate that in case of old age benefits, the contingency covered shall be survival beyond a prescribed age, which, as a rule, shall not be more than 65 years. However, the ILO convention No.102 and the Code further stipulate that the age limit may be extended beyond 65, provided that the number of residents having attained such higher age is not less than 10 percent of the number of residents under that age but over 15 years of age. The conventional approach thus regards the ‘risk’ of old age as survival beyond a certain age. As Ferrera et al (2000:90) note, traditional definition of the ‘old age’ risk has been “life beyond 60 or 65”. The age limit is a matter of social contract, subject to some guidelines stemming from international treaties. It may be observed that the possible extension of the age limit beyond 65 permitted by the ILO convention No.102 and the European Code of Social Security with reference to demographic factors is based on an implicit assumption of acceptable burden for younger generations1. The legal phrase ‘survival beyond prescribed age’ refers to yet another aspect of the old age ‘risk’, namely the uncertainty of remaining lifetime. Old age pension schemes pay benefits until the rest of the beneficiaries’ lives, thus effectively pooling the life expectancy risk of scheme participants. In any society, there is considerable heterogeneity in life expectancy. However, for a scheme as a whole, 1
The 10% benchmark of the ILO convention No.102 and the Code is actually very low compared to the effective age dependency ratios (the ratio of persons in pension age to persons in working age) in European countries. According to this criterion, the legal pension age in Estonia in 2005 could be as high as 72 years.
13
the average life expectancy at pension age determines the average duration of pension payments. As Myles (2002) notes, in many affluent democracies retirement is increasingly regarded as an extended period of labor force withdrawal. In Western countries accumulation of sufficient retirement ‘wealth’ – in most cases, these are pension rights acquired from the state rather than accumulated real assets – has made work unnecessary for many persons. The substance of the ‘old age’ contingency has thus transformed in the context of social and economic developments. Raising life expectancy and raising economies have de-commodified old-age pension systems (see Esping-Andersen 1990). However, as shown by Esping-Andersen, the actual degree of decommodification of old-age pension schemes varies greatly across countries. 1.2. PARAMETERS OF PENSION SCHEMES AND BASIC POLICY CHOICES Pension systems are comprised of a number of characteristic elements – also referred to as parameters. The constituent elements of pension systems relate to: 1) the personal scope of the pension scheme: coverage rules determining the group of insured persons; 2) eligibility conditions: criteria determining entitlement to pensions (e.g. pension age, qualifying period, residence status etc.); 3) the material scope of the pension scheme: types of benefits, benefit calculation rules, benefit adjustment mechanism etc.; 4) the financing rules: e.g. financing instruments and financing methods; 5) administration: arrangements relating to granting and delivery of benefits. Each of these parameters may take different values. It is the totality of parameters, which defines a pension scheme. The choices of scheme parameters are the basic policy decisions in designing and reforming pension systems (Table 1). Parameters of pension schemes are not mere technical aspects, but elements, which influence distributional effects of the pension system. Therefore, the study of changes of these elements reflects important aspects of social policy developments. Table 1. Basic policy choices in old-age pension schemes Parameters Coverage Type of benefit Level of benefit Type of scheme Financing instrument Financing principle Administration
Basic alternatives Compulsory or voluntary Flat rate or earnings-related Replacement rate (from 0% to over 100%) Defined-contribution or defined-benefit Contributions, ear-marked tax or general taxes Pay-as-you-go or pre-funding State, autonomous public-legal body or private
Sources: Müller (1999); Schmähl (2002); own compilation
14
Most commonly, the reasons that have been pointed out for endorsing compulsory participation in pension schemes relate to individual myopia, unpredictability of the future and inability to save (e.g. Barr 1998; Müller 1999). In the absence of a compulsory pension system, some individuals would make no provision for their retirement due to either shortsightedness or deliberate free-riding. However, even without myopia there are individuals who are unable to save sufficiently for their own pension due to persistently low income during working years. Furthermore, unpredictability of the future – e.g. future economic developments or individual life expectancy – makes conscious planning of retirement a difficult task even for the most rational person. These factors give a rationale for pooling of individual retirement risks in the format of compulsory old age pension schemes. In practical terms, the question is therefore not whether there is a compulsory pension system, but rather what the level of protection is for which participation is compulsory. The level of compulsory coverage determines inter alia the remaining scope for voluntary schemes (Müller 1999). The policy choice concerning the type of benefit – flat-rate or earnings-related pension – refers to an underlying social-political distribution principle. Egalitarian principles behind flat-rate benefits imply extensive vertical redistribution from high-income to low-income earners. In a pension scheme with earnings-related benefits, pensions primarily carry the function of deferred earnings and maintain the basic structure of income distribution among working-age population2. Based on the basic type of benefit formula combined with the financing principles, pension schemes have often been categorized as Bismarckian or Beveridgian – the former referring to earnings-related pension schemes financed from earningsrelated contributions, the latter relating to flat-rate pension schemes financed from general taxation (e.g. Bonoli 1997; Müller 1999; Cremer and Pestieau 2003; Villa 2004). However, as many pension systems, at least in Europe, feature a mix of flatrate (e.g. minimum pension guarantees, like national pension) and earnings-related elements, this broad categorization based on the names of ‘founding fathers’ of modern social security is becoming increasingly outdated. The main question is no longer what category the pension system belongs to, but rather the extent of flat-rate elements versus earnings-related elements and the role of contributions (or earmarked taxes) versus general taxation in financing. Nevertheless, there is a qualitative difference between contributions and taxes as financing instruments for pension schemes. Contributions are an instrument for acquiring pension claims, while paying taxes does not create specific rights (Schmähl 2002). Ear-marked tax as a financing instrument is the middle ground between the two. Replacement rate – defined here as the ratio of average pension to average earnings from work – is a key social policy indicator in terms of generosity of pension 2
However, depending on benefits calculation rules the earnings-related pension scheme may also have some degree of vertical redistribution.
15
schemes3. The average replacement rate also indicates to what extent pension replaces former work income for a person with an average length career and average wage. Under pay-as-you-go (PAYG) financing, pensions are paid from current contribution revenues (and possibly, transfers from general tax revenues). Pay-as-you-go financing implies intergenerational redistribution. At the same time, intergenerational solidarity is a necessary precondition for functioning of such financing mechanism (Müller 1999). Under pre-funding, contributions are invested to finance benefits in the future. While pre-funding could be both collective and individual, the latter mode has been preferred in pension reforms of Central and Eastern Europe countries. Respective pension schemes are often called individually fully-funded schemes. In such schemes, individually recorded contributions are accumulated over the work career to finance pension payments for the same individual upon retirement. Pension schemes are often categorized into defined-benefit (DB) or definedcontribution (DC) type (e.g. World Bank 1994). In defined-benefit schemes, the benefit calculation rules are pre-determined, whereas the contribution rate may vary from year to year to collect necessary revenues to finance the benefits. In definedcontribution schemes, the contribution rate is fixed, while the amount of benefit depends on contributions made into the scheme, plus (or minus) any interest earned on contributions (see e.g. Iyer 1999). In other words, in defined-benefits schemes contributions are a dependent variable, whereas in defined-contribution schemes benefits are a dependent variable (Myles 2002). Defined-benefit and defined-contribution principles are essentially alternative ways to achieve financial equilibrium of the pension scheme. In the former case, the financing side is modified, while in the latter case the benefit side is modified to achieve financial balance. However, these categories refer to ideal types. In practice, often over time both sides of the pension scheme – benefit rules as well as contribution rates – have been modified. Defined-benefit and defined-contribution schemes may operate both under pay-asyou-go and fully funded financing principles (Table 2). Table 2. Defined-benefit versus defined-contribution pension schemes Definedbenefit Definedcontribution
Pay-as-you-go Benefits are pre-determined, financing is to be secured. Current revenues finance current benefits. Contributions are pre-determined, benefits depend on (either individual or total) contributions made into the scheme. Current revenues finance current benefits.
Funded Benefits are pre-determined, contribution rate may fluctuate. Collected revenues are invested to finance future benefits. Contributions are pre-determined, benefits depend on contributions, plus (or minus) any interest. Collected revenues are invested to finance future benefits.
Source: own compilation 3
16
Later we use also different replacement rates, e.g. the individual replacement rate, which reflects the ratio of individual pension to former individual earnings of the beneficiary.
Again, in real life, demarcation lines between types of schemes are not always clear. Pension schemes may be partially funded, i.e. combine pay-as-you-go and pre-funding, or combine defined-benefit and defined-contribution principle. Defined-contribution schemes are often privately managed and fully funded. In individually fully-funded defined-contribution schemes there are individual accounts to record acquired pension rights in the form of accumulated contributions. However, pay-as-you-go schemes may also be built on a defined-contribution principle. One example of a pay-as-you-go defined-contribution scheme is so-called notional defined-contribution (NDC) scheme where information about individual contributions is recorded on ‘notional’ accounts and the balance of these accounts serves as the basis for calculating the benefit at the time of retirement (see e.g. Palmer 2000; Holzmann and Palacios 2001; Börsch-Supan 2003). Such schemes have been introduced in Latvia, Poland, Sweden, and Italy. NDC schemes mimic fully-funded DC schemes as individual pension depends exclusively on past contributions (Müller 1999). However, in reality, in the NDC scheme the ‘accounts’ are only a book-keeping mechanism in a scheme that operates on pay-as-you-go principles. Another type of pay-as-you-go DC scheme is labeled here as macro definedcontribution (MDC) scheme. In such a scheme, the real value of pensions is determined by the total revenues available from contributions made by the current working population. As illustrated later in this volume, the Estonian state pension system operated on those principles during the period 1994–2000. The strengths and weaknesses of two alternative financing methods – pay-as-yougo versus funded – has received paramount attention in the literature (e.g. World Bank 1994; Thompson 1998; Holzmann 1999; Ribhegge 1999; Barr 2000; Orszag and Stiglitz 2001). It goes beyond the ambit of the current volume to enter into details of this extensive academic debate, which is often heavily ideological and normatively loaded. I shall agree with Barr, who notes (2000:34) that “the debate over PAYG and funding concentrates on a very narrow part of the pension's picture.” As Barr outlines (2000:5), in economic terms there are only two principal options for seeking security in old age: 1) to store some current economic production or 2) to claim share of future production. The use of the first option for individuals is limited, basically, to securing housing and some consumer durables before retirement. All pension systems are based on the second option. Different financing methods – pay-as-you-go and funding – are simply different ways of putting a claim on future economic production. In the case of pay-as-you-go financing, the claim, mediated by the government, is for a share from future tax/contribution revenues. In case of funded schemes, the claim on future production is in the format of accumulated assets, which are to be exchanged for cash at the time of retirement (see also Holzmann and Palacios 2001). 17
However, the issue at stake clearly is not only about financing, but the choice of financing method also implies important distributional aspects. While pay-as-yougo schemes imply intergenerational redistribution, the dominant redistribution mechanism in funded schemes is intertemporal – over the life-cycle of individuals. In many defined-benefit schemes, there are inherent intragenerational redistributions. However, even funded defined-contribution schemes may imply some intragenerational redistribution, although often to a substantially lesser degree than in definedbenefit schemes4.
1.3. VULNERABILITY OF PENSION SCHEMES TO VARIOUS RISKS Types of pension schemes have relevance also in the context of vulnerability to different external risks. Different types of pension schemes are vulnerable to different risks at varying degrees. As summarized by Gillion et al (2000), the major risks for pension systems are: • demographic risks (e.g. changes in birth rates or mortality rates); • economic risks (e.g. decreasing employment, low wage growth, high inflation, low investment rates of return); • political risks (e.g. sudden changes in scheme parameters); • institutional risks (e.g. failures of public or private institutions managing pension schemes). In particular, the first two categories of risks have received considerable academic consideration. For the time there appears to be a rather broad consensus, that demographic and economic developments principally affect all types of pension schemes – pay-as-you-go and funded, defined-benefit and defined-contribution (see e.g. Holzmann 1999; Müller 1999; Barr 2000; Holzmann and Palacios 2001). However, these factors influence pension schemes in different ways, depending on the policy design of the scheme (Table 3).
4
18
Many DC schemes prescribe the use of unisex life tables to calculate annuities. In a situation of gender differences in life expectancy, this leads to redistribution from men to women.
Table 3. Vulnerability of pension schemes to main risks Demographic risks decline of birth rates
Pay-as-you-go DB
Fully funded DC
negative impact on future revenues, but effects are gradual
no direct effects on financing of the scheme, but indirect negative effects on future benefits are possible no direct effects on financing of the scheme, but lower future benefits no direct effects, but indirect negative effects are possible
increase of life expectancy
increase of expenditures
emigration of economically active persons
decrease of revenues
Economic risks lower employment
low wage growth high inflation low rate of return
Political risks change of key parameters fiscal pressures
decrease of revenues
lower revenues higher revenues, but real value of benefits may decline minor effects5
relatively easy high vulnerability
no direct effects on financing of the scheme, but concerned individuals receive lower future benefits no direct effects real value of accumulated assets and pensions in payment declines no effects on financing of the scheme, but lower future benefits more difficult no direct effect
Sources: Holzmann (1999); Müller (1999); Gillion et al (2000); own modifications
For pay-as-you-go schemes, the total effect of different economic and demographic influencing factors combine in an indicator of system dependency ratio, which is the ratio of pensioners to employed persons contributing to the system. Under increasing system dependency ratio, the principal policy options are limited. However, economic and demographic factors do not pre-determine a particular policy choice; there are always alternatives. The choice between alternatives has multiple dimensions: e.g., normative, distributional, and financial. One of the alternatives is shifting from defined-benefit to defined-contribution scheme. A policy shift from DB to DC schemes considerably modifies the pattern of risk-sharing. Some risks (e.g. increasing life expectancy or low wage growth), which are borne collectively in DB schemes, are transferred to individuals in DC schemes. At the same time, DC schemes provide individuals with greater flexibility, e.g. between time of retirement and the amount of pension. 5
There may be some effects to the extent the scheme holds some current reserves.
19
The strategies available to mitigate such vulnerabilities differ according to the type of pension scheme. In a situation where pension systems are confronted with demographic and economic challenges, a key theoretical quest of recent years has been to find sustainable designs for pension systems – i.e. designs, where parameters of the pension scheme take into account changing demographic and economic circumstances, so that such changes would not require ad hoc political adjustments. NDC schemes have been ascribed such properties (see e.g. Palmer 2000). Indeed, the automatic stabilizers written into the NDC design – adjustments with changes in life expectancy, demographic buffer fund, and indexing of total pension volume with total payroll – address several vulnerabilities of other pay-as-you-go schemes. However, qualitatively, these risks are mitigated in a similar way to funded DC schemes. In addition, as elaborated later in this volume, the difference between NDC and some earnings-related DB schemes is often ‘optical’, rather than substantive (see also Börsch-Supan 2003). According to Barr (1998), there are only three basic alternatives for any pension scheme to cope with demographic ageing – to increase contributions, to reduce benefits or to increase pension age6. It is a matter of a different dimension – that of policy process – whether under some system design the necessary adjustments are more palatable than under a different design. According to Myles (2002), the main policy challenge posed by population ageing is neither demographic nor economic but distributional – if retirement costs rise, how are they to be allocated within and between generations? While Myles (2002) agrees with Barr (1998) in that additional costs can be partly offset by working longer and postponing retirement, he points out that there are also distributional effects to this strategy. It is likely that the welfare losses (reductions in leisure time, i.e. shortened period of retirement) resulting from increased pension age affect disproportionately the least advantaged, those with shorter life expectancies and with low income (who are often the same people). Myles (2002) illustrates the underlying intergenerational dilemma of pay-as-you-go pension schemes by contrasting two ideal types of policy designs. In a definedbenefit model with a fixed replacement rate (FRR), pensioners are entitled to a certain fraction of their former earnings, adjusted regularly with some index7. Under such a design, an increase of the system dependency ratio transfers into higher contribution rates. As benefits drive contributions, all costs associated with demographic ageing fall on contributors and their dependants. An alternative is a 6
Reduction of benefits does not necessarily mean declining nominal value of benefits, but mostly refers to declining real value of benefits or declining replacement rate, i.e. relative value of benefits. Concerning increase of pension age it shall be noted that increasing the legal pension age would ease the burden of ageing only if it is accompanied by increase in the effective retirement age.
7
However, in practice there are also defined-benefit schemes, where the replacement rate is not fixed.
20
design, based on a fixed contribution rate (FCR)8. Here contributions drive benefits and if system dependency ratio increases, benefits decline9. As a result, all costs associated with demographic ageing fall on pensioners. To solve this policy dilemma, Myles (2002), supported by Esping-Andersen (2002) and Hemerijk (2002), endorses a fixed relative position (FRP) model, which was earlier advocated by Musgrave (1986). In the FRP model, contributions and benefits are fixed in order to maintain a constant ratio of per capita net earnings of those in the working population to the per capita net benefits of pensioners. As they advocate, this method allows for proportional sharing of the risk of population ageing between generations10. Under FRP design, essentially both benefits and contributions are ‘indexed’ to population ageing. Myles (2002) also points to the original rationale of Musgrave (1986), that in practical terms, neither FRR nor FCR designs are politically sustainable under conditions of population ageing since both designs presume an intergenerational contract, which probably cannot be kept or at least, would generate uncertainties about its future. However, the FRP proposal has also received criticism. Vandenbroucke (2002) indicates that focusing only on the relative income position of pensioners, potential expenditure differences between the working age population and pensioners may be neglected, such as expenditure differences related to health care or social care. Secondly, as he points out, pensioners and working age population have different effective consumer price indices. As a result, it may be that unchanged relative incomes yield a fall in relative well-being for the elderly. Therefore, according to Vandenbroucke, a ‘fixed relative position’ between the generations should rather be addressed more in terms of well-being than in terms of income. When analyzing the options for different types of pension schemes to mitigate risks, many scholars (including Barr 2000; Myles 2002 etc.) limit their analysis to the case of closed systems. Indeed, their conclusions on available strategies hold under the assumption of closed systems; however, in an increasingly globalizing world, this assumption cannot be maintained. Under open systems, the range of alternatives is wider. In case of small countries – like the Baltic states – this extension becomes even more relevant. Under the assumption of open systems, for pay-as-you-go systems one of the alternatives is to import labor to increase the number of taxpayers11. For funded schemes, the parallel alternative is to export capital during the accumulation period and/or sell the assets of pension fund abroad during the payout period (see also Ribhegge 1999).
8 9 10
11
Notably, Myles discusses fixed contribution design in pay-as-you-go scheme, not in a funded scheme. This is not necessarily nominal decrease, but rather relative decrease compared to wages. Myles (2002:142) points out that while advocating for the FRP method, they do not take a normative position on the issue of what the relative income position of pensioners to working population should be, but only suggest a method for allocating the additional costs of demographic change between generations once an acceptable ratio is established.
However, ageing of migrants may simply postpone problems. 21
In any case, the set of alternatives is more complex than originally suggested by Barr (1998). Clearly, the issue of sustainability can not be downsized to the choice of financing mode or DB versus DC design. Both pay-as-you-go and funded schemes may be either sustainable or unsustainable. Scheme parameters and context factors are important in this respect. Inter alia, benefit adequacy is a vital element of system sustainability. If the results produced by policy change are not acceptable to the population, the reform will not be politically sustainable. This in turn may have consequences for financial sustainability. Thus, there is no ultimate sustainability. As Gillion et al (2000:302) write, “No pension scheme in an unpredictable world can completely succeed in providing a predictable source of retirement income.” Nevertheless, this does not imply that increasing the sustainability of pension systems has little relevance. Clearly, some system designs are more sustainable compared to others – both in terms of financial as well as social and political sustainability. Furthermore, Esping-Andersen (2002:23) points out that failure to reform pension systems may produce negative secondary consequences e.g. jeopardize job growth. Reforms may also be necessary to address ambivalences inherent in a pension scheme. 1.4. MULTI-PILLAR PENSION SYSTEMS A particular strategy to cope with uncertainties is a deliberate ‘mixing’ of different principles and design elements. Such strategies refer to multi-pillar pension systems, comprising of several pension schemes (or sub-schemes), based on different principles. The metaphor of pillars is widely used in social and political sciences, but has different connotations in different fields. In studies of welfare states, the notion of pillars has also several meanings. For example, Esping-Andersen (2002:11) refers to three welfare pillars – market, families, and state. Much of the welfare is purchased in the market. The reciprocity of kinship is another important source of welfare and security, e.g. in terms of income pooling. The traditional role of governments in welfare production is based on a redistributive ‘social contract’ that reflects some form of collective solidarity, i.e. is based neither on purchase nor on reciprocity. As Esping-Andersen points out, the three welfare pillars are mutually interdependent. Families and government may absorb market failures. The market or government may compensate for family failure. In pension literature, the notion of a pillar commonly refers simply to a subsystem of the total pension system, i.e. a pension scheme with a particular parameters and policy design. Some authors (e.g. Barr 2000; Schmähl 2002) prefer to use of the term of ‘tier’ rather than ‘pillar’, on the grounds that the notion of ‘tiers’ more adequately illustrates that different schemes build upon each other rather than stand apart, or pointing out that intuitively the pillars need to be of a broadly similar size (which is rarely the case) to provide stability. 22
However, in this volume I stick to the notion of pillars, mainly for the convenience of compatibility with the prevailing literature. Furthermore, the metaphor of pillars merely refers to different sources of welfare at old age. In this respect, there is a compelling parallel between welfare pillars as described by Esping-Andersen (2002) and pension pillars. In practice, at least in more advanced countries pension systems are rarely purely single-pillar systems. Rather, what characterizes the so-called single-pillar systems is the prevailing dominance of one scheme, which is most commonly a pay-as-yougo state pension scheme, leaving a relatively modest scope for other pension arrangements. Therefore, the demarcation line between ‘single-pillar’ and ‘multipillar’ systems is often seen in the absence or existence of any other compulsory pension schemes besides the state-managed pay-as-you-go scheme. Table 4. Multi-pillar framework I pillar mandatory public pay-as-you-go redistribution and deferred earnings
II pillar mandatory private funded individual savings or occupational pensions
III pillar voluntary private funded individual savings or occupational pensions
Source: World Bank (1994); own modifications
The multi-pillar framework clearly has acquired some ideological load, in particular due to the World Bank (1994) famous textbook. However, there is no normative standpoint associated with the multi-pillar framework per se. What has been controversial and contestable is the particular format of the second pillar – mandatory defined-contribution scheme – introduced by the World Bank typology. Furthermore, the relative size of different pillars is contestable. There are many scholars of welfare state, who recognize the advantages of multipillar arrangements over single-pillar systems. For example, Ferrera et al (2000:59) point out advantages of separating different distributional functions: “pension systems which display an institutional separation between a pay-as-you-go financed basic pension and a fully funded private mandatory insurance, also allow for a more targeted assignment of the various redistributive and insurance functions of the welfare state and are thus less likely to generate distributive conflicts than is the case for pension systems which combine these functions within one tier.” Ferrera et al (2000:59) also outline a cognitive dimension, indicating, “contributions for private pensions are perceived as part of private consumption rather than as part of the tax wedge and thus are likely to generate fewer work disincentives.” The importance of cognitive factors is recognized also by Barr (2000:25) who notes, “Because of citizens’ perceptions, funded schemes in some countries at some times in their history might have greater legitimacy.”
23
From the perspective of sustainability, the heavily debated question has been, whether the multi-pillar design has advantages in coping with various risks (as referred to in Table 3). Many authors suggest that a multi-pillar system provides better security through higher risk diversification (Chłon et al 1999; Holzmann 1999; Ferrera et al 2000; Fox and Palmer 2001). At the same time, Barr (2000:24) notes, “the proposition that multi-pillar pension systems diversify risks holds only if those risks are negatively correlated or are at least orthogonal to each other.” It is also notable that there is a large group of scholars who advocate for the prevailing predominance of the pay-as-you-go paradigm inter alia on normative and distributional grounds. The pay-as-you-go paradigm does not preclude the existence of private supplementary pension schemes, but according to this approach, these ought to be voluntary. However, in practical terms the issue of compulsory or voluntary participation in the ‘second pillar’ may sometimes be of secondary importance. More relevant is the actual weight, and at the cognitive level, the perceived role of the second pillar in the overall pension system. In this respect, the actual coverage and the (future) average replacement rate provided by the second pillar are important. Voluntary funded schemes with broad actual coverage may have more significant practical implications than relatively small compulsory funded schemes. Broadly speaking, the two paradigms – single-pillar social insurance and multipillar – favor different public-private mixes in pension systems and represent different normative views on the role of collective and individual responsibility in pension systems (see also Bönker 2004). In none of the actual pension reforms in Central and Eastern Europe or elsewhere has there been a complete shift from collective responsibility to an individual responsibility. Albeit, there have been partial shifts to the extent that the new second pillar is ‘carved out’ from the previous state system when transferring to a mixed system. However, even a partial shift is perceived as a threat to solidarity features of the welfare state by some commentators (e.g. Casey 2004). Nevertheless, it shall be pointed out that when public trust in the existing pay-asyou-go pension pillar is low, the funded pillar may help to restore trust in the overall pension system, which inter alia is important for contribution compliance. Pay-as-you-go and funded pillars shall therefore not necessarily be seen in terms of trade-offs. The interplay between different pillars may be also mutually reinforcing.
24
2. THEORETICAL EXPLANATIONS OF PENSION REFORM PROCESS 2.1. CONTRIBUTIONS FROM WELFARE STATE STUDIES The analysis of pension policy can benefit from a wealth of literature in different fields, in particular political science, economics, and sociology. From the perspective of social policy, theories of welfare state development provide useful frameworks, since the pension system is a core part of the welfare state. The notion of welfare state is used here in its conventional meaning, denoting programs of statutory benefits and services (see e.g. Barr 1998; Wilensky 2002; Clasen and Siegel 2005). This non-normative definition is chosen here deliberately. Several authors offer normative definitions of welfare state, e.g. Esping-Andersen (1990) associates a welfare state with a certain degree of decommodification, allowing persons to maintain a livelihood without reliance on the market. However, the aim here is not to engage in discussions as to what extent benefits and services in Estonia and other Central and Eastern European countries achieve the aim of decommodification or other normative tasks. The focus of this study is on policy changes and for that purpose, the narrow definition of the welfare state is sufficient. By this, I assume that the basic mechanisms and processes of policy changes are not qualitatively different depending on the normative tasks associated with welfare state benefits. Even if not explicitly recognized, this assumption is de facto shared by a vast array of researchers in the field of comparative social policy studies. It could also be pointed out that many authors have categorized Central and Eastern European states as welfare states (e.g. Orenstein and Haas 2002; Cerami 2005), albeit sometimes as emerging welfare states (e.g. Aidukaite 2004). Furthermore, several authors have regarded the socialist welfare state of the former Soviet Union and its Eastern European allies as a variant of the welfare state in general (Wilensky 1975; Wagener 1999; Guardiancich 2004). From this perspective, developments in Central and Eastern Europe may be regarded as a transformation from one type of welfare state to another. In studies of welfare states, there have been considerable efforts to understand the process of policy changes: which factors trigger changes, what is the role of different policy and political actors, which conditions enable and shape the policy processes and outcomes, etc. Besides studies which have focused on welfare states in more affluent Western European countries, there have been several studies (e.g. Müller 1999; Grimmeisen 2004; Schubert 2005), which in particular have addressed pension systems in CEECs, trying to explain why some CEECs took a more radical reform path than others, despite the fact that all CEECs were characterized by rather similar pension systems under the communist legacy.
25
2.2. POLICY CHANGES, PENSION REFORMS AND TRANSFORMATION Hall (1993:278) has suggested to “think of policymaking as a process that usually involves three central variables: the overarching goals that guide policy in a particular field, the techniques or policy instruments used to attain those goals, and the precise settings of these instruments …”. Based on these three central variables, Hall (1993) categorized policy change according to its magnitude into three orders: 1) change of the level or settings of policy instruments, while the overall goals and instruments of policy remain the same; 2) change of policy instruments and their settings, while the overall goals remain the same; 3) change of policy instruments and their settings, accompanied with a shift of policy goals (i.e. all three policy components are changed). Although Hall elaborated this framework for analyzing macroeconomic policy changes, it has become widely used also in the field of social policy (e.g. Palier 2000; Stone 2001; Natali 2001; Pfau-Effinger 2004; Béland 2005). Pfau-Effinger (2004) suggests complementing Hall’s categorization with the fourth order of changes – a shift in the basic values of the welfare state. However, since the current thesis is limited to the analysis of a single sub-system of the welfare state (i.e. pension system), I have chosen to stay with Hall’s original framework of three orders of change. At the same time, I assume that a shift in the policy goals reflects a shift in the underlying values. Transferring Hall’s categorization of policy changes to the area of pension policy (see also Palier 2000; Natali 2001), the first order of changes – change in the settings or level of policy instruments – could imply, e.g. adjustments (increase or decrease) of the level of pensions or contribution rates. Such adjustments, as a rule, do not alter the general principles and logic of a pension system12. The second order of changes involves the change of pension policy instruments, e.g. a change of pension calculation rules (pension formula) or eligibility rules (pension age, qualifying period). The third order of change involves a simultaneous change in both the policy instruments and policy goals. Examples of such changes are a shift from a defined-benefit to a defined-contribution pension scheme or a shift from a pay-as-you-go financing to pre-funding. In studies of pension reforms, a distinction is commonly made between parametric and paradigmatic reform (e.g. Holzmann et al 2003; Hinrichs 2003). Parametric reform is a significant change in the elements (parameters) of the pension system, while the basic structure of the system remains untouched. Such a reform may seek to make alterations in eligibility rules, benefit rates, or financing side to the pay-asyou-go pension system. Common elements of parametric reform include increasing the pension age, modifying pension indexation, curtailing privileges of special groups, shifting financing from contributions to general taxation, etc. Parametric 12
26
However, it may be claimed that a very radical increase or retrenchment of the benefit level essentially changes the underlying goals of the pension system, e.g. whether the system aims to secure only basic minimum protection or more significant income replacement.
reform may also include expanding voluntary private pension provisions, e.g. through tax advantages, tripartite agreements, or other means. In contrast, a paradigmatic reform is “a deep change in the fundamentals of pension provision” (Holzmann et al 2003:8). As the name suggests, this type of reform relates to the change of a paradigm in which a pension system operates. Typically, paradigmatic pension reform entails moving away from the monopoly of a pay-asyou-go pillar to a multi-pillar pension system – introduction of a mandatory funded pension pillar, along with a reformed PAYG pillar and the expansion of voluntary pension schemes. Noticeably, the categorization of pension reforms into parametric and paradigmatic is well compatible with Hall’s categorization of policy changes, corresponding respectively to the second and third order of policy change. The first order of pension policy changes could be labeled as incremental. The second order – parametric pension reforms – constitute adjustments to pension policy without challenging the overall terms of a given policy paradigm (see also Palier 2000). The third order entails a paradigm shift. Such reforms seek to alter the underlying policy goals, but it is important to note that such policy changes are often influenced by preceding changes in value systems. According to Hall (1992:91) the policy paradigm is an “overarching set of ideas that specify how the problems /…/ are to be perceived, which goals might be attained through policy and what sorts of techniques can be used to reach those goals.” Müller (1999, 2003) indicates that the paradigmatic pension reform may entail an important shift from an intergenerational contract to individual provision for old age, as well as from the state to the market as the supplier of retirement pensions. The paradigm change “amounts to a substantial rewriting of the underlying social contract, which usually does not occur in the case of a mere change of entitlement conditions.” (Müller 1999:38; 2003:48). Further, the notions of reform and transformation need some clarification. Wagener (1999) notes that as a rule, pension systems are reformed and not transformed. According to Wagener (1999:35), a reform “is based on pragmatic constructivist intentions”, but lacks the holistic systemic aspect, and is concerned only with partial system elements. Transformation concerns the overall paradigm and thus relates in particular to the macro level. However, paradigm change obviously influences the meso and micro levels as well. For Wagener (1999), the notion of reform relates to changes within a paradigm, while transformation entails a change of the paradigm. Earlier, Pierson (1996:173) had regarded transformation as a “fundamental shift”. In a similar vein, Caldwell (2003:2) defines transformation as a change that is “significant, systematic, and sustained”. Following these interpretations, the notion of transformation is used here as an overarching concept, which encompasses the process of changes in socio-economic context, institutions, structure, normative legitimation and values (see also Bönker et al 2002). Transformation is commonly regarded a process, not a one-time event, since for fundamental changes to take effect and be sustained, these need to extend over a longer period. 27
As Wagener (1999) indicated, there is no generally accepted theory of transformation from a socialist to a capitalist pension system. This is still the case in 2006. Furthermore, as Wagener pointed out, it is not entirely clear whether such a transformation (i.e. a paradigmatic change) is needed or if simply a (parametric) reform of the socialist pension system will do. At least in some Central and Eastern European states – the Czech Republic and Slovenia – pension systems have survived the context of broader socio-economic transformation with a series of parametric reforms (see e.g. Fultz 2002b; Müller 2002). A further significant contribution to conceptualization of reforms, in particular the substantive side of them, has been given by Ferrera et al (2000). To capture the basic logic of welfare reforms, Ferrera et al introduce two metaphors: reexperimentation and re-calibration. The former refers to the reform process and the latter to the substantive direction of reform. Ferrera et al (2000) suggest re-experimentation as a label for system-wide search for novel, economically viable, socially acceptable and politically feasible policy solutions. The adoption of an ‘experimental’ approach – policy innovation through new institutional settings – is often prompted by practical policy puzzles, sensitive to empirical evidence and based on policy learning and lesson drawing (id. p.88). To conceptualize substantive direction of reforms, Ferrera et al (2000:89) introduce the notion of re-calibration, which is meant to suggest an act of institutional reconfiguration and re-balancing. This is characterized by: 1) the presence of a set of constraints that condition developments, stemming from the interaction of new external pressures and domestic challenges; 2) the interdependence between additions (or upgrading) and subtractions in the policy menu under review, as a consequence of such constraints; and 3) a deliberate shift of weight and emphasis among the various instruments and objectives of social policy. According to Ferrera et al (id. p.89), re-calibration is a balancing act, and as such, it “is a delicate operation, which raises difficult dilemmas, involves tough political choices, and is inherently exposed to twin risks of either excessive or insufficient transformation.” Re-calibration is considered to have functional, distributive, normative, and politico-institutional dimensions. Functional re-calibration addresses the “goodness of fit” between the welfare state and a socio-economic reality. Ferrera et al (2000) indicate that this ‘fit’ is never constant, but is always evolving. Social and economic transitions gradually erode this congruence, generating a mismatch between the supply and demand for institutionalized welfare. Distributive re-calibration has to do with restructuring benefits schemes in terms of how these address the needs of particular social groups. Normative re-calibration addresses the “goodness of fit” between broad value premises in society and the policies in practice. Politico-institutional recalibration has to do with actors and institutions involved with the governance of the welfare state. 28
In summary, the conceptualization elaborated by Ferrera et al (2000), allows distinguishing between various dimensions of policy changes. However, it is not merely a descriptive tool, but also indicates the possible causal factors leading to reforms – a need for reform may arise from the erosion of the ‘fit’ between the welfare state and an evolving socio-economic reality. As they formulate (2000:88), welfare reform is necessitated by “a complex set of exogenous and endogenous challenges that are throwing the inherited design of the welfare state into question.” In the following, I shall look into details of some explanatory models for welfare state developments, which have a potential to offer explanations on pension reform in Estonia. 2.3. THEORETICAL APPROACHES TO EXPLAIN PENSION REFORMS Three medium-range theoretical frameworks on policy change in welfare states are reviewed here: historical institutionalism, actor-centered institutionalism, and ideational approach. These approaches shall not be perceived as mutually exclusive. However, in practice, the focus of each of these frameworks is different, as will be explained in more detail below. When reviewing these theoretical frameworks, the focus is on the following questions: what triggers a policy reform and what is conditioning the choice of a particular reform design? The three theoretical frameworks reviewed here are clearly not the only available explanatory models.13 Over the 1970s and 80s, different theoretical approaches were developed in attempts to explain the emergence and expansion of welfare states, and in the area of comparative studies, the diversity in welfare state models (e.g. Titmuss 1974; Esping-Andersen 1990). However, this analytical work was predominantly based on studies of welfare states in Western Europe and North America. I have chosen to focus on the theoretical frameworks of historical institutionalism, actor-centered institutionalism, and ideational approach. Historical institutionalism has become one of the most influential theoretical perspectives in policy studies, in particular in the area of social policy research since the beginning of 1980s, and gained a new momentum in 1990s in studies of welfare cutbacks. From the second half of 1990s, the approach of actor-centered institutionalism moved to the forefront and was subsequently used in several studies to explain pension reforms in Central and Eastern Europe. Ideational approach has been (re)discovered quite recently in studies of welfare state developments. Reviewing the Estonian case of pension reform against the background of these frameworks allows comparing the findings with some of the earlier results from studies of other Central and Eastern European countries. However, theorizing on pension reforms of Central and Eastern European countries is only gaining 13
More extensive overviews of theoretical approaches to explain welfare state developments in the field of pension can be found, e.g. in Müller (1999) and Schubert (2005).
29
momentum for the obvious reason that many of the reforms have been implemented only very recently. In this context, the findings of this study also serve as an additional test of validity of the existing theoretical frameworks and allow indication of some possible improvements to the existing explanatory models. Briefly, institutionalism claims that institutional settings influence the choice of the reform path. However, there are many different facets of institutionalism. The theoretical approach of institutionalism was reinforced in the 1990s, in particular by Pierson (1996), who shifted the analytical focus from welfare state expansion to retrenchment. 2.3.1. Historical institutionalism The key concept in historical institutionalism is path dependency. While there are somewhat different interpretations as to what precisely constitutes a ‘path’, here reference is made to the definition offered by Deeg (2001:36): a path is defined by the logic – predictable strategies and routines – generated through the operation of a given institution or institutional system. Pierson (1996) describes how large social programs enacted during the postwar era have created vested interests and ‘armies of beneficiaries’ that generally prevent widespread explicit attacks against the welfare state. The central idea is that decisions made earlier in history have a strong impact on the options and incentives available to contemporary decision-makers. Once a particular path is chosen, it becomes increasingly difficult to change it to another (Pierson 2000). Historical institutionalism is based on the assumption that a historically constructed set of institutional constraints and policy feedbacks structure the behavior of ‘actors’ and interest groups during the policy-making process (Immergut 1998). Most explanations stress the influence of the legacy of past decisions. As summarized by Deeg (2001), path dependence involves three phases. The first is the ‘critical juncture’ in which events trigger a move toward a particular path out of several (at least two) alternatives. The second is the period of ‘path reproduction’ – the period in which positive feedback mechanisms reinforce the movement along one path. Finally, the path ends when new events dislodge the long-lasting equilibrium. Thus, every path begins and ends with a ‘critical juncture’, marked by specific triggering events (Pierson 2000). As can be observed from this life-cycle approach to the concept of path, path dependence entails resistance to change, but it does not preclude change. Nevertheless, the theoretical approach claims that the pattern of those changes is dependent upon the previously chosen path. Path dependence clearly has significant relevance in respect to pension systems – probably more than in respect to other types of welfare state programs. Myles and Pierson (2001) suggest that pension policy is a prime example of path-dependent social processes. The relevance of path dependence is evident, inter alia since 30
pension rights are accumulated over a relatively long period of 30–40 years. The necessity to recognize earlier acquired rights often prescribes very gradual reform scenarios14. Palier (2001), analyzing pension policy developments in European countries with Bismarckian (pay-as-you-go earnings-related defined-benefit) pension systems (e.g. Germany, France), noted that for a number of years governments responded to increasing demographic and economic pressures by increasing contribution rates. This strategy was consistent with the underlying logic of defined-benefit schemes, but was also politically more feasible than cutting back pension entitlements. The benefit side of the system was touched only once the increase of contribution rates reached a ‘critical’ limit. Palier (2000) links the concept of path dependency to Hall’s (1993) distinction between different orders of policy change and suggests to reserve the label of ‘path dependency’ for changes at the second order (i.e. changes of instruments), which nevertheless may at times be quite far-reaching (Andersen 2001). While some authors have employed path dependency to account for the relative inertia of welfare states, others focus on the ‘critical juncture’ points and analyze what brings about path departure. In simple terms, the question is how far does history shape human choice? (Pfau-Effinger 2004). Andersen (2001) suggests using the concept of path dynamics to describe developments that are path dependent but at the same time turn out to be path breaking. In addition, Pfau-Effinger (2004) points out that processes of path deviation and path dependence can coincide. From the perspective of historic institutionalism, one could have expected the Central and Eastern European countries staying with the Bismarckian pension arrangement, because of the role of path-dependence: socialist and pre-war systems were employment-based, while high fixed costs, determined by large inherited payas-you-go defined-benefit schemes arguably prevent radical policy shifts (see e.g., Myles and Pierson 2000). However, as shown by a number of authors, this proposition did not hold true in several Central and Eastern Europe countries (see e.g. Müller 1999; Grimmeisen 2004; Cerami 2005 etc.). 2.3.2. Actor-centered institutionalism The theoretical framework of actor-centered institutionalism emerged in the 1990s because of an integration of the earlier structuralist-institutionalist and actorcentered approaches, which were perceived as an analytical dichotomy (Scharpf 1997). Müller (1999) was first to apply this framework to analyze pension reform dynamics. To explain policy choices and processes, this approach first identifies “relevant actors, whose choices will ultimately lead to the policy outcome under consideration” (Müller 1999:50). These actors may be individual, corporate, or collective and are characterized by specific action resources, perceptions, and preferences. 14
However, there is some political discretion on the real value of previously acquired rights.
31
The approach assumes that the action resources, which may include personal properties or access to information, enable actors to influence a policy outcome and are largely determined by the prevailing institutional rules. The actors’ perceptions and preferences influence their specific orientation, their evaluation of the situation and desirability of certain courses of action. However, as Müller further points out, individual, corporate or collective actors do not proceed in isolation, but interact with other relevant actors in a dynamic way. Therefore, “policy outcomes are shaped by actor constellations, involving all actors participating in the respective political process, as well as their strategy options and strategic preferences” (Müller 1999:51). Actor constellations may entail a variety of modes of interaction. These could be co-operative, non-co-operative, or hierarchical. The actual policy choices are determined by both the specific actor constellation and the respective mode of interaction. Beyond the focus on actors and their dynamic interaction, the analytical framework of actor-centered institutionalism also pays attention to the structural-institutional context in which the policy process takes place (Müller 1999). However, it is assumed that neither institutional setting nor contextual factors actually determine a particular policy choice or outcome. Rather – as the argument of Müller goes – the contextual setting stimulates, enables, or restricts the strategy options available to the involved actors. Furthermore, contextual setting may also shape the actors’ perception of the situation and the action that needs to be taken. At the same time, actors may also shape institutional rules, and their policy decisions have an impact on the structural environment. As summarized by Müller (1999), the actor-centered institutionalist approach assumes it is the constellation of actors, shaped by a given structural-institutional setting, which determines paradigm choice in old-age security. More specifically, Müller claimed that structural and institutional factors condition largely, which actors become involved in the process of pension reform, as well as their relative strength. In particular, Müller paid attention to the level of external indebtedness and the financial situation of the existing old-age security system. Müller (1999) applied this approach on the analysis of three cases of pension reform – in Hungary, Poland, and the Czech Republic. Based on this analysis, Müller claimed that the main actors shaping the old-age security in Central and Eastern Europe are two Ministries within government – the Welfare and Finance Ministries, complemented by an important external actor, the World Bank15. Secondary internal actors – trade unions, employers’ organizations, pensioners’ associations, social insurance bodies, financial institutions, constitutional courts – may influence the details of the eventual policy design, but not the basic paradigm choice.
15
32
The ’Welfare’ Ministries are called differently in different countries, e.g. Ministry of Labor, Ministry of Social Affairs etc.
From the other side, Orenstein (2000) has indicated that the success or failure of radical reform proposals depends on whether all potential veto actors (i.e. actors who have the power to block the reform) have been included in the process from the reform design phase, as otherwise their exclusion from an early phase may result in harsher opposition at a later stage. Müller ascribes particular importance to the World Bank as an external actor favoring partial privatization of the pension system, indicating that the Bank’s impact is not limited to binding conditionality; rather, what matters is the general level of external indebtedness. According to Müller, the Bank’s influence on the pension reform is exerted first and foremost as an agenda setter, or rather, agenda shifter in the local pension reform debate, but channels to support ongoing pension reforms include expert-based knowledge transfer and loans to strengthen local capacities. The role of the World Bank in shaping pension reforms of Central and Eastern European countries has also been stressed by many other authors (e.g., Deacon and Hulse 1997; Mouton 1998; Orenstein 2000; Nickless 2002; Schmähl 2002; Madrid 2003; Casey 2004; Grimmeisen 2004; Orenstein 2006). Virtually, analysis of the influence of the World Bank on pension reforms has become a separate study-line under the general approach of actor-centered institutionalism. Nevertheless, in spite of this emphasis on the role of the World Bank as an external actor, Müller (1999) views the domestic actors and political process as decisive and focuses attention on the domestic factors that trigger reform (see also Holzmann et al 2003). In respect of the structural-institutional setting, Müller (1999) employs the ‘benefit of crises’ argument, which holds that radical reforms are feasible under critical conditions (see e.g., Drazen and Grilli 1993).16 Müller argues that the fiscal situation of the existing PAYG scheme – whether it is in deficit or surplus – makes an impact on the perceived urgency of radical reforms. According to Müller, the structural setting determines the set and relative strength of main pension reform actors. Radical pension reform becomes feasible when the Ministry of Finance and the World Bank have stakes in and are able to leverage the reform process, whereas radical reform does not proceed when the Welfare Ministry is the key reform actor. A severe financial crisis may strengthen the position of the Ministry of Finance, and high debt may enhance the advantage of international financial institutions that advocate paradigmatic reform (Holzmann et al 2003). Müller (1999) also argues that new institutional arrangements, such as special pension reform commissions, may play an important role in advancing a radical policy change, e.g. by circumventing opposition to reform. 16
The ‘benefit of crises’ argument in actor-centered institutionalism can be seen as a counterpart of the ‘critical juncture’ point argument in historical institutionalism.
33
structural factors
problems
main actors in constellations in situations
basic paradigm choice
interaction of reform team with secondary actors
policy outcome
institutional factors Sources: Scharpf (1997:44), Müller (1999:174)
Chart 1. Explanatory framework of actor-centered institutionalism
Müller (1999) demonstrates that this explanatory model is successful in explaining why a radical, paradigmatic pension reform was chosen in Hungary and Poland, while it was refused in the Czech Republic. At the same time, the findings of Müller seem to indicate the non-applicability of historical institutionalism to the country cases of Hungary, Poland, and the Czech Republic, as these countries shared a common legacy in old-age security but opted for markedly different pension paradigms. Chłon-Dominczak and Mora (2003), addressing the questions of what triggers pension reform, what the motives are behind it, and what are the necessary preconditions for adoption of paradigmatic pension reform, conducted a survey of policymakers and pension experts from 25 different countries, of them 13 from countries, which had introduced multi-pillar systems and 12 from ‘non-reformer’ countries. The focus of the study was on Central and Eastern Europe, with some examples from the EU countries and Latin America. Based of their analysis, Chłon-Dominczak and Mora (2003:134) indicate that the probability of radical pension reform taking place seems to be independent of the age structure of the population because population ageing varies substantially among countries that have implemented a paradigmatic reform. Put into the language of actor-centered institutionalism – the actual demographic context factors have little influence on the reform decision and therefore cannot explain the extent of reform, its timing, or the shift to funding (see also Holzmann et al 2003). This seems to confirm the assumption of Müller (1999) that contextual factors do not actually determine a particular policy choice17.
17
34
However, as indicated above, the contextual factors (including the demographic situation) clearly influence the available policy alternatives.
Using a probit regression analysis of survey results Chłon-Dominczak and Mora (2003) showed that the size of fiscal deficit in the pension schemes was statistically significant as a trigger of pension reform. On this, Chłon-Dominczak and Mora (2003:143) conclude that when the deficit is higher, there is a higher probability that a country will reform its pension system. The latter observations however conflict with the arguments of Müller (1999), who claimed that structural settings matter primarily insofar as these determine the relative strength of main pension reform actors: fiscal imbalances of the pension system for the Ministry of Finance, external indebtedness for the World Bank. In contrast, Chłon-Dominczak and Mora (2003) find that the fiscal deficit of the existing pension system is a significant trigger of pension reform on its own, regardless of the institutional setting. Chłon-Dominczak and Mora (2003) find little explanatory power in claims about the impact of institutional arrangements on pension reform. They demonstrate that paradigmatic reforms have occurred in a variety of institutional formats. Instead, they emphasize the role of leadership as a key variable. Studying the relative importance of domestic and international actors, the findings of Chłon-Dominczak and Mora (2003) from survey data broadly confirm observations of Müller (1999), in that the Ministry of Welfare (whatever its actual name) and Ministry of Finance were the two most influential institutions in the reform process. However, contrary to the arguments of Müller (1999), the Ministry of Welfare received highest ranking as the main agenda-setter or coordinator of the reform process according to the survey data. While survey respondents confirmed that the World Bank was the single most active international institution in the field of pension reform (having been active in 16 out of 25 countries covered by the survey), in contrast to Müller (1999), Chłon-Dominczak and Mora (2003) observed no systemic difference in reform outcomes that could be explained by the involvement of international financial institutions (see also Holzmann et al 2003). Later, Guardiancich (2004) has used the theoretical framework of actor-centered institutionalism to compare the Polish and Slovenian cases of pension reform. While Guardiancich finds this framework suitable for studying welfare state developments in Central and Eastern Europe, he indicates that the explanatory model of Müller does not fully fit to the Slovenian case, relating this to the ‘small N’ problem which affects Müller’s research and resulting early theories18. In particular, Guardiancich indicates that positions of the Ministry of Welfare and the Ministry of Finance towards paradigmatic pension reform could be much more ambiguous than the straightforward defense of their constituencies as ascribed by Müller. In Slovenia, the standard roles (as suggested by Müller 1999) of the Minister of Welfare and the Minister of Finance were inverted – there was a case where the Minister of Welfare was pro-reform, while the Minister of Finance was not.
18
By 1999 Hungary and Poland were the only countries in Central and Eastern Europe which had implemented a paradigmatic pension reform.
35
According to Guardiancich (2004), the ambiguous stance of the Minister of Finance towards a radical reform proposal can be explained by various trade-offs the Ministry has to face, e.g. while preferring reforms, which reduce future fiscal obligations, short-term fiscal conversion of the implicit pension debt (i.e. shouldering of transition costs) may prove extremely difficult. In an attempt to explain the surprising position of the Minister of Welfare, Guardiancich (2004:56) points to a possible cognitive factor – long periods of immobility and stalemate in the reform process in Slovenia may have led the Minister of Welfare to prefer radical reforms. Guardiancich further points out that in contrast to the situation in Poland, the Slovenian system was not perceived as unfair by the population. He ascribes this to the fact that in Slovenia both contributions and benefits were individualized already during the 1970s, employers’ contributions reflecting the wages of individual employees. In essence, the theoretical framework of actor-centered institutionalism is also supported by Clasen (2005). Based on comparison of welfare state reforms in Germany and the United Kingdom, he claims (id. p.23) that a particular alignment of actors’ preferences (motives), institutional capacities at their disposal (means), and contextual conditions (opportunities) impinges on the timing, but also the type and scope of policy change. However, it is noteworthy that in contrast to Müller’s analysis, which was based on pension reforms in three Central and Eastern European countries, Clasen finds (2005:9) that in Germany and the United Kingdom institutional aspects were generally less relevant than contextual changes (e.g. altering employment patterns and household structures) in influencing the dynamics of policy preferences. 2.3.3. Ideational approaches Andersen (2001) has noted that dominant approaches in welfare state studies have often difficulties to explain why some of the welfare states that have undergone the most far-reaching changes are those where the actual challenges appear rather modest. According to Andersen, research often seems to over-emphasize economic challenges and under-emphasize the importance of interests, and in particular the formation of ideas, perceptions, and preferences among decision-makers. Andersen argues for a more constructivist approach to the ‘challenges’ facing contemporary welfare states. Andersen (2001) observed that welfare states find quite different solutions to the challenges they are facing. What derives from this is that these challenges do not clearly point in one direction. Therefore, as Andersen notes, more emphasis should be given to the social construction of challenges – which may be quite diverse – rather than to the economic challenges of welfare states. Andersen (2001) indicates that in a situation, where the level or quality of current public welfare support is perceived as insufficient, or its future uncertain, one of the possible paths is an increased demand for private welfare. Andersen argues that if 36
trust in public welfare deteriorates, it will increase demand for private welfare among those who can afford it, which in turn can trigger lower support for public welfare as people do not want to pay twice for welfare. Also the mechanisms of policy learning through dissemination of ideas point to the relevance of constructivist approaches – if policies move in one direction across countries, this may stem not only from the nature of challenges encountered but also from a similar construction of challenges (Andersen 2001).
Sources: Andersen (2001); own modifications
Chart 2. Possible alternative impacts of perceived quality/sufficiency of public pensions on willingness to contribute to the public pension system and preference for private alternatives
37
Aidukaite (2004) in her study of emerging post-socialist welfare state in the Baltic states also notes that people’s attitudes towards distributive justice and the shift to individualism in people’s mentality has played an important role in shaping social policies. Aidukaite argues that social policy should be studied not only as embedded in the welfare-economy nexus, but also in the societal and cultural nexus of a given society. Also Jochem (2005) criticizes the dominant institutional theories, which, according to him are increasingly questioned because of ‘empirical puzzles’, i.e. inability to explain why welfare reforms went in different directions in countries where the institutional designs of the welfare state were previously similar, the institutional decision-making frameworks were similar, and even the challenges to these welfare state were largely similar? While he agrees that institutional theories of welfare change have several merits, he argues that these theories also suffer from at least two major weaknesses (id. p.: firstly, the micro-logic of the reform process is often faded out by institutional explanations and secondly, institutional approaches often pay only limited attention to the functional environment in which welfare institutions are embedded. According to Jochem (2005), institutional theories cannot provide us with causal mechanisms; rather, they should be regarded as frameworks for empirical research19. Jochem suggests that explanations of welfare state reform should pay more attention to the causal chain in the reform processes and need to be accomplished by empirical observation of the reform process and the embedding of social security schemes into the broader functional framework. Béland (2005) suggests that a more careful study of ideational processes can shed new light on crucial empirical puzzles that traditional historical institutionalism is unable to solve alone – for example, why are some issues becoming important for policymakers while others are not even considered by them? According to Béland, the answer to this question lies in the study of the policy agenda and its change over time. However, Béland notes that ideational processes are not always decisive for explaining policy outcomes. Chłon-Dominczak and Mora (2003:135) point out that ideology has been the least investigated factor behind pension reform. The main reason for this is that despite its importance, it is difficult to identify the driving forces behind changes in ideology over time. Chłon-Dominczak and Mora emphasize the importance of ideas and, in particular, as a crucial determinant of adopting or refusing the paradigmatic reform option, the presence, or absence of a reform consensus among key pension experts and policymakers.
19
38
Jochem (2005) differentiates between theories that inform about generalized causal mechanisms and approaches that in contrast structure the perspectives and methods of research.
Earlier, Pierson (1994) had claimed that the 1986 pension reform in United Kingdom was ideologically driven. However, according to Chłon-Dominczak and Mora (2003), it would be superficial to subscribe pension reforms in Central and Eastern Europe simply to the spread of neoliberal ideology. They point out that in many cases, pension reforms in CEECs were adopted by a non-conservative government, which often opposed neoliberalism. Furthermore, they argue that the idea of mandatory fully funded pensions is contentious in the neoliberal school of thought. Therefore – they conclude (id. p.135) – the influence of neoliberal ideology can only explain the general perception of the need to reform pension systems and preference for more reliance on private savings, but not the whole reform. Bönker (2004) in studying the spread of multi-pillar systems in Europe identified four explanatory factors – the exhaustion of the old paradigm, the compatibility of the new paradigm, the presence of role models and the support by major interest groups. He claims that these factors help to understand why the multi-pillar paradigm has gained more ground in Germany and Sweden than in Finland, France, and Spain. According to Bönker, in the German case, trust in the social insurance paradigm has been especially low, while a fertile ideological ground, the existence of a ‘positive’ foreign model and the ambivalent position of trade unions have further favored the spread of the multipillar system. In Sweden, the paradigm shift was favored by an extraordinarily strong decline in trust in the old paradigm in the early 1990s and a number of supportive ideological traditions. In addition, Schmähl (2002:14) recognizes that “after the overwhelming dominance of public pensions, a shift towards private pension arrangements could be expected.” What therefore is largely missing in the analysis of pension reforms from the actorcentered institutionalist approach is the cultural context – the dominant value positions in countries concerned, their perceptions towards existing and prospective social policy arrangements. These elements are emphasized by Pfau-Effinger (2004; 2005). Although this dimension is sometimes recognized as a part of the contextual setting, which determines the cognitive approaches of actors, it has not been used as an explicit factor explaining differences in reform courses across countries. Pfau-Effinger (2004) demonstrates that causes for path deviant process and starting the dynamics of a new path do not always relate to any specific ‘critical juncture’ points. Instead, the process may be based on conflicts, negotiation, and compromises and the main cause of the whole process may be the increase of a ‘basic contradiction’20.
20
Pfau-Effinger analyzed changes in German and Dutch family and eldery care policies, in which case the basic contradiction was seen as: the contradiction between the individualized, autonomous citizen who integrates into society by the labor market on one hand, and the construction of the housewife marriage on the basis of personal dependence of the housewife and exclusion from the labor market on the other.
39
On this basis, Pfau-Effinger (2004) identifies the elements that are decisive for the start of a new path: contradictions at the level of culture, and between culture and structure. Simultaneously, a group of social actors is needed who refer to the conflict and try to establish a new cultural basis of welfare state policies. Earlier, De La Porte (2000) identified the cultural code – ethos – as the underlying force driving the selection of policies. ‘Welfare culture’ as defined by Pfau-Effinger (2005:4) comprises the relevant ideas in a given society surrounding the welfare state and the way it is embedded in society. It includes the stock of knowledge, values, and ideals to which the relevant social actors, the welfare state institutions, and concrete policy measures refer. The values and ideals, which predominate in the welfare culture, restrict the spectrum of policy alternatives that are taken under consideration. Finally, as noted above, the three theoretical frameworks of historical institutionalism, actor-centered institutionalism, and ideational approach should not be regarded as mutually exclusive. Although they have emerged relatively lately and are still rare, there are combined approaches. For example, Pfau-Effinger (2004) has combined ideational approach with historical institutionalism, whereas Bönker (2004) combines ideational approach with actor-centered institutionalism.
40
3. METHODOLOGY AND RESEARCH METHODS OF THE THESIS 3.1. RESEARCH STRATEGY, DESIGN, AND METHODOLOGICAL ISSUES The multi-dimensional nature of pension systems allows for a number of different research approaches to be taken to study them. While historically pension systems are a product of the industrial revolution of the late 19th century, research of pension systems has intensified only over the last two decades. This has been largely due to the new challenges pension systems in many countries have to face, in particular the ageing of populations and changing labor markets. Research has responded to these practical needs, looking for more sustainable policy designs, but also studying the process of policy changes. At the same time, policies have responded to practical challenges, often using a “trial-and-error” method. As pointed out by Roland (2000), any process of large-scale institutional change involves a genuine uncertainty about outcomes. Even when carefully planned and designed, pension reforms are large-scale social experiments, because multiple influencing factors make precise prediction of reform outcomes virtually impossible. In this context, the role of research is to monitor the reform process to provide useful feedback. Clearly, there are research interests with more modest practical value, which can also broaden our understanding of social processes. The current study combines the research strategy of explanatory case study (Yin 1994) with elements of cross-case analysis (see e.g., Huberman and Miles 1998). The case, which is analyzed here, is the transformation of the pension system in Estonia that took place during 1990–2005. Case studies and cross-national comparative studies are the dominant strategies in existing studies of welfare state transformation, including pension reform. However, cross-national comparison is not the aim here. For the time being, there are already a number of studies on pension reforms of Central and Eastern European countries (e.g., Müller 1999; Fultz 2002a, 2002b, 2006; Schmähl and Horstmann 2002; Chłon-Dominczak and Mora 2003; Schmähl 2003; Casey 2004; Grimmeisen 2004; Whitehouse 2004; Schubert 2005). Against this background and given the common caveats of crossnational comparisons of policy reform studies, the author believes that a detailed study of a single case may give a greater contribution to the already extensive literature than a repeated exercise of cross-national comparison. As the case study approach implies, transformation of pension system is investigated in its’ ‘real-life’ context, i.e., in the context of simultaneous economic, political and demographic changes. The study goes beyond descriptive overview of changes in the Estonian pension system during the transition period, to explain which factors have triggered reforms, to explore the rationale behind policy choices and to evaluate the impacts of pension policy.
41
The key unit of analysis of the thesis is pension policy: capturing here policy alternatives, policy design elements (parameters), as well as policy outcomes. While studying policy changes over the period of 1990-2005, I am also interested in reasons for these changes. The approach applied here is consistent with GreenPedersen (2004), who argued that the conceptual capturing of welfare reform could not be isolated from causal accounts of change. The basic proposition of the study is that the transformation process is a complex phenomenon where multiple influencing factors are in play (see e.g., Aidukaite 2004): domestic (including economic, demographic, historic, social, political, cultural) and international (e.g. international organizations, reforms in other countries etc.).21 Therefore, different aspects and their interplay have to be addressed to explain pension reform. Siegel (2003) has raised the question about the ‘key independent variable’, i.e. what are the principal determining causes of change among various influencing factors, e.g. structural socio-economic forces, political actors, influence of institutions or policy legacies (see also Amenta 2003). However, the ‘key independent variable’ may not always exist, as the determining causes can be a constellation of several factors with simultaneous impact accounting for a policy change. While studies of welfare state developments have a history of longer than three decades, methodological aspects in this field of research have only received deeper consideration relatively recently. Several authors have pointed out that welfare state analysis should be disaggregated to investigations into the dynamics of particular policy domains in order to be better able to capture the details of policy outcome (Huber and Stephens 2001), as well as processes of reform (Pierson 1996; Ferrera et al 2000). The current study follows this advice and concentrates on pension policy. Obviously, pension policy is embedded in the broader context of social policy and influenced by this context. However, at the same time pension policy has its specific features. Since the aim is to capture within the field of pensions the specificity of policy choices and policy processes, the broader social policy context and policy changes in other social security schemes have been largely omitted. In addition, the focus on main policy changes does not allow going into multiple technical details. To some surprise, questions related to dependent variables in welfare state studies have been raised only quite recently. Green-Pedersen (2004) was among the first to raise the ‘problem of dependent variable’ of welfare state changes. According to Green-Pedersen (2004), the ‘dependent variable problem’ is a problem of theoretical conceptualization rather than a problem of data. He points out that different conceptualizations lead to different evaluations of the same changes in welfare schemes. For Green-Pedersen, the key to solving the ’dependent variable problem’ relates to appropriate conceptualization and operationalization of welfare state changes. 21
42
Following the case study strategy, I do not formulate a hypothesis and do not apply hypothesis-testing. Instead, the basic proposition is formulated (based on existing knowledge and earlier studies of pension reform) to give general directions for the analysis.
Clasen and Siegel (2005) suggest that insufficient discussion of the dependent variable problem is a general shortcoming of many studies that have described and analyzed welfare state change. They agree with Green-Pedersen in that the underlying problem of the dependent variable is linked to the question of finding an appropriate concept for welfare state change. They also point to the potential problem that inappropriate conceptualization may exclude some important dimensions of changes from the analysis. However, at the same time Clasen (2005:15) warns: “any approach that aims at reducing the multifaceted reality of welfare state changes to one dominant concept of change pays a high price: it runs the risk of oversimplifying patterns and contents of change”. In the context of the current study, the dependent variable problem translates into questions like what constitutes a change of pension policy and how can these changes be measured? The approach of Hall (1993) is exemplified to determine what counts as policy change (see Chapter 2). In respect of the measurement problems, recommendations of Siegel (2003) have been followed. He has suggested combining different research strategies and levels of analysis for ‘poly-centered’ perspectives on welfare state change. As noted by Siegel (2003:31), macro analyses could be used for some basic descriptive and analytical purposes, combined with an analysis on the lower level of abstraction, e.g. with a program specific case oriented analyses to “accumulate substantial knowledge about the political logic of welfare state change”. In essence, the ‘poly-centered’ approach has also been advocated by Green-Pedersen (2004) and Clasen (2005), who suggest that analysis of quantitative indicators (e.g. expenditure) shall be accompanied by indicators which signify changes in ‘social rights’, (e.g. entitlement criteria, benefit level), because quantitative elements can hardly assess – even at the most disaggregated level of analysis – the implications of legislative changes for the actual scope of programs. This advice has been largely followed here. 3.2. DATA AND METHODS While, in principle, transformation of pension systems can be studied from different perspectives, e.g., from economic (incl. political economy), political science, sociology etc., the current thesis attempts to take primarily a social policy perspective. In this respect, reference is made to Walker (1983:141), who defines social policy as “the rationale underlying the development and use of social institutions /…/ which affect the distribution of resources, status, and power between different individuals and groups in society. Social policy is concerned with both the values and principles that govern distribution as well as their outcome. The task of the social policy analyst is to evaluate the distributional impact of existing policies on social welfare, their implicit and explicit rationales, their impact on social relations, and the implications of policy proposals.” Along these lines, the current study analyses the development of ideas on pension policy, the underlying principles which govern distribution mechanisms inherent in pension policy as well as distributional outcomes of reforms on macro and (to a lesser extent) micro level. 43
In this perspective, the study could be also categorized as a policy analysis. The basic methods of policy analysis are applied (see Patton and Sawicki 1986) using both qualitative and quantitative data. The analysis of pension policy developments combines historical reconstruction with descriptive statistics. In respect of analysis of policy developments, the ontological position of the author is that of constructivist. The author of the current study has been personally involved in the pension reform process during the period 1997–2001 – first in the capacity of a Deputy Director of the National Social Insurance Board and later as a Social Security Policy Advisor of the Estonian Ministry of Social Affairs, simultaneously being a Member of the Government Social Security Reform Commission. It is clear that at least in respect of the abovementioned period, the author can only offer his specific version of social reality, rather than a definitive account of past events. Although not being directly connected to policy-making in the area of pensions before 1997 and after 2001, the ´insider´ position of the author in 1997–2001 obviously has some influences on interpretations of policy developments also in respect of these periods. The background of the author bears both advantages and limitations. From one side, having been closely associated with the policy-making process, the advantage of the author is a much closer insight into the process compared to those researchers who have merely observed these developments from the outside. To overcome the limitation – possible bias in respect of certain policy choices or explanations – a triangulation method (see e.g., Webb et al 1966; Denzin 1970; Bryman 2001) is used, where possible, to compare the observations of the author with findings from other observers. Triangulation is also used in the sense that multiple sources of evidence are used, when possible. The sources of analysis on evolution of the Estonian pension policy include: • legal texts: over 50 acts and decisions adopted by Riigikogu (Estonian parliament), Government regulations and other implementation acts (e.g. decrees by Ministers); • coalition agreements; • policy papers and conceptual documents of the Government and Social Security Reform Commission; • annual state budgets, incl. pension insurance budgets; • reports of the European Commission, IMF press releases etc. The analysis of qualitative material focuses on changes in key pension policy parameters: eligibility rules (e.g. pensionable age, qualification period), benefit calculation rules (pension formula), financing arrangement, and institutional set-up. This analysis takes a dynamic perspective, looking at the background factors and pre-conditions of reform, identifying the influencing factors and significant actors that have accounted for evolution and transformation of the Estonian pension system. This is followed by analysis of main trends in pension policy outcomes, using descriptive statistics. For the latter, secondary data is used – mainly macro-level aggregate statistical data from the Estonian Statistical Office, Ministry of Social Affairs, and Social Insurance Board. The main indicators here are the total number 44
of pensioners and breakdown by types of pension; newly granted pensions (influx of new retirees); system dependency ratio; revenues and expenditures of the pension system (level and structure); benefit level (in nominal and real terms); average replacement rate. The analysis of policy outcomes on micro level is based on individual data of pension recipients and insured persons, obtained from administrative databases, i.e. from the Pension Insurance Register of the National Social Insurance Board, Estonian Central Depository of Securities, and Financial Supervision Authority. The main indicators analyzed include coverage rates of different pension pillars, distribution of pension amounts, and individual replacement rates. In this respect the analysis builds upon a pension study by the PRAXIS Center of Policy Studies (Tiit et al 2004), where the author of the current thesis was involved as a member of the expert team.
45
PART II. THE STUDY OF THE ESTONIAN PENSION SYSTEM 4. PENSION REFORM PROCESS IN ESTONIA 4.1. BACKGROUND FACTORS The pre-reform scene in Estonia, like the other Baltic countries, differed from that in other countries of Central and Eastern Europe, which are now new members of the European Union. The Soviet legacy of nearly 50 years shaped not only the pension system, but virtually all facets of society. Although countries like Hungary and Poland also belonged to the socialist block and were heavily influenced by the Soviet Union both through economic ties and general politics, they nevertheless had some degree of autonomy, including autonomous pension systems. In addition, the adjustment to market conditions and related social reforms after the breakdown of the Soviet Union followed somewhat different patterns in the Baltic states. The economic recession in the first half of 1990s was more severe in the Baltic countries compared to Poland, Hungary, and the Czech and Slovak Republics. Also, the demographic changes in the Baltics were more dramatic, including decline in the birth rates and increase in the mortality rates.22 4.1.1. Demographic changes The processes of political, economic, and social transition associated with Estonia’s regaining of independence had major demographic impacts. The steady growth of the population in the 1970s and 1980s – caused by relatively high birth rates and immigration – was replaced by a rapid decline beginning in 1991. This was rooted in a host of negative developments: negative net migration, decline of birth rates, and increase of mortality rates (Estonian Statistical Office 2003). The most drastic changes occurred in 1990–95. Later the situation has gradually stabilized: migration declined, birth rates stabilized (albeit on a considerably lower level) and life expectancy even increased. However, as a net effect of demographic changes, by 2005 the total population had declined to the level of the early 1970s (Chart 3).
22
46
For further details, see, for example, Schmähl and Horstmann (2002).
1600000 1550000 1500000 1450000 1400000 1350000 1300000 1970
1975
1980
1985
1990
1995
2000
2005
Source: Estonian Statistical Office
Chart 3. Population of Estonia (at the beginning of the year), 1970–2005
Parallel to the decrease of total population, the population also began to age (see Katus et al 1999). The median age increased from 34 in 1989 to 38 in 2005, while over the same period the share of the population aged 65 or over increased from 11.5 % to 16.5 % (Chart 4). 80% 70% 60% 50% 40% 30% 20% 10% 0% 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 0-14
15-64
65+
Source: Estonian Statistical Office
Chart 4. Population age structure, 1989–2005 47
Ageing of the population is caused mainly by a decrease in the birth rate and increase in life expectancy23. The total fertility rate was above the rate of reproduction needed to maintain the population (i.e., 2.1 births per female) in 1990, but declined rapidly in the following years to an all-time lowest 1.28 in 1998. However, the rate had begun to stabilize in the second half of 1990s and from 1999, a slight increase in the total fertility rate has been observed (Chart 5). However, the improvement is still far too modest to change the long-term impact of more than a decade of low birth rates. 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1.0 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Source: Estonian Statistical Office
Chart 5. The total fertility rate, 1987–2004
The average life expectancy at birth declined in the first half of the 1990s, reaching the lowest level in 1994, when it was 60.5 years for men and 72.8 for women. However, since then, life expectancy has been increasing, and in 2004 the average reached 66.3 years for men and 77.8 years for women. The decline of average life expectancy in mid 1990s was primarily due to rising mortality rates for the age group 40–64. At the same time, mortality rates for persons aged 70 and over did not change significantly24.
23
24
48
Katus et al (2002:48) point out a particular factor which contributed to an increased rate of population ageing in 1990s – the first cohorts of post-Second-World-War Russian immigrants reached old age. However, see Katus (1998) on reliability of population data, including mortality rates for higher ages.
The difference between the life expectancy of men and women at birth is considerable, nearly 12 years. By the age of 60, however, this difference declines to 6 years25. Table 5. Life expectancy at birth and at age 60, 1990–2004
Men Women
0 60 0 60
1990 64.5 14.7 74.7 19.4
1992 63.4 14.5 74.6 19.7
1994 60.5 13.8 72.8 19.2
1996 64.1 14.7 75.4 20.1
1998 64.4 14.8 75.5 20.3
2000 65.1 15.2 76.0 20.7
2002 65.2 15.4 77.0 21.2
2004 66.3 15.4 77.8 21.7
Source: Estonian Statistical Office
Net migration has been negative since 1990, as emigration exceeded immigration considerably. Immediately after Estonia regained independence, a large proportion of the non-Estonian population moved, mostly to Russia and other parts of the former Soviet Union. Emigration peaked in 1992 and has shown a decreasing trend since then. In the second half of the 1990s, net migration approached zero as emigration had declined to the level of immigration (Chart 6). 40000 30000 20000 10000 0 -10000 -20000 -30000 -40000 1989
1990
1991
1992
1993
Emigration
1994 Immigration
1995
1996
1997
1998
1999
Net migration
Source: Estonian Statistical Office
Chart 6. Emigration, immigration and net migration26
25
26
For a comparative analysis of main demographic indicators (total fertility rate, life expectancy, median age, share of population over 65 etc.) of Estonia and other EU Member States, see Leetmaa et al (2004). In 2000, the Estonian Statistical Office stopped publishing migration data, because of the low quality of administrative data. The main problem is that changes of residence are underreported, as registration of the place of residence is not compulsory.
49
4.1.2. Economic background Transition to a new economic system and reconstruction of the economy caused sharp decline of Estonia’s GDP during 1991–94. After regaining its independence, Estonia chose a radical approach to economic restructuring, re-shifting economic ties from East to West. The cornerstones of economic reform were a currency board arrangement, a liberal trade policy, a balanced budget doctrine with very limited state borrowing, and far-reaching privatization (see e.g. Eamets 2002). These principles were followed by successive governments in spite of changes in coalitions. A major factor in stabilizing the economic situation (see Table 6) was the introduction of the national currency – Estonian kroon (EEK) – in June 1992 based on a currency board arrangement with a fixed exchange rate.27 Table 6. Real economic growth and inflation, 1992–2005 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Real GDP growth (%) -14.2 -9.0 -1.6 4.5 4.4 11.1 4.4 0.3 7.9 6.5 7.2 6.7 7.8 9.8
CPI change (%) 1076.0 89.8 47.7 29.0 23.1 11.2 8.2 3.3 4.0 5.8 3.6 1.3 3.0 4.1
Source: Estonian Statistical Office, Bank of Estonia, data on GDP growth for 1992 and 1993 from Sillaste (1998)
The measures undertaken helped the small open economy of Estonia to overcome the recession phase by 1995 (see e.g. Hansson 1997). In the following years, the economy recovered quickly. Economic growth was mainly driven by a rapid growth in exports to Western countries, reinforced by a high level of foreign investments. Economic growth reached 11.1% in 1997. In 1998, Estonia experienced an economic slowdown due to a crisis in the financial sector combined with a decline in foreign demand (including a major crisis in the Russian market). In 1999, there 27
50
On 20 June 1992, Soviet roubles were converted to Estonian croons (EEK) at the rate of 10 roubles per kroon. The exchange rate for the kroon was pegged to the German mark (DEM) at the rate of 1 DEM = 8 EEK. From 2002, the Estonian kroon is pegged to Euro at the rate of 1 Euro = 15.64664 EEK.
was an economic stagnation, GDP increased by mere 0.3%. In 2000, the economy recovered rapidly and posted a GDP increase of 7.9 %. High growth continued in 2001–2005. Currency reform cut the inflation rate, which was as high as 1100 percent in 1992, to 90 percent in 1993 (see e.g. Korhonen 2003). The inflation rate continued to decline steadily through 1999. Economic revival after the 1999 economic recession and the increase of some administratively regulated prices (e.g. electricity, heating, public transportation) increased the inflation rate in 2001, when it reached 5.8%. However, in 2002-2003 the inflation rate declined again, reaching a historic low of 1.3% in 2003. Estonian inflation has in recent years followed the dynamics of the Euro area (however, at slightly higher level), indicating the close links of the Estonian economy with the EU through trade channels and capital markets (Masso and Staehr 2005). The economic downturn of the first half of the 1990s was coupled with declines in both the labor force participation rate and the employment rate. This adjustment to the market economy marked the end of the Soviet period of full-employment (see e.g. Eamets 2001). The employment rate in the age group 15–64 declined from 77.4% in 1990 to 60.7% in 2000. Over the same period, the unemployment rate increased from 0.6 to 13.8% (Chart 7).
90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1990
1991
1992
1993
1994
1995
1996
1997
Employment rate
1998
1999
2000
2001
2002
2003
2004
2005
Unemployment rate
Source: Estonian Statistical Office Note: Data on unemployment rate for 1990–1992 is for the 15-69 age group
Chart 7. Employment and unemployment rates for the age group 15–64 (percent), 1990–2005
51
Notably, the employment rate continued to decline in the second half of the 1990s in spite of the economic recovery, i.e., Estonia experienced jobless growth. Even the very high growth rate in 1997 did not significantly affect the employment situation. After a period of a relative stability in 1995–1998, the employment rate declined and the unemployment rate increased again in 1999 and 2000. Some correlation between GDP growth and the employment rate can been observed only beginning in 2000. Annual growth rates of around 5–7% were accompanied by an increase in the employment rate from 60.7% in 2000 to 64.0% in 2005, while over the same period the unemployment rate declined from 13.8% to 8.1%. The favorable economic developments of last years have taken the Estonian unemployment rate below the EU average (8.7% in 2005). In absolute numbers, the total labor force in 2005 was 654,000 persons, down from 831,000 in 1990. The number of employed persons was 602,000 in 2005 as opposed to 826,000 in 1990. In 2005, the number of unemployed persons was around 52,00028. As a part of the Soviet legacy, the employment rate of women has been relatively high. In fact, over the transition period their employment rate declined less that the employment rate of men. In 2005, the employment rate of men was 66.2% whereas for women the rate was 61.9%, which is over the EU average29. Contrary to the situation in several other European countries, women have a slightly lower average unemployment rate than men. In 2005, the unemployment rate of men was 9.0% as opposed to 7.2% for women. A similar gap has existed since the mid 1990s. The decline of the employment rate in the first half of 1990s was most significant for older workers, i.e. the age group 55–64. However, beginning in 1995, this trend reversed itself and by 2003, the employment rates in the 55–64 age groups rebounded to the level of 1991–1992 (Leetmaa et al 2004). At the same time, employment rates in the 15–29 age groups continued to decline until 2004. Other noteworthy trends in the 1990s were the restructuring of the labor force between economic sectors and changes in the types of employment. Namely, employment in the agricultural and manufacturing sectors decreased while the service sector grew (Eamets 2001). While total employment declined, the number of self-employed persons increased and comprised in 2004 about 6% of total employment. Maintaining the competitiveness of Estonian economy has been a major concern of all governments starting in the early 1990s. In addition to the policy measures mentioned above, tax policy has been a used in this respect. Estonia abandoned progressive income tax and introduced a proportional income tax in 1994. In 2006, the tax rate is 23% beyond an annual threshold, which is 24,000 EEK. In 2000, the 28 29
52
Here, the total labor force, number employed and unemployed persons are for the age group 15–69. Employment rates and unemployment rates are shown for the age group 15–64.
corporate income tax on reinvested profits was abolished30. According to adopted legislation, the income tax rate is set to decline gradually to 20% by 2009. The strict principles of a balanced budget, limited borrowing, and lower taxes have obviously limited available public finances. Overall public spending in the second half of the 1990s was around 37–40% of GDP. In 1999, it rose to 40.4% due to optimistic assumptions about economic performance in the state budget, while in the reality the GDP increased by marginal 0.3% in 1999. In the following years, the conservative approach to drafting state budget was restored, resulting in a surplus in the total government sector balance beginning in 2001. As a result of tax cuts (abolition of the corporate income tax on reinvested profits), total government expenditures declined to 35% of GDP in 2001 and 2002. Economic growth, the increase of employment rate, and transfers from the EU funds increased state revenues in 2002–2005. On the other hand, EU and NATO-related commitments increased also expenditures (Chart 8). 41% 40% 39% 38% 37% 36% 35% 34% 33% 32% 1996
1997
1998
1999
Total government revenues
2000
2001
2002
2003
2004
Total government expenditures
Source: Ministry of Finance, Estonian Statistical Office
Chart 8. Total government sector expenditures as a percentage of GDP, 1996–2004
Chart 9 illustrates the government fiscal policy. From 2000, Estonia has followed the EU’s Maastricht criteria on the budget deficit (not to exceed 3% of GDP). Over the last 5 years, the total government sector has run a surplus (Eesti Vabariik 2005).
30
See Funke and Strulik (2003) on welfare effects of corporate income tax reform.
53
3.0% 2.0% 1.0% 0.0% -1.0% -2.0% -3.0% -4.0% -5.0% 1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Surplus/deficit
Source: Estonian Statistical Office Note: 2005 – preliminary data
Chart 9. Balance of the total government sector as a percentage of GDP, 1994–2005
The total government sector debt has been rather low, around 5–6 % of GDP (Eesti Vabariik 2005). In 2004, the total government sector debt was 5.4% of GDP, which is the lowest among the 25 EU Member States and over 10 times below the EU average. In 1997, a year of very high economic growth, the Government formed a stabilization reserve, to which it directed the surplus of revenues – a buffer fund to soften the effects of possible macroeconomic shocks and to finance major structural reforms, e.g., pension reform. The reserve was further increased in 2001–2005. By the end of 2005, the stabilization reserve amounted to 5.4 billion EEK or about 3.5% of GDP. About 95% of its assets are invested in German, French, Dutch, and Belgian government bonds. The total reserves of the government (including shortterm cash reserves) amounted to 14 billion EEK (or 10% of GDP) by the end of 2004, or more than double the total government debt (Eesti Vabariik 2005). 4.1.3. Pension system up to 1990 Estonia, like the other two Baltic countries, inherited its benefit systems from the Soviet Union. Until 1990, pension arrangements in Estonia were a part of the Soviet pension system. The Soviet legacy of the pension system included the following features: 1) low general pensionable age – 55 for women and 60 for men; 54
2) privileged retirement rules for several occupational groups, including lower pensionable ages; 3) entitlement to a pension based on previous work, benefits linked to the former wage; 4) a relatively high replacement rate ranging from 100 percent for low-income earners down to 50 percent for higher-income earners; 5) separate schemes for workers and farmers; and 6) financing from the general state budget, no individual contributions by workers. Thus, whereas entitlement rules had Bismarckian features, pensions were financed from the state budget, not from contributions as in a typical Bismarckian scheme. Table 7. Statutory replacement rate under the Soviet pension law Former monthly wage (in roubles) Up to 35 35 to 50 50 to 60 60 to 80 80 to 100 100 or over
Pension as percentage of former wage 100 85 75 65 55 50
Source: Soviet Union Pension Law 1956
By 1990, Moscow had recognized the need to reform the pension system, and under the glasnost campaign a public debate on a draft of a new Soviet pension law was commenced. However, a strong national aspiration for independence had developed by that time, and the political aim of the Estonian government was to separate all social systems from the rest of the Soviet Union (Leppik and Männik 2002). Yet despite this goal, the features of the Soviet pension system influenced people’s image of the optimal pension arrangement, including such features as the pensionable age, benefit rates, and the willingness to pay contributions (or rather the lack thereof). The Soviet heritage of a low pensionable age and relatively high replacement rate was clearly not a favorable starting point for an autonomous pension system, especially given that the cost of pensions had been largely hidden in the socialist period.
55
4.2. THE EARLY REFORMS 1990–199331 Early transformation of the pension system in Estonia may be characterized by the following stages (also see Leppik 2002): 1) Financial separation of the benefit system (1990); 2) Failed attempt to liberalize benefit rules (1991); 3) Benefit retrenchment with introduction of flat-rate pensions (1992); 4) Benefit restructuring with the State Allowances Act (1993). 1990 – Financial separation Separation of the Estonian pension system from the Soviet arrangement started from the financial side in 1990 through creation of the Social Fund, while the main benefit rules remained unchanged for another year.32 This was because the deepening economic and fiscal problems in the Soviet system forced the Estonian government to take steps to consolidate the financing of the pension system even before it formally regained independence.33 With the underlying aims of limiting transfers to the Soviet budget and limiting damage from economic turbulence caused by price liberalization, increasing inflation and disturbed cash flows, the government separated the financing of the Estonian pension system from the rest of the Soviet Union (Leppik and Männik 2002). The adoption of the Social Tax Act in 1990 introduced a social tax of 20% of gross payroll to be paid by employers as the financing instrument of the state pension system.34 Revenues collected by the Social Fund were earmarked for this purpose, and pension expenditures were separated from other budgetary expenditures. 1991 – Failed attempt to introduce a new pension law Although the Soviet pension system provided rather high replacement rates, the Estonian government’s first attempts to reform the system were partly motivated by a desire to raise the level of social protection even further. There was a common belief that financial separation would allow for the provision of improved benefits. 31
32
33
34
56
It shall be noted that there are some overlaps between periodizations applied in this study. This relates to the fact that some policy decisions are implemented only after a certain period. As the study analyzes both the policy changes and policy outcomes, these overlaps can not be avoided. From the perspective of policy process, the time of making policy decisions (e.g. time of adopting a concept paper or law) has relevance. However, in the time-series analysis of policy outcome indicators, the time when a particular policy change became effective is relevant. There were a few changes adopted already in 1990, e.g., personal pensions for Communist Party officials were ceased beginning in July 1990; and minimum pension benefits were increased beginning in October 1990. Estonia regained independence on 20 August 1991. See Leppik and Männik 2002 for a more detailed description of the situation in 1989 and 1990. The act was adopted on 12 September 1990 and came into effect 1 January 1991.
The new Pension Act of Estonia, adopted 15 April 1991, had two main objectives – to separate the benefit side of the Estonian pension system from the Soviet system and to increase coverage and the level of benefits. Prime Minister Edgar Savisaar (the leader of the National Front) ordered that no paragraph of the new pension law should be a copy of the text of the Soviet law (Leppik and Männik 2002). The formerly separate schemes of workers and farmers (kolhoz members) were unified into a single system with universal coverage. The new act liberalized eligibility rules, broadening the coverage of the pension scheme to all residents, and prescribed higher pension rates (Leppik 2002). The pensionable age and qualification period for an old-age pension both remained unchanged. However, the qualification period for the invalidity pension was abolished, making access to invalidity pensions easier. The act prescribed a mixed pension formula – a flat rate base amount supplemented by an earnings-related component. The calculation of pensions was based on two factors: the minimum wage and the worker’s former earnings. For example, old-age pensions were calculated as 60% of the minimum wage supplemented by 40% of the average former earnings of the beneficiary.35 The previous supplements for uninterrupted service were abolished. The minimum old-age pension was set at 85% of the minimum wage. The act also introduced a social pension equal to 70% of minimum wage for persons not eligible for an old-age pension. However, the payment of social pension was deferred for 5 years after the general pensionable age. The new act created a quite typical defined-benefit pension scheme financed by contributions. However, high expectations soon collided with economic reality and the act had a very short life, being implemented for only few months (see also Püss 1995). Because of the total neglect of financial calculations, implementation of the act turned out to be unaffordable.36 The failure of the first reform was mostly due to a striking lack of qualified staff able to develop policies and legislation in a coherent way, while the situation was further exacerbated by serious economic crisis at the time of collapse of the Soviet Union (see Leppik and Männik 2002; Leppik and Kruuda 2003). 1992 – Flat rate pensions Parliament suspended the Pension Act in February 1992, and pensions were replaced by flat-rate state living allowances. The introduction of flat rate allowances was a temporary rescue measure to help cope with deep economic crisis rather than
35
36
Similar to the Soviet system, the ‘best years’ approach was used to determine former earnings. These were now determined on the basis of the five best consecutive years within the 15 years preceding the pension application or the end of the working career. Another problem was the need for multiple recalculations of pensions in an environment of very high inflation in 1991–1992. This had to be done manually, as the level of computerization of pension offices in the early 1990s was still very low.
57
an intentional shift toward egalitarian principles.37 With flat-rate benefits, payments were easier to administer and calculations easier to make. The levels of pensions were linked to the minimum wage. To keep pace with the very high inflation (see Table 6), the minimum wage, and the rates of pensions were increased five times during the course of the year. However, in spite of these nominal increases, the real values and the replacement rate of pensions decreased considerably. Katus et al (2004) calculated that the gross replacement rate of the average old-age pension declined from 36 to 16 percent over these turbulent times. The flat-rate pension package thus entailed a substantial benefit retrenchment. The later developments of pension system were strongly influenced by two important reforms in 1992, setting the broader context for various national policies – monetary reform and the adoption of the constitution.38 The currency board arrangement, which limited public expenditures to available tax revenues, set strict budgetary limits for the pension system. The constitution, inter alia, laid down the general principles of social security: Estonian citizens shall be entitled to state assistance in case of old-age, inability to work, loss of provider, and need. The types of assistance, their level, eligibility conditions and procedures shall be established by law. Unless otherwise determined by law, this right shall exist equally for Estonian citizens, citizens of foreign states and stateless persons who are present in Estonia. Notably, social rights were formulated in a rather weak manner in the constitution. Although referring to the traditional social risks insured by the pension system (old age, invalidity, survivors), there is no explicit reference to the concept of ‘pension’. Instead, the formulation refers to ‘state assistance’ with the implicit assumption that the primary responsibility lies with the individual. 1993 – State Allowances Act After the first democratic elections of the Estonian parliament in September 1992, a national conservative government came in power. Public dissatisfaction (in particular from pensioners’ organizations) with the system of flat-rate benefits forced the preparation and adoption of a State Living Allowances Act, which differentiated old-age pensions on the basis of length of
37
38
58
The failure of 1991 pension law was not admitted by politicians. The resolution on flat-rate allowances was supplemented with a decision to reintroduce the pension law three months after the introduction of Estonia’s own currency. However, in the reality, the 1991 pension law was never put back into force. The constitution was drafted by a Constitutional Assembly, adopted by referendum on 28 June 1992, and entered into force 3 on July 1992.
service, marking a shift in the social-political distribution principle.39 The pension formula comprised two elements: a flat-rate base amount and a component depending on the years of pensionable service. Pension amounts still related to the minimum wage, with the individual variable being the length of service. Originally, the base amount was calculated as 85% of the minimum wage. If the length of service exceeded 40 years, 1% of the minimum wage was added to the base amount for each year of service, whereas in case of length of service 15 years, the value of one service year was only 0.5% of the minimum wage (see Table 8 and Chart 10). In the other words, longer service periods had higher values. Notably, the pension had no maximum value. Invalidity pensions continued to be paid at a flat rate, depending only on the extent of invalidity. As an essential change, the 1993 State Allowances Act introduced a gradual increase of the pensionable age by 6 months each year with the target of reaching 65 for men and 60 for women by 2003. With the experience of the failed 1991 pension law, the need to take steps for financial consolidation of the pension system was recognized. The increase of the pensionable age was a reaction to the increase in the number of pensioners and rapid decline in the number of insured workers caused by labor market restructuring, but also reflected a recognition that demographic ageing is on the horizon for Estonia. In fact, in the draft act presented to Parliament, the Government suggested equalizing the pensionable age of men and women at 65. However, when the vote was put to Parliament, female politicians succeeded in changing the scale to 65/60. Although the pensionable age was increased while maintaining gender differences, the qualification period for old-age pension was equalized for men and women: it was reduced from 25 for men and from 20 for women to 15 years for both sexes. Working pensioners were paid full pension only if their earnings were below the minimum wage, otherwise the benefit was reduced. The existing social pension was renamed national pension, with payment levels set at 85% of the minimum wage. Age criteria for the national pension were set at 65 years for men and 60 for women, or five years higher than for a normal old-age pension at that time. However, as the general pensionable age was to be increased, the age criteria for an old-age pension and a national pension would eventually become identical. From the financing side, the act authorized a pension scheme that could be classified as a defined-contribution scheme at the macro level (Leppik 2002). The revenues of the system were set by a fixed contribution rate (i.e. social tax rate), and benefit levels were adjusted according to the revenues available. This closed-budget approach introduced clear fiscal boundaries to the pension system. 39
The Act was adopted 17 March 1993 and entered into force on 1 April 1993. The use of the term ‘state living allowances’ in the title was suggested by politicians who were behind the 1991 pension law, but who landed in the opposition after 1992 elections. The use of this term was intended to indicate that the benefit rules of the law were still temporary, and a true ‘Pension Law’ was missing.
59
4.3. PENSION POLICY OVER THE TRANSITION PERIOD 1993–2000 The pension formula in the 1993 State Allowance Act was broadly considered (by political parties as well as by pensioners) as a temporary solution for a period of economic transition. The political aim was to reintroduce earnings-related pensions in a period of few years. However, the benefit rules, which were established as temporary, survived a period of 7 years – from April 1993 to April 2000 – and exerted a heavy influence on subsequent pension rules for the first pillar. Despite the longer-than-expected duration of these rules, the period of their existence was also marked by some important parametric changes. Beginning in July 1994, pension amounts were decoupled from the minimum wage. This increased flexibility in both the pension system and national wage policy, allowing pensions to increase without changing the minimum wage and vice versa. The calculation of pensions was based on a fixed rate – national pension rate (NPR) – to be established by Parliament for each fiscal year.40 The formula, which was used to calculate the amount of old-age pension from July 1994 to April 2000, may be represented as:41 P = B + E × α (E) × B, where B is the flat-rate base amount, E is the length of service (years of employment and equalized periods), and α is a coefficient which varies with the length of service E. Starting in September 1996, pensions could be received simultaneously with income from work without any restrictions. Accordingly, working pensioners could receive full benefits. Also starting in 1996, the timetable for increasing the pensionable age was modified to remove a perceived injustice between those born in the first half of the year and those born in the second half, inherent in the previous timetable. To provide a smoother and more gradual increase, the rate of increase was reduced from 6 months a year to an average of 4 months a year. The target pensionable age remained at 65 for men and 60 for women, but with the modified scale, these targets were to be achieved by 2007 instead of 2003. From the financing side, the Estonian pension system operated on the macro-level defined-contribution principle, as previously described, during 1994–2000. The rate of social tax to finance the pension system did not change, remaining at 20% of an employer’s gross payroll. In spite of frequent changes in government, successive
40
41
60
The only additional restricting clause was that the new value of the national pension could not be less than that of the previous year. In reality, however, the NPR remained unchanged at EEK 410 from 1996 to 2000. This is a mathematical illustration, the act contained no formula. The values of B and α were determined by Parliament. For the development of these values, see Table 7.
coalitions followed a conservative fiscal policy, increasing pensions only when sufficient reserves were available from social tax revenues.42 At the micro level, i.e., concerning the benefit calculation rules, the pension scheme could still be seen as a defined-benefit scheme. Although the level of benefits could also fluctuate downward in such a system in theory, in practice this did not happen, as social tax revenues increased due to economic growth and inflation. In the absence of any pre-determined rules for increasing pensions (e.g., indexation), two methods were used: increasing the base amount (the flat-rate component of the pension) and increasing the coefficients for years of service. Each increase required an ad hoc legislative amendment by Parliament modifying the pension benefit formula. Over the period of 1993–1999, the benefit formula was modified ten times (Table 8). Table 8. Changes in the pension formula, 1993–99 April 1993 Minimum wage (MW) EEK 300 National pension rate (NPR) – National pension (NP) 85% MW Old-age pension base amount: 85% MW the value of one service year: 15–19 years 0.5%MW 20–24 years 0.6%MW 25–29 years 0.7%MW 30–34 years 0.8%MW 35–39 years 0.9%MW 40 years and over 1.0%MW
September January 1995 1996 EEK 450 EEK 680 – EEK 410 EEK 410 110% NPR
April 1994 EEK 300 – 85% MW
July 1994 EEK 300 – EEK 300
85% MW
NP
NP
NPR
1.0%MW 1.1%MW 1.2%MW 1.5%MW 1.6%MW 1.7%MW
1.0%NP 1.1%NP 1.2%NP 1.5%NP 1.6%NP 1.7%NP
1.7%NP 1.8%NP 1.9%NP 2.2%NP 2.3%NP 2.5%NP
2.3%NPR 2.4%NPR 2.5%NPR 2.8%NPR 2.9%NPR 3.1%NPR
April January 1996 1997 Minimum wage (MW) EEK 680 EEK 680 National pension rate (NPR) EEK 410 EEK 410 National pension (NP) 120% NPR 135% NPR Old-age pension base amount: NPR NPR the value of one service year: 15–19 years 2.6%NPR 3.2%NPR 20–24 years 2.7%NPR 3.3%NPR 25–29 years 2.8%NPR 3.4%NPR 30–34 years 3.1%NPR 3.7%NPR 35–39 years 3.2%NPR 3.8%NPR 40 years and over 3.4%NPR 4.0%NPR
November March January 1997 1998 1999 EEK 680 EEK 1100 EEK 1250 EEK 410 EEK 410 EEK 410 145% NPR 160% NPR 195% NPR NPR
NPR
NPR
3.9%NPR 3.9%NPR 3.9%NPR 3.9%NPR 4.1%NPR 4.3%NPR
4.7%NPR 4.7%NPR 4.7%NPR 4.8%NPR 4.8%NPR 4.9%NPR
6.4%NPR 6.4%NPR 6.4%NPR 6.4%NPR 6.4%NPR 6.4%NPR
Source: adapted from Leppik and Männik 2002
42
Over the period of application of the State Allowance Act, six different governments were in power.
61
In principle, pensions could also be increased by raising the value of the national pension rate (NPR), rather than changing the coefficients. This method was not used, however, because the NPR also served as the flat-rate base amount. With respect to old-age pensions, the political aim was to emphasize the work-related length-of-service component. Therefore, starting in 1996 the basic amount was left unchanged and only the coefficients were increased. As a result, the relative importance of the flat-rate base amount declined, while the importance of work-related length-of-service component increased. For a person with 40 years of service, the share of the flat-rate component in the total pension declined from 71% in 1993 to 28% in 1999. In this way, the system became more advantageous for people with long periods of service. At the same time, progressivity was reduced in the length-of-service component; and by 1999, the values of service years were equalized. Chart 10 illustrates the impacts of changes of the pension formula on the values of pensions, indicating that differentiation of pensions increased. 2000
1800
1600
1400
1200 1993 1997
1000
1999 800
600
400
200
0 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
Source: calculations by author
Chart 10. Values of pensions (in EEK) depending on the length of service (in years) in April 1993, January 1997 and January 1999
4.4. OPERATION OF THE STATE PENSION SYSTEM IN 1990–2000 Administratively, starting in 1993 the state pension system was operated by the National Social Insurance Board (ENSIB), a government agency under the Ministry of Social Affairs. However, since at the Ministry there was no department dealing with social security matters, the ENSIB had to grapple with a broad range of issues 62
ranging from policy development to administration of the pension insurance budget to the supervision of regional pension offices.43 The latter were responsible for collection of the pension insurance part of social tax and the payment of benefits. During 1992–93, there was an increase in the total number of pensioners due to the 1991 broadening of pension coverage and abolition of the qualification period for invalidity pensions. The increase in the pensionable age, enacted with the 1993 State Living Allowances Act, stabilized the total number of pensioners around 375,000 in the second half of the 1990s (Chart 11). The number of old-age pensioners decreased by nearly 24,000 during 1994–2000, whereas the number of invalidity pensioners increased by nearly 17,000 over the same period. 450000 400000 350000 300000 250000 200000 150000 100000 50000 0 1990
1991
1992
Old age pension
1993
1994
Superannuated pension
1995
1996
Disability pension
1997
1998
Survivors' pension
1999
2000
National pension
Source: Ministry of Social Affairs Note: The so-called superannuated pension was inherited from the Soviet pension system (labeled also as service pension by some authors). It is a special early retirement pension for some occupational groups, e.g. artists, miners, pilots etc.
Chart 11. Total number of pensioners and distribution by types of pension (at the beginning of the year), 1990–2000
The rate of influx of new pensioners was reduced by the increase in the pensionable age (Chart 12).44 Here, changes in the timetable for increase of pensionable age and 43 44
The Social Security Department at the Ministry of Social Affairs was established only in 2000. Due to the timetable for increasing the pensionable age (on average by 4–6 months a year), in some years only half of the age cohort (with the same year of birth) could retire, whereas in some other years (e.g. in 1998) a full cohort reached pensionable age.
63
differences in cohort size caused some yearly fluctuations in the number of new pensioners. In general, however, the number of newly granted old-age pensions has decreased. An offsetting trend can be observed for invalidity pensions. As the incidence of disability increases with age, raising the pensionable age caused a simultaneous increase in the number of new invalidity pensions. 20000 18000 16000 14000 12000 10000 8000 6000 4000 2000 0 1990
1991 Old age pension
1992
1993
1994
Invalidity pension
1995
1996
Superannuated pension
1997
1998
1999
Survivor's pension
2000 Other
Source: Ministry of Social Affairs
Chart 12. The number of newly granted pensions by type, 1990–2000
However, in spite of the increase in the pensionable age, the system dependency ratio (ratio of pensioners to insured persons)45 increased from 50% in 1992 to 63.6% in 1999 (Chart 13). Although the number of pensioners declined beginning in 1994, the number of insured persons declined even more rapidly due to shrinking employment. Recovery in the labor market can be observed only beginning in 2000 (see Chart 7).
45
64
When calculating the system dependency ratio, the annual average number of pensioners (total number recipients of all types of state pension) is divided by the annual average number of insured persons (employees on whose behalf employers paid social tax was paid and selfemployed persons who themselves paid social tax to the pension insurance budget).
70.0
65.0
60.0
55.0
50.0
45.0
40.0
35.0
30.0 System dependency ratio (%)
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
50.0
55.1
57.2
57.7
58.7
60.4
61.3
63.6
61.9
60.7
62.5
Source: National Social Insurance Board, calculations by author
Chart 13. System dependency ratio (%), 1992–2000
The increasing system dependency ratio however did not translate immediately into financial problems of the pension system. Quite the contrary, in spite of the decline in the number of insured persons, the pension insurance budget maintained reserves (Table 9). Table 9. Revenues, expenditures and reserves of the state pension system (in million EEK), 1992–2000
Social tax revenues State budget allocations Other revenues Total revenues Total expenditures Cash reserves at year end Annual change in reserves
1992 731 731 694 67
1993 1514 49 1563 1440 190
1994 2170 30 162 2362 1970 582
1995 2917 214 3131 2908 769
1996 3844 26 73 3917 4067 618
1997 4637 19 198 4855 4728 744
1998 5339 150 38 5527 5306 965
1999 5520 176 15 5711 6460 216
2000 6297 254 3 6554 6504 20
+36
+123 +392 +186 -151 +127 +221 -749
-196
Source: National Social Insurance Board, calculations by author
This was because, as previously explained, the state pension system was operating on a macro-level defined-contribution principle, whereby expenditures were largely determined by available revenues from social tax. However, annual expenditures were also influenced by the timing of pension increases – pensions were increased by ad hoc political decisions in the absence of any pre-determined rules on the time or size of the increase. 65
During 1992–2000, pension expenditures exceeded revenues in two years, 1996 and 1999, reflecting political attempts to use pension increases to attract pensioner voters. On 1 January 1999, pensions were increased over 20% in anticipation of general elections in March 1999. Against the backdrop of negative economic growth and changes in the social tax collection procedures, pension expenditures exceeded social tax revenues by over 750 million EEK in 1999, nearly exhausting the reserves which had been accumulated in 1997–1998.46 However, as pensions were not increased in 2000 while the economy recovered, the balance between revenues and expenditures was restored by the end of 2000.
9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 Pensions in GDP (%)
1993
1994
1995
1996
1997
1998
1999
2000
6.3
6.3
6.8
7.3
6.9
6.8
7.9
7.0
Source: Ministry of Social Affairs, Estonian Statistical Office, calculations by author
Chart 14. State expenditures on pensions as a percentage of GDP, 1993–2000
Pension expenditure as a percentage of GDP increased from 1993 to 1996 even though the rate of social tax remained unchanged (Chart 14). The reasons for this varied. In 1992–94, the increase was driven mainly by the increase of real wages, which offset the impact of a declining number of wage-earners and resulted in the growth of the total wage bill in GDP from 34% in 1992 to 40% in 1994. Consequently, social tax revenues increased, allowing for an increase in pension expenditures and the establishment of some reserves. In 1995–97, the total wage bill as a share of GDP decreased. However, pension expenditures continued to increase in 1995 and 1996, mainly due to the use of reserves from earlier years and
46
66
As of 1 January 1999, collection of social tax was transferred from pension offices to the Tax Office. At the same time, dates of payment of social tax by employers were changed. Due to this transition, there was no deadline for payment of social tax in January 1999. As a result, in 1999 social tax was paid by employers only for 11 months.
improved collection of social tax.47 In 1997–98, the expenditure level stabilized around 7% of GDP. In 1999, this percentage jumped to 8.0% due to large increase of pensions in a period of negative economic growth. In 2000, pensions were not increased while at the same time the economy quickly recovered, resulting in the decline of pensions in GDP to the level of the mid 1990s. Old-age pensions account for nearly 85% of total pension expenditure. Accordingly, the share of old-age pensions in GDP has been around 5.5–6%. 1600 1400 1200 1000 800 600 400 200 0 01.04.1993
01.04.1994
01.07.1994
01.09.1995
Minimum wage Old age pension (40 years)
01.01.1996
01.04.1996
01.01.1997
National pension Disability pension (1st group)
01.11.1997
01.03.1998
01.01.1999
Old age pension (15 years)
Source: Ministry of Social Affairs, calculations by author
Chart 15. The amounts of pensions (EEK) under the State Living Allowances Act, 1993–1999
Chart 15 illustrates changes in pensions over the period when the State Allowances Act was in effect. The differentiation of pensions increased over the time span. Oldage pensions for persons with a service record of 40 years quadrupled in nominal terms, while the increase for those with a service record of 15 years tripled. It should be noted, however, that the average length of service was quite extended, exceeding 40 years. As Table 10 indicates, average old-age pensions increased in real value by over 50% during 1993–2000, as pension increases steadily exceeded the inflation rate.
47
The collection rate of social tax (revenues from social tax divided by social tax calculated from the reported wages) increased from 87% in 1994 to 97% in 1997 (see Leppik and Männik 2002 for further details).
67
Table 10. Average old-age pension versus the consumer price index, 1993–2000 Cumulative CPI growth (1993=100) Average old-age pension (EEK) Old-age pension in real terms (1993=100)
1993 100
1994 148
1995 191
1996 235
1997 261
1998 283
1999 292
2000 304
320
453
671
953
1110
1247
1545
1532
100
96
110
127
133
138
165
157
Source: Estonian Statistical Office, National Social Insurance Board, calculations by authors
The net replacement rate of the average old-age pension (that is average pension as a percentage of average net wage) increased from 40% in early 1990s to 45% in mid 1990s. The replacement rate exceeded 50% in 1999. However, this level was not sustainable due to the fixed revenue base of the pension system; and the replacement rate dropped to 46% in 2000 (Chart 16). 60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
1993
1994
1995
1996
1997
1998
1999
2000
Gross replacement rate
32.8%
29.1%
31.8%
36.0%
35.1%
34.0%
41.1%
36.5%
Net replacement rate
40.0%
36.8%
40.9%
45.7%
44.9%
43.9%
53.1%
46.3%
Source: National Social Insurance Board, calculations by author Note: Replacement rates are calculated on the basis of the average earnings upon which social tax was paid
Chart 16. Gross and net replacement rate of average old-age pension, 1993–2000
While most pensioners have incomes below the median, poverty rates among pensioner households appear lower than for other vulnerable groups like the unemployed, single-parent households, and large families (see Kutsar and Trumm 1999, Kuddo et al 2002, Tiit et al 2004). This is explained by the relatively flat structure of pensions, which results in most pensioners being in the second income 68
quintile – above the poverty level, but below the average income. Puur (2000) has shown that the relative incomes of oldest olds (75 years and over) somewhat improved during the transition period. 4.5. REASONS FOR MULTI-PILLAR REFORM AND EXPECTED RESULTS 4.5.1. Pension reform debate in the mid 1990s Starting in 1994, i.e. the year after adoption of the State Allowances Act, pension reform was promised by different political groups. There seemed to be an implicit political consensus that a pension reform is necessary but rather different views on what the reform should accomplish. In 1994–1997, every Minister of Social Affairs – there were 4 different ministers over the period – promised to present a new draft pension law and did so. Two more draft laws were presented by members of Parliament. In total, over the period of 4 years, six different pension bills were presented to Parliament. Two main themes dominated public debate during the mid 1990s – the low replacement rate provided by state pensions and questions about the fairness of the pension benefit formula. On the first issue, pensioner organizations demanded that the average old-age pension be increased to 50% of the average wage. During the 1995 Parliamentary election campaign, several political parties promised to implement a pension reform to achieve this if they were elected. As a result of the public discourse on this issue, the general public came to associate pension reform with an increase in benefits. Concerning the calculation of pensions, the dominant perception held that pensions should be calculated based on former earnings48. This was related to the issue of ‘rouble salaries’ – that is, the demand by pensioner organizations to recalculate pensions on the basis of salaries earned during the Soviet period. In essence, the issue of rouble salaries was a veiled demand to increase pensions, since it was believed that such a recalculation would boost pension amounts. The plan was however rejected by national conservative parties, who considered the idea as a throwback to the Soviet legacy. A further complication was due to competing proposals within the coalition, which took power after the 1995 elections.49 In addition to Government proposals, the Social Commission of Parliament put forward its own drafts. The pension debate broadened to include other policy matters and other actors, including the social partners. For example, with respect to the administrative 48 49
There were no public opinion polls on this matter, but the desirability of earnings-related pensions was a frequent leitmotif in many newspaper articles and political discussions. After the general elections in 1995, the government was formed by the centre-right Coalition Party and centre-left agrarian parties.
69
management of the pension system, the question was raised whether to establish an autonomous public legal body with a tripartite council or to continue with the government institution. The social partners advocated an autonomous institution, while the government was in favor of a governmental institution. However, there was broad consensus on pension financing, namely, that state pensions should continue to be financed from ear-marked social tax and that the budget must continue to be balanced. In this context, the claims of pensioner organizations to increase pensions translated directly into an increase in the rate of social tax, which was opposed by employers and rejected by successive ruling coalitions. Nevertheless, the possibility of dividing the burden of social tax between employers and employees was debated, primarily among the social partners. Employers’ organizations suggested sharing the social tax burden, while trade unions were opposed. These reform debates often took place without the benefit of background analysis. Policy alternatives were mostly debated in abstract terms, often referring to the experience of particular countries, without quantitative or qualitative analysis in the Estonian context. In addition to the reform of the state pension system, public discussions also focused on possibilities for establishing private pension schemes. By the mid 1990s, the banking and insurance sectors had already been privatized, yet there were no private pension products available. This was largely because it was unclear whether there would be a sufficient market for them. The discussions thus concentrated on a two-tier system with the state pay-as-you-go pillar and voluntary private pillar. To conclude, the main obstacles to reform in the early and mid 1990s were the existence of multiple competing ideas, a lack of political consensus on the aims of pension reform (including conflicts inside the ruling coalition), the short life span of governments, and the absence of background analysis on various proposals. Together these factors created a stalemate that extended to 1997. 4.5.2. The 1997 reform blueprint Prime Minister Mart Siimann, who headed the minority government that took power in March 1997, quickly took steps to overcome this stalemate. By a decree issued on 5 May 1997, the Government appointed a Social Security Reform Commission (SSRC) with the mandate to prepare an outline for pension reform. The expert commission was headed by Mr. Ardo Hansson, an economic advisor to the Prime Minister, and included 4 experts (Chairman of the Social Commission of Parliament, representatives of the National Social Insurance Board, the Health Insurance Fund, and the Ministry of Finance). In less than a month, this expert commission elaborated a reform proposal – a policy paper entitled Conceptual Framework for Pension Reform – approved by the Government on 3 June 1997.
70
The drafting of this paper marked a shift in the strategy and tactics of preparing the reform. Given the existence of numerous pension reform proposals, the SSRC suggested that before advancing to the stage of drafting legislation, a political agreement should be reached on the basic policy choices. The concept paper also presented an analysis of the existing problems of the pension system, in particular drawing attention to the worsening of the demographic situation and its long-term consequences for pension financing. By this, the SSRC broadened the focus of the pension reform debate from the concerns of current pensioners to the long-term sustainability of the system and intergenerational equity. The paper also set objectives for the new pension system. It stated that pension reform ought to balance the interests of various groups and create political and legal stability. As a social policy objective, it suggested that the compulsory pension system should secure at least the European minimum standard of social security, defined by the European Social Charter and the European Code of Social Security (see below). At the same time, it stressed that the reform should maintain the financial stability and sustainability of the pension system, as well as of public finances. The paper argued that in the long run, these objectives could essentially be achieved only by a multi-pillar pension system. In particular, it proposed to introduce a compulsory pre-funded second pillar with the following features: • I pillar: a state-managed compulsory pension scheme, operating on the pay-asyou-go principle, financed by the employer-paid social tax, and offering earnings-related benefits; • II pillar: a privately-managed, compulsory, and fully-funded pension scheme, financed by employees’ individual contributions; • III pillar: privately-managed voluntary pension schemes, in the form of pension funds or insurance policies offered by insurance companies. The first pillar was to be created by reforming the existing state pension scheme, while the second and the third pillars were to be introduced as new schemes. The first pillar was to cover the risks of old age, invalidity, and survivorship; the second pillar, only old age; and the voluntary third pillar, old age and invalidity. The SSRC took also a position on several of the issues under public debate. It showed that a 50 percent replacement rate could not be achieved under existing financial constraints (i.e., maintaining the pensionable age and the rate of social tax). It also showed that, due to the increase of the system dependency ratio, even the existing replacement rate could not be maintained in the long run without tightening eligibility criteria. It weighed in against recalculation of existing pensions on the basis of Rouble salaries. In terms of dividing the social tax burden between employer and employee, it took the position that the burden for the PAYG system should remain on the employer, whereas the pre-funded second pillar should be financed from individual contributions by the employee. In effect, this amounted to postponing the division of the social tax burden until introduction of the second pillar. 71
Concerning the timetable for reform, the Commission did not consider it realistic to undertake the reform as a single step. Rather, it suggested the reform of the first pillar as the first order of business, followed by introduction of the framework legislation for the voluntary third pillar. The introduction of the second pillar was scheduled only for 2001. The main considerations for this decision were: • obtaining first some experience with the voluntary, third pillar pension funds – both for the state as the regulator and supervisor as well as for the fund managers as administrators; • expected further development of the local financial market; and • expected further decline in the inflation rate, which in 1997 was still at 11%. The Commission set the following primary objectives for the first pillar reform: 1) Introducing stronger financial incentives for participation in the pension system and decreasing labor market distortions, especially the phenomena of ‘envelope salaries’, by relating each worker’s benefits more closely to the actual contributions made on behalf of him/her;50 2) Combating an expected increase in the system dependency ratio (beneficiaries to contributors) due to demographic aging by tightening eligibility criteria. In this way, a decline of the relative value (replacement rate) of pensions could be avoided; 3) Increasing financial transparency by shifting the financing of non-insurance pensions or pension supplements to general state revenues; 4) Guaranteeing compliance with the EU’s acquis communautaire by securing the equal treatment of men and women in all aspects of the pension system. The first objective entailed establishing individualized records of the amounts of social tax paid by employer on behalf of each employee, since previously, employers had paid social tax on the total payroll. The fourth objective was directly related to Estonia’s application for EU membership in November 1995. In spite of proposing to strengthen the earnings-benefit link in the first pillar, the concept paper held that an important feature of the first pillar was solidarity, both between and within generations. Intra-generational redistribution was to be achieved by minimum pension guarantees and a flat-rate base amount for pensions. The main declared aim of the second pillar was to increase individual responsibility by providing a benefit that was based entirely on each worker’s own contributions. In addition, the new compulsory pillar would, in the longer term diversify retirement income, since pensioners would rely on at least two different sources of income. The second pillar was characterized in the concept paper by the following principles: • compulsory participation of all persons covered under the first pillar; • benefits determined solely by the contributions paid, i.e. defined-contribution principle; 50
72
The term ‘envelope salaries’ refers to employers’ practice of keeping books and paying taxes only on a part of a worker’s total salary (often only on the minimum wage). The other portion is paid out (and accepted by the employee) ‘in an envelope,’ thus avoiding taxes.
• contributions paid by individuals (i.e. not by employers); • fully-funded financing principle; pension funds open to all workers and administrated by private asset management; and state supervision. The Commission’s recommendation to make the second pillar mandatory as a way of increasing individual responsibility was based on the underlying logic was that once people start to make individual contributions towards their private pensions, they will also revise their expectations towards the state pension system, thus easing the financial pressure on it. The Commission judged that universal mandatory coverage under the second pillar would facilitate such a paradigm shift more quickly. While providing a broad outline for reform, the Commission left several practical issues unaddressed. It did not take a firm position on the second pillar contribution rate, instead citing a need for further demographic and financial projections. It argued that the second pillar contribution rate should not be too small; otherwise, the new system would not be able to provide reasonable replacement rates. At the same time, it said, the contribution rate should not be too high either, as a high rate could be perceived more like a tax rather than a contribution. The Commission also did not take a firm position on the possible division of the social tax rate between the first and the second pillar. It did make it clear however, that introduction of the second pillar might be financed in part by an additional contribution. The aim of the third pillar was to provide instruments for additional savings for old age, to allow workers to maintain their earlier living standard. The SSRC supported an individual savings approach (similar to the second pillar), objecting to employerbased schemes on the grounds of potential negative effects on labor market flexibility. However, while it conceived of the second pillar pension as providing a single standardized product, it saw the third pillar as developing in two main forms: pension insurance policies offered by life insurance companies or participation in voluntary private pension funds. According to the concept paper, both forms were to be promoted by excluding contributions from taxation, whereas the benefits should be taxable. The concept paper was strongly backed by the Prime Minister and the leading Coalition Party. Approved by the Government on 3 June 1997, it served as a basis for drafting new pension legislation. Beyond the declared objectives and arguments of the Conceptual Framework for Pension Reform, the author’s personal conversations point to several other reasons why the three-pillar reform was suggested by the SSRC. First, the SSRC considered that without the second pillar, the average replacement rate would fall below the level of social adequacy in the long run, which would be neither socially nor politically sustainable. Second, the SSRC considered that the second pillar was 73
necessary to prevent an increase in the social tax for the purpose of increasing first pillar pensions. In this sense, it considered the second pillar as a sort of insurance against possible future increases of social tax, which in turn would have increased labor costs. These considerations were not imaginary problems of the future, but a reflection of the actual situation in 1997. In spite of following a macro-level defined-contribution principle with the rate of social tax being unchanged throughout the 1990s, there had been repeated claims to increase social tax in order to increase state pensions. The SSRC considered that with the ageing of population these pressures were only likely to increase. In this context, the second pillar was seen as a necessary tool with which to illustrate the cost of pensions for the broad public and, in this way, to add an element of realism and balance to the public debate. The SSRC also considered that when the option of increasing the social tax is evaluated, for longer-term sustainability it would be preferable to increase the total contribution rate in order to create the second pillar rather than expand the first. 4.5.3. Pension reform preparations in 1998–2002 The reform of the state pension scheme was initiated by a new Social Tax Act, adopted on 15 April 1998 and implemented from 1 January 1999. Most crucial changes were enacted with the State Pension Insurance Act, adopted on 26 June 1998, with gradual implementation foreseen during 1999-2000. The legal framework for the third pillar was also enacted in 1998 in the Pension Funds Act, adopted on 10 June 1998 and entering into force on 1 August 1998.51 It is noteworthy that these changes were legislated by a minority coalition. Although there was no formal agreement between the coalition and opposition, the opposition did not challenge the principles of the reform. After general elections in March 1999, the former opposition gained a majority in Parliament.52 A new three-party coalition was established comprising the liberal Reform Party, the national-conservative Pro Patria, and a "third way" socialdemocratic party of Moderates. In spite of political changes, the three-party coalition followed the broad reform outline accepted by the previous government. Moreover, in the coalition agreement signed by the three parties, the new coalition promised to continue to finalize pension reform, establishing the compulsory funded pension scheme. At the same time, the coalition committed itself not to increase the rate of the social tax.
51 52
74
The legislated changes are described in 4.6 and 4.8.4. At the end 1998, the ruling Coalition Party – hoping for re-election – took the controversial step of increasing state pensions over 20% beginning in January 1999 in a situation of economic recession. However, this did not buy political support for the party – on the contrary, the Coalition Party failed in the 1999 elections and was dismantled couple of years later.
The new coalition soon restructured the Social Security Reform Commission. Two cabinet members – the Minister of Social Affairs (Mr. Eiki Nestor) and the Minister of Finance (Mr. Siim Kallas) – joined the commission, with the Minister of Social Affairs taking the chair. The reconstituted commission – comprising a mix of politicians and experts – opened a debate on the key policy issues related to introduction of the second pillar: • who should be covered and whether the coverage should be voluntary or compulsory; • the second pillar contribution rate; • the management of the second pillar; • guarantees to fund participants; • tax treatment of second pillar contributions and benefits; • how to cover the transitional costs of establishing the second pillar; and • impacts of the second pillar on first pillar benefits. It is noteworthy that the three political parties in essence represented a broad political spectrum – liberals, conservatives, and social democrats. Simultaneously, the Government held trilateral consultations with employers and trade unions, which further broadened the debate. While none of these parties questioned the necessity of the reform, there were rather different views on how it should be implemented. The related political process was therefore complicated and finding compromises acceptable (or at least palatable) to different reform actors became necessary. Reform deliberations lasted nearly two years. Over this period, different alternative concerning division of contribution rates between the first and second pillars were debated, ranging from a radical 10+10 (i.e.,10% for the first pillar and 10% for the second pillar) to 16+6 (see also Oorn 2004). Another aspect of the debate concerned a possible division of social tax burden between the employer and employee, with a simultaneous increase in nominal wages. The population to be covered by the second pillar was also debated, with proposals ranging from compulsory participation for everyone under 50 years of age to voluntary participation for all. Finally, the coalition reached a compromise solution in January 2001. It included the formula ‘16+4+2’, indicating the new division of the contribution burden – 16% for the first pillar, 4% for the second, and an additional 2% contribution by employees for the latter.53 According to the compromise, the second pillar was to be voluntary for all workers regardless of age and broad participation was to be achieved by attractive switching rules. Under this proposal, the transitional financing costs of creating the second pillar were estimated to be about 0.3 – 0.8 % of GDP per year depending on how many from the current work force decided to join the second pillar.54 The methods 53 54
Full details of the plan are described in 4.6.2. Here transition costs refer to the “hole” in first pillar financing created by diverting a portion of social tax revenues to the second pillar.
75
suggested by the Social Security Reform Commission to cover these transition costs included the use of the stabilization reserve in the short run. In the longer-term, it suggested transfers from the state budget and possibly issuing of Government bonds (borrowing). It is noteworthy that the Government’s approach clearly softened in the course of debates. Debates started with a radical approach of 10+10 (with simultaneous division of contributions between employers and employees) and compulsory participation for everyone less than 50 years of age – this position was held by the Government in July 1999. Debates ended with a plan in which: (1) the first pillar contribution (social tax) was reduced only by 4 percentage points, (2) an additional individual contribution of 2% was introduced, (3) there was no division of the first pillar contribution (social tax) between employers and employees, and (4) participation was to be voluntary for all workers. Among the three coalition members, the second pillar reform was strongly backed by the social democratic Moderates, with Mr. Eiki Nestor, the Minister of Social Affairs and chairman of the Social Security Reform Commission, being the main spokesman for the reform. Mr. Siim Kallas, the Minister of Finance, and the leader of the liberal Reform Party, although originally a supporter of the radical 10+10 approach, later took a more cautious position towards the reform, due to the high transitional financing costs. To understand the development of the compromise, the positions of the different parties and dynamics of the public debate have to be examined. The Moderates regarded the second pillar reform as feasible only if the total contribution rate were increased, and thus supported the top-up element. They felt that future pensions for current workers could be increased only if some new resources were pulled into the system and therefore agreed to introduce an additional contribution for workers. This additional contribution was also acceptable to Pro Patria. However, the liberal Reform Party had concerns, as it was ideologically opposed to increasing taxes and contributions. Therefore, it insisted that the second pillar should be voluntary, leaving to each individual a free choice of whether to pay higher contributions. To move forward with the pension reform, Moderates and Pro Patria together agreed to this condition. Once the reform plan was made public, it received generally positive evaluations from the media – the main daily newspapers, TV, and radio commentators. However, the issue of voluntary participation was questioned by several commentators and the Government was criticized for taking a soft position in this respect. In the public debate that followed, the idea of compulsory participation was backed in particular by trade unions and from the other side by potential market players – financial institutions. In this situation, the Reform Party agreed to make participation in the second pillar compulsory for all new entrants to the labor market. The new compromise allowed all parties to save face. The Reform Party could still point out that for all current workers the choice was free, whereas the Moderates could point to the top-up element of the second pillar. 76
The draft Funded Pensions Act was presented to Parliament in April 2001. In Parliament, the draft was reviewed jointly by two committees – the Social Affairs Commission and the Finance Commission – a rare procedure for handling draft legislation. It was also defended before Parliament by two ministers – the Minister of Social Affairs and the Minister of Finance. The Estonian Parliament adopted the Funded Pensions Act on 12 September 2001. The Act was supported by 47 (from the total 101) Members of Parliament from the coalition of Moderates, Reform Party and Pro Patria. 26 Members of Parliament from the Center Party, agrarian Peoples’ Union, and predominantly Russianspeaking Estonian United People's Party were against it, and the others were absent. However, even the representatives of opposition parties indicated support for the second pillar while at the same time criticizing the Government bill for weak guarantees to fund participants and the potential negative impact of transition costs on the future increases of state pensions for current pensioners. Peculiarly enough, the strongest critics of the second pillar were not the opposition parties, but insurance companies who were concerned about the short-term weakening of their market position in selling voluntary pension products once the compulsory pension funds entered the market. Before the second pillar reform was implemented, the Government faced a crisis. Loss of trust in their coalition partner and mutual accusations on controversial privatization issues (namely, the failed privatizations of Estonian energy companies and problems associated with privatization of Estonian Railways) led Pro Patria and Moderates to step down from the Government at the end of 2001. In January 2002, the liberal Reform Party formed a new coalition with the Center Party, which had been its main ideological opponent in the 1999 election campaign. With the Center Party entering the coalition, some of the former critics of the reform were now in the Government. However, the new coalition agreement of the Center Party and Reform Party included a commitment to guarantee stability of the three-pillar pension system.55 As the Center Party enjoyed rather strong support from the pensioner population, the coalition also promised an additional pension increase on top of the regular indexation. The new Minister of Social Affairs from the Center Party (Ms. Siiri Oviir) personally participated in the awareness campaign, stating that she was concerned about her future and was therefore joining the second pillar. Concerning the broader context of ‘welfare culture’, public views on pension reform under planning are noteworthy. According to a public opinion survey 55
The reasons why the Center Party changed its earlier positions relate to political circumstances. The party had been in opposition for a long time in spite of rather strong electoral support. As the party was in conflict with almost all other major parties, it was generally not accepted as a coalition partner. The government crises in 2001 provided a window of opportunity for the Center Party to enter into government, but because of its weak negotiating power, it was forced to accept the proposals of the Reform Party.
77
conducted by Saar Poll (a private polling company) in May 2000, about 33% of respondents with household income per person in the bracket of 1000–3000 EEK (which at that time included about 70% of the population) supported the idea of compulsory funded pensions (Kirch 2000). Among persons with household income over 3000 EEK per persons (about 20% of the population), the idea was supported by 53% of respondents. These support rates were recorded at the time when the discussion on pension reform had been held only at a very general level, the government coalition had not yet reached an agreement on voluntary versus compulsory participation or contribution rate for the second pillar, nor had there been any public information activities. It therefore appears that the general idea of accumulating own’s pension besides relying on the public pension system had a significant number of supporters already at the early stage of reform debates. Another public opinion poll in June 2001 by ES Turu-uuringute AS (2002) studied the general attitudes of the population in respect of retirement and the pension system56. 82% of respondents considered the possibilities of copying with the current level of state pensions as unsatisfactory. Two-thirds (about 65%) did not believe that the situation of persons who rely only on the state pension could significantly improve in the future. Over half of respondents agreed with the statement that a person should also take steps his/herself to secure his/her retirement income. Over one-third of respondents expressed that they are prepared to save for their own retirement. Concerning personal expectations towards one’s own retirement the most common theme was a wish to avoid the misfortunate situation of current pensioners. 4.6. MAIN ELEMENTS OF THE MULTI-PILLAR REFORM The implementation of series of acts establishing the new pension system over the period 1998–2002 can be seen as a second wave of transformation. While the first wave separated the Estonian pension system from the Soviet one, the second wave entailed a move from a single-pillar to a multi-pillar system. The reform made changes in the state pension system but, even more significantly; it supplemented the state system with privately managed pre-funded pension schemes. 4.6.1. Changes in the state pension system Reform of the first pillar entailed parametric changes in both pension financing and benefits.
56
78
Note that the survey was conducted at the time when the draft law on funded pensions was discussed in Parliament.
Changes in the financing of state pension insurance The 1999 Social Tax Act brought about some important changes in the collection of social tax. Up to 1999, pension offices had collected the pension insurance part of social tax (20%), while sickness funds had collected the health insurance part (13%). As the collection was separate, some employers treated the two parts as separate taxes. Occasionally some employers transferred only one part, falling into arrears on the other. While the 1999 Social Tax Act maintained the rate of social tax unchanged, it made a significant change in the method of tax collection. This function was unified under the Tax Office. Starting 1 January 1999, employers were required to pay the total rate of social tax (33% of gross wage) to the Tax Office accounts. Under this new arrangement, the Tax Office then transfers 20 percent to the account of the pension insurance budget and 13 percent to the health insurance budget.57 The Tax Office performs the control functions previously performed by the pension and health insurance offices, as well as pursuing payments in arrears. While previously employers had calculated and paid social tax on the total payroll without providing any information on individual earnings, since 1999 employers have been required to provide data on the amount of social tax paid on behalf of each insured person. This information is reported on monthly basis to the Tax Office, which transmits the information to the Social Insurance Board. As mandated by the new State Pension Insurance Act, a State Pension Insurance Register was established as a structural unit of the Social Insurance Board, to record data on insured persons, including the amounts of social tax paid on their behalf. The new procedures thus required individual registration of social tax paid by employers on behalf of their employees. This was an important prerequisite for the introduction of the new individual contribution-related component of state pensions. Individual recording of social tax started on 1 January 1999, and the new benefit rules were to be applied beginning in 2000. The new legislation also prescribed integration of the formerly autonomous social insurance budgets with the general state budget. Starting in 1999, the state pension insurance budget was adopted annually by Parliament as a part of the general state budget. However, this was a technical, rather than a substantive, change since the earmarked nature of social tax had always been maintained, and revenues from social tax were held strictly separate from other state revenues. Also according to the State Pension Insurance Act, revenues from the pension insurance component of social tax cannot be used for any other purpose except payment of state pensions. The 1997 concept paper had proposed that beginning in 1999 all new pension rights would be based only on the payment of social tax. Earlier, certain periods were 57
The accounts of the pension insurance budget are administrated by the National Social Insurance Board. Health insurance revenues are managed separately by the Health Insurance Fund.
79
considered as covered employment even though no contributions were paid. These credited periods were financed by intra-generational redistribution. However, the cost of this redistribution was not transparent. The SSRC suggested that the state pay contributions on a deemed amount equal to the minimum wage on behalf of parents raising children, conscripts in compulsory military service, and unemployed persons. Essentially, the payment of contributions by the state would strengthen the financial base of the pension system. In the course of passing this proposal into law, Parliament modified the original SSRC plan. The list of categories of persons for whom the state pays social tax was extended, while the amount of social tax contributed by the state was reduced. The list of persons for whom the state pays social tax originally included the following categories of persons58: • parents with a child of up to 3 years of age who are on parental leave or who are receiving a child-care fee pursuant to the Family Benefits Act; • conscripts in compulsory military service; • persons with disabilities working in enterprises listed by the Minister of Social Affairs; • non-working spouses of diplomats working in a foreign representation;. However, the state paid social tax on behalf of these persons only on a deemed amount of EEK 700 per month, while the minimum wage in 1999 was EEK 1400. As a result, pension rights for periods outside the labor market are modest. 59 The 1998 State Pension Insurance Act also shifted a part of the burden for funding the state pension system to general taxation. The state budget became responsible for financing of non-contributory pensions (national pensions), certain politically motivated pension supplements that were unrelated to past earnings (e.g., to compensate for the periods of repression), and the administrative costs of the pension offices. Changes in the benefit side of state pension insurance The 1998 State Pension Insurance Act made a number of policy changes compared to the previous State Living Allowances Act, most importantly: 1) equalization of the pensionable age for men and women; 2) provision for early retirement with a reduced pension; 58 59
80
The list was later extended to cover also caregivers of disabled persons, non-working Chernobyl veterans etc. The amount of social tax paid by the state remained unchanged from 1999 to 2005, while the minimum wage increased to EEK 2690 by 2005. As a result, the pension insurance coefficients for the periods when the state pays social tax are only in the range of 0.1, i.e., about 10% of the pension right acquired by a person earning an average wage. This would have significant impact inter alia for parents with many children, staying away from the labour market, e.g. because of parental leave. However, the amount of social tax paid by state was doubled from 2006.
3) introducing a contribution-related element in the pension formula by linking the acquisition of new pension rights to social tax paid on behalf of the person; 4) replacement of disability pensions with so-called work incapacity pensions; 5) introduction of qualification periods for work-incapacity and survivors’ pensions; 6) calculation of old-age, work-incapacity and survivors’ pensions on similar principles. Most of these changes entered into force on 1 April 2000. However, the counting of pension insurance periods based on registered social tax payments started 1 January 1999. The pensionable age had first been increased in 1994 by the State Allowances Act. However, the earlier legislation maintained different target ages for men and women (65 and 60 respectively) to be reached by 2007. The State Pension Insurance Act stipulated equalization of the pensionable ages of men and women at 63 years. In essence, the target pensionable age of men was reduced from 65 to 63 (the pensionable age of men reached the target level in 2001), whereas the pensionable age of women was further increased, so that it will reach 63 in 2016. With demographic aging on the horizon, the increase of statutory pensionable age was seen as a key cost-containment measure. Parallel to increasing the pensionable age, the new act also provided an option for early retirement up to 3 years before the normal retirement age, with a reduction in benefits of 0.4 percent for each month of earlier retirement (i.e., 4.8 percent per year).60 The declared objective of this provision was to allow for greater flexibility in retirement decisions and in particular, to provide an alternative for persons who lost their jobs shortly before reaching pensionable age. To counterbalance the options of earlier retirement, a deferred old-age pension was introduced beginning 1 January 2002. Under this option, pensions were increased by 0.9 percent per month of postponed retirement (i.e., 10.8 percent per year). This is considerably more than an actuarially fair adjustment and provides a strong incentive to continue in work without drawing a pension.61 Continuation of work yields a double benefit, as the full career is taken into account for calculation of the pension and the additional working years increase the pension amount. The new formula to calculate old-age pensions consists of three additive elements: • a flat-rate base amount; • a length of service component applying to service periods through 31 December 1998; • a pension insurance component applying to contributions made after 1 January 1999.
60
61
In the draft law, the Government suggested to reduce the pension by 0.5% per month, which was the actuarially fair reduction. However, an amendment by trade union representatives reduced the coefficient to 0.4, making early retirement more attractive. However, in contrast to regular old-age pensions, these early retirement pensions are not paid to working persons. An actuarially fair increase would be about 0.6% per month.
81
The pension formula could be expressed as:62 P = B + s × V + ΣI × V, where – B = base amount; s = pensionable length of service of the applicant until 31.12.1998; ∑I = sum of pension insurance coefficients of the applicant, where I is annual coefficient accrued from 01.01.1999 onwards; V = the cash value of one year of pensionable service and the pension insurance coefficient 1.0. The flat-rate base amount constitutes the solidarity element in the state pension system and provides vertical redistribution from higher-income earners to lowerincome earners. The length of service component is also redistributive as it takes into account only the number of service years but not the former earnings. However, this component applies only to pre-reform ‘old service’ up to the end of 1998. From 1999 onwards, pension rights are acquired only based on social tax paid. The acquired rights are assessed through annual pension insurance coefficients, which indicate the social tax paid on behalf of the person as a proportion of the average amount of social tax paid for a worker in the given calendar year. Hence, the insurance coefficient 1.0 corresponds to the payment of social tax on the average contribution wage (i.e., the average earnings upon which social tax is paid). In fact, since the cash value of one year of pensionable service is the same as for the pension insurance coefficient 1.0, all service years up to 1998 are treated as if all persons earned an average wage. Thus, the amount of a pension depends on two individual variables – length of pensionable service before 1999 and the sum of pension insurance coefficients accumulated thereafter. Longer service before 1999 and higher amounts of social tax paid (i.e., higher legal wages) from 1999 onwards are the factors which contribute to a higher individual pension – there is no maximum pension. As pensions are essentially calculated on lifetime earnings, there are presumably no adverse incentives on labor market behavior. In essence, the new formula entails gradually increasing differentiation of state pensions as pension insurance coefficients vary more than the length of service. At the same time, a minimum pension guarantee was introduced, which limits the differentiation toward the lower end – the old-age pension for a person who has fulfilled the qualification period (which remained at 15 years of pensionable service) will not be less than the national pension.
62
82
Starting in April 2006, the base amount B is 1001.41 EEK and the value of the service year V is 48.51 EEK. Let us take, for example, a person with service of 40 years before 1999 and pension insurance coefficients acquired since 1999 summing to 4.0. According to the formula, the pension would be calculated as P = 1001.41 + 40 × 48.51 + 4.0 × 48.51 = 1001.41 + 1940.40 + 194.04 ≈ 3136 EEK.
For the majority of current pensioners who withdrew from work before 1999, the pension amount depends only on the flat-rate base amount and the number of service years. For persons who entered the labor market in 1999 or later, the state pension will consist of two parts: the flat-rate base amount and a contributionrelated insurance component. In fact, the three-part pension formula applies only to the 'transition generations', who have worked before as well as after 1999. The real amounts of pensions in payment depend on the values of B (the base amount) and V (the value of one service year and pension insurance coefficient 1.0). Until 2002, these values were determined annually by Parliament and the Government, respectively, within budget constraints. From 2002, pensions in payment as well as components determining the amounts of newly granted pensions (i.e., B and V) are indexed annually on 1 April.63 The index is an arithmetic average of the annual increase of the consumer price index and the increase of social tax revenues.64 In an important sense, the introduction of indexation changed the Estonian first pillar into a defined-benefit pension scheme, since benefits are now adjusted based on a formula in the law rather than on an ad hoc basis according to available revenues.65 This new index, which gives equal weight to price increases and increases in social tax revenues, determines changes over time in individual pensions and the state’s total pension obligations. Invalidity pensions were replaced by so-called work incapacity pensions. Whereas the former invalidity pensions could be paid regardless of age (from birth to death), 63
The introduction of pension indexation was suggested in the 1997 concept paper. In the course of drafting the law, the idea of indexation was rejected by the political coalition, maintaining the earlier ‘macro level defined-contribution’ approach. It was argued that with indexation, social tax revenues might be insufficient to finance pensions, depending on the development of factors determining the index and considering possible increases in the number of pensioners. In reality, the main driving force against indexation was politicians’ desire to leave their hands free to increase pensions before the next general elections, which were scheduled for March 1999.
64
In principle, social tax revenues and the price index could also decline. However, according to the law, pensions can not be reduced and in these cases indexation is not applied. When calculating the index, the whole pension insurance part of social tax – 20% − is taken into account. Therefore the partial loss of social tax revenues due to transfers to the second pillar does not affect the development of first pillar pensions.
65
It may be argued however that qualitatively the determination of pension rights on the basis of pension insurance coefficients calculated on the basis of social tax (i.e., the Estonian first pillar) is similar to determination of pension rights in the notional-defined-contribution (NDC) systems (e.g., the Polish and Latvian first pillar). Furthermore, the NDC systems also use indexation for revalorisation of the paid contributions as well as for pensions in payment. The difference in recording individual contributions lies only in the aspect that in the Estonian first pillar, contribution data are recorded in relative amounts (weighted against the average contribution) whereas in the Latvian and Polish first pillars contribution data is recorded in absolute amounts. Obviously, it may be argued that the latter method is more transparent. Moreover, the qualitative similarity of acquisition of pension rights does not make the Estonian first pillar a NDC scheme, as a crucial element of the latter is also the demographic adjustment factor (so-called G-factor).
83
work incapacity pensions are limited to persons of working age (from 16 to the pensionable age). Three previous invalidity groups were replaced by a new assessment of work incapacity in percentages. Furthermore, an age-related qualification period was established for work incapacity and survivors' pensions.66 Calculation of work-incapacity and survivors' pensions is generally based on the same formula as the old-age pension. There are, however, certain deviations. First, the higher of the following two amounts is determined: • The amount of an old-age pension calculated from the individual’s years of service and pension insurance coefficients (i.e., the amount of a standard old-age pension),67 or • The amount of an old-age pension for a person with 30 years of pensionable service.68 For a work incapacity pension, the calculation base, as derived above, is multiplied by the percent of person’s work incapacity. To create a floor below which workincapacity pensions cannot fall, it is further stipulated that the work-incapacity pension may not be less than the rate of the national pension (that is, the minimum old-age pension). Otherwise, individuals with a low degree of work incapacity (4060%) would end up with very low pensions. In fact, this calculation algorithm created a 2-level floor for the amounts of workincapacity pensions depending on the level of work-incapacity. In practice, most persons under 50 years of age – about two-thirds of all beneficiaries – receive the fixed rate, as their insurance record is relatively short. The standard old-age pension formula has relevance only for persons who are closer to the pensionable age and have insurance record longer than 30 years or whose sum of pension coefficients exceeds 30. The survivors' pension depends on the number of dependant family members. According to the new law, it amounts to: • 100 percent of the calculation base, in the case of 3 or more dependant family members; • 70 percent of the calculation base, in the case of 2 dependant family members; 66 67 68
84
In the latter case, the qualification period related to the insurance record of the breadwinner before death. In the case of a work-incapacity pension, this applies to the applicant, whereas in case of a survivors' pension it applies to the breadwinner. The choice of the second component of the floor, as described above – an old-age pension for a person with 30 years of pensionable service – bears an indirect relation to the requirements of the European Code of Social Security, signed by the Government in January 2000. The Code requires that the old-age pension for a standard beneficiary – a person with an insurance period of 30 years – shall correspond to at least 40% of the wage of an ordinary adult male labourer. The same standard – 40% of the wage of an ordinary adult male labourer – also applies for an invalidity pension for a person with a total loss of earnings capacity and for a survivor's pension for a widow with 2 children. In establishing this particular floor, the Government linked these requirements, so that the pension for a person with total work incapacity must at least equal the old-age pension for a person with 30 years of service.
• 40 percent of the calculation base, in the case of 1 dependant family member.69 Again, the floor established in the calculation base – equal to the old-age pension for a person with 30 years of pensionable service – provides the minimum survivor's pension. However, given that many breadwinners with minor children have relatively short insurance records, the majority of survivors' pensions are paid in fixed amounts. The State Pension Insurance Act also made significant changes in the role of the national pension. Previously, the national pension addressed only the risk of old-age for persons without a sufficient qualification period. Under the new rules, it is also granted to persons who lack the required insurance period for receiving a disability or survivors' pension as well. While national pensions for old-age are paid at a flat rate,70 for work-incapacity they are calculated to reflect the loss of capacity for work and the national pension rate. In contrast with work-incapacity pensions, there is no second floor established; and as a result, national pensions for persons with a low degree of work incapacity (40–50%) are modest, falling even below the social assistance benefit level. For survivors, the national pension also depends on the number of dependant family members and is calculated using the same percentages as in the survivors' pension, but again applying these to the national pension rate. Thus, in essence, the first pillar comprises of 2 separate tiers: • residence-based national pensions; and • employment-based old-age, work incapacity, and survivors' pensions.
69
70
These percentages were changed on 1 April 2004, when the rate of pension for a single survivor was increased to 50% and for 2 survivors, to 80%. As explained previously, the same rate – national pension rate – also serves as the minimum pension guarantee for old-age pensions.
85
Table 11. Comparison of the old and new rules for the first pillar Collection of pension insurance part of social tax Payment and declaration of social tax by employers Target pensionable age
Until 1999 Pension offices
Starting with 1999/2000 Tax Office
Paid on total payroll, no individual registration of wage data 65 for men, 60 for women to be reached by 2007
Amounts of social tax indicated separately for each employee 63 for both genders, to be reached by 2001 for men and by 2016 for women On the basis of social tax paid Flat-rate base, variation on the basis of social tax paid over the full career Indexation (from 2002) Defined-benefit principle: initial benefits determined by the amounts of social tax paid, pension adjustments and total expenditures determined by the index
Acquisition of pension rights Old-age pension formula
On the basis of years of service Flat-rate base, variation on the basis of length-of-service
Increase of pensions The key principle of the first pillar
Ad hoc political decisions Macro-level definedcontribution: the rate of social tax was fixed (at 20% of gross wages), the level of pensions depended on the resulting revenues
Source: own compilation
The state pension scheme also provides old-age pensions on favorable conditions and superannuated pensions. The former are paid to workers in occupations that are considered hard or hazardous (e.g., workers in chemical, metal, glass, pulp industry, mining, etc.) They may retire 5 or 10 years before the normal retirement age if they have fulfilled certain requirements – i.e., from 15 to 25 years of pensionable service of which at least half was in the given profession. In addition, parents of disabled children, parents who have raised 3 or more children, and some other categories may retire before the normal pensionable age. Superannuated pensions are in essence early retirement pensions for certain professional groups, like pilots, mariners, miners and artists, whose professional abilities are assumed to have declined before the normal pensionable age. Most of these privileged rules were inherited from the Soviet pension system.71
71
86
The Soviet pension law included a lower pensionable age (by 5 years) for mothers with 5 or more children. The 1993 State Allowance Act extended this right to one of the parents at the choice of the family. The 1998 State Pension Insurance Act broadened early retirement to parents of 3 or more children – that is, retirement 1 year before general pensionable age for parents of 3 children, 3 years for parents of 4 children, and 5 years for parents of 5 or more children.
In the 1997 reform proposal, the SSRC had suggested limiting these privileged rules. Its rationale was that the actual working conditions in many privileged professions had improved. It also noted that the state’s acceptance of certain professions as unhealthy provided no incentive for employers to improve working conditions. Concerning superannuated pensions, the SSRC also questioned stateprovided early retirement for pilots and mariners, since these sectors had been privatized and favorable rules on retirement could be regarded as an indirect subsidy for some private enterprises. The SSRC suggested applying universal rules in the state pension system, with any special early retirement provisions to be financed separately by employers through voluntary insurance. These proposals however have so far not been implemented. They were strongly opposed by trade unions, which demanded the introduction of a separate scheme on work accident and occupational disease insurance as a precondition to any changes.72 This in turn would necessitate the introduction of an additional contribution, i.e., an increase of labor costs, which was opposed by both employers and the Government. Until 2002, state pensions constituted non-taxable income. Since then, pensions have been treated as a taxable income, but a higher non-taxable allowance applies.73 4.6.2. Implementation of the second pillar The second pillar became operational on 1 July 2002, when collection of contributions for the new individual savings schemes started. The main second pillar rules were enacted by the Funded Pensions Act (adopted 12 September 2001) and subsequent regulations of the Government and the Minister of Finance. However, in addition to the Funded Pensions Act, the Guarantee Fund Act (adopted 20 February 2002), the Investment Funds Act, and amendments to the Estonian Central Register of Securities Act (adopted 12 September 2001) had relevance for implementation of the second pillar. Moreover, as the first pillar reform was legislated before the introduction of the second pillar, further amendments to the first pillar rules were introduced to deal with the consequences of its introduction.
72
73
There is no separate scheme for these risks in the Estonian social security system. The risks of work accidents and occupational diseases are covered respectively by the general health insurance and pension insurance schemes, and on top, employers’ civil liability applies. The general non-taxable income in 2006 is 24,000 EEK per year. For pension income, the non-taxable allowance was further increased by 36,000 EEK. As the overwhelming majority of state pensions remain below 5000 EEK per month, they are not taxed.
87
Coverage The second pillar addresses only the risk of old-age and does not provide pensions for the risks of invalidity and survivorship. Participation in second pillar is compulsory for new entrants to the labor force beginning in 2002, while voluntary for all current workers, i.e., for those already in the labor market before 2002.74 In contrast to Poland and Latvia, Estonia opted not to apply explicit age restrictions for limiting access to second pillar for older cohorts. Instead, a qualification period of 5 years was used to prevent early access to second pillar benefits. However, policy makers believed that the qualification period might also discourage older persons from joining the new system. All persons born before 1 January 1983 were granted the option to join the second pillar voluntarily. Operational rules for exercising this right nevertheless contain some age discrimination – namely, older cohorts were provided shorter deadlines (Table 12). Table 12. Deadlines for submitting applications for cohorts, whose participation in the second pillar is voluntary75 Deadline 31 October 2002 31 October 2003 31 October 2004 31 October 2005 31 October 2006 31 October 2007 31 October 2008 31 October 2009 31 October 2010
Cohorts Persons born 1942–1956 Persons born 1957–1961 Persons born 1962–1964 Persons born 1965–1967 Persons born 1968–1970 Persons born 1971–1973 Persons born 1974–1976 Persons born 1977–1979 Persons born 1980–1982
Source: Funded Pensions Act
The underlying thinking was that age-specific deadlines might implicitly discourage older persons from joining since they had to make a decision based on limited information.
74 75
88
In operational terms, participation in the second pillar is compulsory for young persons aged 18 entering the labor market in 2002 or later – i.e. persons born after 1 January 1983. Deadlines for younger cohorts were shortened in 2004. According to the previous timetable, persons born in 1982 could join the second pillar until 2024. The open season was shortened since, in 2004, over half of the eligible persons had already joined the second pillar. Another factor was the Government’s interest in getting a clear picture of transitional financing costs by determining with certainty who is participating in the second pillar and who is staying out.
Overview of legal rules on participation in the second pillar Participation in the second pillar entails two basic choices: • first, a decision to join the system, accepting an obligation to pay an additional contribution of 2% of the wage; • second, the choice of a pension fund, to which contributions are directed for investment. The first decision is irreversible: there is no option of ceasing payment of contributions, neither switching back to the first pillar. After joining the system, the payment of contributions becomes a legal obligation. The main incentive to join is the fact that individual contributions are supplemented by the state with 4% of gross wage, re-directed from the pension insurance part of social tax paid by employers. Put in the other words, joining the second pillar and paying an individual contribution of 2% of gross wage provides a right to re-direct 4 percentage points of social tax to an individual pension account in a private pension fund. The second decision, choice of a fund, may be altered from time to time by the individual. The new division of the pension insurance part of social tax is best captured by the formula '16+4+2'. Participants in the second pillar pay an individual contribution of 2% of their gross wage, which is supplemented by the state with 4 percentage points of the individual gross wage posted on the social tax account of the employer.76 In total, 6 percent of the gross wage is accumulated in individual account in the second pillar pension fund, whereas 16 percentage points continue to finance current state pensions and to serve as the basis for computing each worker’s contribution-related state pension benefits.77 Social tax 33 % Health insurance 13 %
Pension insurance 20 % I pillar 16 %
II pillar 4% +2% individual contribution
Chart 17. Distribution of social tax after second pillar pension reform
76 77
The individual contribution of 2% reduces taxable income (i.e., contributions are not taxable); therefore, the net wage decreases less than 2%. It should be noted here that the first pillar contribution 16% covers three risks – old age, disability and survivors – whereas the second pillar contribution covers only the risk of old age.
89
Consequently, participation in the second pillar diminishes pension rights in the first pillar. Since only 16 percentage points of social tax (instead of the previous 20%) goes to the first pillar, the annual pension insurance coefficient – determining the size of the contribution-related component of the first pillar pension – is proportionally reduced (by 4/20, i.e., 1/5). However, this affects only post-reform periods, leaving previously acquired pension rights unaffected. Two important features of the first pillar benefit structure are unaffected by an individual’s decision to join the second pillar. First, the base amount of the state pension (which in 2002, when the reform was introduced, accounted for about 25% of the average old-age pension) is not reduced. Second, the minimum pension guarantee in the first pillar – that is, the provision stipulating that the old-age pension cannot fall below the national pension rate – is maintained, regardless of an individual’s decision to join the second pillar and consequent lower contributions to the first pillar. The first pillar rules thus contain several additional incentives to join the second pillar. Persons who decide not to participate in the second pillar will acquire rights only from the reformed first pillar. The pension insurance component of the social tax remains at 20%, and their pension will develop only from the first pillar. As described above, the rules of the second pillar provide an opportunity for a partial opt-out from the state system, on the condition that the individual pays additional contributions. In other words, the second pillar was established using both a ‘carve-out’ and a ‘top-up’ method. A portion of the pension part of the social tax was redirected to the second pillar, to be supplemented by an additional contribution of the employee. The reform is neutral with respect to the overall social tax rate of employers, but affects the distribution of social tax revenues, redirecting a portion from the first to the second pillar. Originally, second pillar contributions could be paid only on wage income. Although there were no formal obstacles to joining the second pillar for self-employed individuals (i.e., they could make an application), contributions could not be paid on income from self-employment78. This was changed in 2004, allowing accumulation of a second pillar pension for self-employed persons as well as employees. Periods outside the labor market as a rule do not accrue second pillar pensions, as contributions are paid only on wages (and beginning in 2005 on income from selfemployment). There was however, one exception introduced in 2004 – the state pays 1% contribution on the parental benefit.79
78
79
90
The reasons why second pillar contributions were not permitted on income from selfemployment in the first 2 years of the reform were administrative and technical rather than political. The barriers related to the different taxation period (which for self-employed persons is a calendar year, as opposed to a calendar month for employers) and to the procedures for making quarterly advance payments of social tax by self-employed persons. The parental benefit is an earnings-related benefit introduced in 2004. It provides 100% of previous earnings for one of the parents of a newborn child during its first 11 months. The 1% contribution paid by the state is still only 1/6th of what is accrued on work income.
Types of pension funds and rules for joining them By their nature, pension funds in Estonia are open contractual investment funds. The fund is not a separate legal entity, but a pool of assets owned mutually by fund participants. The assets of a pension fund are formed from contributions and investment returns. The rights of fund participants to their share of the assets of a fund are represented by units of the fund. Pension funds are managed by asset management companies. In 2002, 6 pension fund managers entered the market. Three of these were affiliated with the 3 largest banks in Estonia (Hansa Asset Management, Ühispank Asset Management, and Sampo Asset Management), 2 linked to insurance companies (Ergo Asset Management, Seesam Asset Management), and 1 linked to an investment bank (LHV Asset Management). The pension funds themselves can be classified into 3 different categories according to investment strategy: • lower-risk (or conservative) funds, which may invest only in fixed-interest instruments (bonds, money market instruments, and bank deposits); • medium-risk (or balanced) funds, which may invest up to 25 % of assets in equities; • higher-risk (or aggressive) funds, which may invest the maximum allowed amount – 50 % − of assets in equities. The Funded Pensions Act requires each fund manager to establish a low-risk fund. In addition, fund managers may set up additional second-pillar funds, but the investment strategy of the other fund must be qualitatively different. In practice, each of the fund managers has set up a higher-risk fund, and 3 fund managers offer a medium-risk fund. Accordingly, there are a total 15 different second-pillar pension funds: 6 low-risk, 3 medium-risk, and 6 higher-risk funds. In any given calendar year, contributions can be made to only one fund. However, at the beginning of the following year contributions can be re-directed to a new fund, leaving previously obtained units in the old fund. Thus over a worker’s full career it is possible to accumulate units in many different funds.80 The Estonian second pillar design is in this respect different from many other Central and Eastern European states (e.g. Hungary and Poland), which have followed a ‘one person-one fund’ principle, whereby at any given time all units of the persons are in a single pension fund – if funds are switched, all existing previously acquired units have to be converted. The participation rules in the Estonian case are therefore considerably more flexible. It is also possible in Estonian case to convert the units of one pension fund to units of another one. However, in this case, a minimum threshold of 500 units applies.81
80 81
Shifting of contributions to a new fund became possible as of 1 January 2004. Given the average wage of second pillar participants at around EEK 7000 a month (in 2004), it takes the average wage earner about 1 year to acquire 500 units.
91
Switching funds is possible only once a calendar year, always from 1 January.82 These rules constrain members’ ability to react to short-term market changes. Administration of the second pillar The 2% employee contribution to the second pillar is withheld by the employer and transferred together with the social tax to the Tax Office. The Tax Office supplements the 2% contribution with 4 percentage points from the social tax and transfers the total contribution (6%) to the bank account of the Estonian Central Depository for Securities (ECDS).83 The ECDS calculates the number of pension fund units corresponding to the contribution received, issues the respective number of units, records the relevant information, and transfers the total contribution to the custodian bank of the fund management company. Pension fund managers then decide how to invest the assets. In this way, the functions of account administration and investment management are fully segregated.
82
83
92
Switching funds became an option on 1 January 2005, i.e., it was not possible to change funds in the first years of the reform. To do so, the Tax Office has to identify all participants of the second pillar. This information is provided to the Tax Office by employers, but the Tax Office may also check the information from the Central Register of Securities. In fact, there are different contribution rates for different employees of the same employer, depending on whether the person has joined the second pillar or not.
Employee Contribution 2%
Employer Social tax 33%
Tax Office Controls and registers the first and second pillar contribution data Matching 2%+4%
Central Depository for Securities Opens pension accounts Registers second pillar contributions Issues pension units
Custodian bank Holds accounts of pension fund management companies
Fund managers Makes investment decisions
Chart 18. Administration of the second pillar
The Estonian Central Depository for Securities – the registrar of the Estonian Central Register of Securities – plays a key role in administration of the second pillar. The ECDS is a private company, a part of the Tallinn Stock Exchange group, the latter being the infrastructural centre of the securities market in Estonia. The ECDS opens a pension account and sends an annual account statement to all second pillar participants, either by regular mail or electronically by e-mail. The person can also check the balance in his/her account at any time through Internet bank or the special Internet page of the Register (www.pensionikeskus.ee). When a person switches funds and decides to transfer his/her contributions to the new fund, the ECDS redeems the units of the old fund and issues units of the new fund. In the future, the ECDS will also have to process benefit claims for second pillar pensions, redeeming the pension fund units and transferring the corresponding sum to a life-insurance company. In the case of a programmed withdrawal (explained later in this section), the ECDS will redeem the units and administer the payment of the pension directly. 93
The law authorizes pension fund management companies and their custodian banks to obtain from the ECDS the names and personal identification codes (ID-codes) of their account owners, as well as statistical information on issuing and redemption of pension units for these individuals. The ID-code contains information on the gender and date of birth of the person. In these ways, the fund managers may know the identity of their members. In addition, the law makes it possible for other persons to know whether a particular person has joined the second pillar or not. Inquiries may be submitted to the Register through its web site based on the personal ID-code. The Register will disclose whether a pension account has been opened for that person and the year, in which the obligation to make contributions commenced or will commence, but no information on which fund the person has chosen. The decision to entrust the management of second pillar pension accounts to the ECDS rather than to the social security body (National Social Insurance Board) related to the fact that second pillar pension accounts are a particular type of security account. Before the second pillar was introduced, the ECDS already had the basic infrastructure for managing securities accounts as well as institutional links with the other institutions involved (custodian banks, fund management companies, life-insurance companies). It was thus considered more feasible to broaden the functions of the ECDS than to add some new functions to the Social Insurance Board. Investment regulations Certain qualitative requirements and quantitative limits are applied to pension fund assets. They may be invested only in recognized securities (shares, bonds, units of investment funds, money market instruments, and derivatives), bank deposits and real estate. To ensure risk diversification, a fund’s share in one entity (instruments issued by the same issuer) may not exceed 5% of its assets, except for bonds issued by a Member State of the European Economic Area (EEA) – in this case, the limit is 35%. Among the CEECs, Estonia has one of the most liberal regulations concerning foreign investments (Financial Advisory Ltd. 2003). In contrast to countries like Poland or Hungary which have placed a general limit on pension fund assets available for foreign investments (5% in Poland and 30% in Hungary), there is no general limit in Estonian legislation. Instead, there is a currency matching limit – investments denominated in currencies of third countries may not exceed 30% of funds assets. This limit however does not apply to instruments denoted in Euros, as the Estonian Kroon is pegged to Euro. For geographical risk diversification, the share of pension fund assets invested in instruments of issuers of any single country may generally not exceed 30%, unless a higher share is accepted in the operating conditions of the pension fund.84 There is also a limit on investments outside the EEA and OECD – for companies registered 84
94
Operating conditions of the pension fund are subject to the approval of the Financial Inspectorate.
in such countries, 30% of pension fund assets may be invested in their securities, while for instruments traded only in such countries, 20% of pension fund assets may be invested in those securities. Similar to the situation in Poland and Hungary, Estonia has a limit on equity risk exposure. Pension funds are allowed to invest up to 50% of their assets in shares. The limit for equities also includes the units of mutual funds investing in equities. Investments in money market instruments are allowed up to 35% of pension fund assets. Direct investments in real estate are allowed up to 10% of pension fund assets, but not more than 2% in a single piece of real estate. The aim of applicable investment regulations was to secure an effective diversification of the investment risk, both geographically and by types of instruments, since high risk diversification is presumably in the best interest of participants. On the one hand, the EU free movement of capital principles were applied. At the same time, considering the relatively small size of the local stock market, substantial restrictions on foreign investments would have entailed high local risk concentration. Operating rules for pension fund managers A number of requirements apply to pension fund management companies that seek to establish second pillar funds. The company must first apply for a special license from the Financial Supervision Authority (FSA). To be eligible, the minimum share capital of the fund management company must be least 3 million EUR. Once licensed, the fund management company must present the conditions of the pension fund for registration. If the fund manager intends to run more than one second pillar pension fund, each pension fund must have a different investment strategy, distinguished by the share of pension fund assets invested in equities. For the investment strategy to be considered as qualitatively different, the share of assets invested in equities must differ by at least 25 percentage points. As the maximum allowed share of equity investments is 50%, in essence, one fund management company may run up to 3 second pillar pension funds. Besides the second pillar pension funds, the fund management company is allowed to run other types of investment funds (e.g., voluntary pension funds, open or closed investment funds, individual portfolios). A pension fund management company is obliged to participate itself in the pension funds it manages; at the same time it may not participate in other pension funds. The fund manager is thus obliged to purchase the units of pension funds it manages and keep the share of it units within established limits – within three years of establishing the fund, the fund manager must have at least 2% of the pension fund units. When the number of units of the pension fund exceeds 100 million, compulsory participation is 2% of the first 100 million units (i.e., 2 million units) and 1% of the units exceeding 100 million.
95
As noted above, the units of a pension fund represent the rights of fund participants to a share of pension fund assets. The nominal value of pension fund units is 10 EEK. The market value of units is captured in the notion of net asset value (NAV). The net asset value of units is calculated as the total market value of a fund divided by its total number of units.85 The operational costs of the fund management company are covered by fees. The rate of the management fee is determined as a proportion of the market value of the assets of the pension fund and is set on a yearly basis. The maximum rate of the management fee is established by the Minister of Finance and differs for fixed income funds, versus those that also invest in equities. For fixed income funds, the maximum rate for management fee is 1.5% of assets under management; and for funds investing also in equities, the maximum management fee is 2.0%. The gross management fee includes a registrar charge, custodian charge, guarantee fund charge, supervision charge, and net management fee.86 The costs of organizing the issue and redemption of the units of the pension fund, managing the pension accounts, and performing other duties prescribed by the Act are covered by the registrar's charge, payable by the pension fund management company to the registrar (i.e. the ECSD). Fund participants pay separate fees for issuing and redemption of pension units. In both cases, maximum fees are established by legislation. The maximum subscription fee for pension funds units (applicable until 1 January 2007) is 3% of the net asset value of the unit. The maximum redemption fee is 1% of the net asset value of the unit. Mechanisms to protect the interests of fund participants In contrast with the earlier reformers of Hungary and Poland, the Estonian second pillar has relatively modest explicit guarantees. There are no guarantees on absolute or relative rate of return, leaving the investment risk on participants. The fund management company is even prohibited from making any direct guarantees on rate of return. However, members are protected against any breaches (e.g., violation of investment rules) by the fund managers. For protection against any possible damages caused by the fund managers, the pension fund management company must pay contributions to a Guarantee Fund. In taking this approach, the SSRC’s primary consideration was avoiding the potentially counterproductive effects of guaranteeing the rate of return – i.e., adverse incentives for fund managers and participants. It was considered more appropriate to build into the system an appropriate incentive structure. One example is compulsory participation of the fund manager in the pension fund. In this way, the fund manager is motivated to make reasonable investment decisions. The lack of guarantees must also be evaluated in light of the option provided for participants to choose among pension funds with different levels of risk. 85
86
96
When calculating the market value of a fund, the management fee is deducted from the value of its assets. For the structure of the management fee, see Põld 2002.
Although in principle the participants of a pension fund can ‘vote with their feet’, i.e., leave pension funds with poor investment returns, this option is limited since switching can occur only at the beginning of the next calendar year. In this context, a crucial role in protecting the interests of pension fund participants has been assigned to the Financial Supervision Authority, which supervises pension funds. Established in 2002, the FSA is an independent legal authority, which consolidates the formerly separate inspections of banking, insurance, and the securities market. The FSA issues activity licenses for pension fund management companies and approves conditions of pension funds and any amendments thereto. The FSA checks all information provided by fund management companies when applying for licenses or registration of conditions. The FSA also controls the application of funds’ investment policies and their adherence to investment limits prescribed by legislation. To perform these functions, the FSA may demand any necessary information, statistical data, or explanations from fund managers, custodian banks, or any other relevant party. In the case of violations, the FSA may impose penalties. Any mergers of fund management companies are subject to the approval of the FSA. All fund management companies have to submit semiannually activity reports to the FSA. Second pillar benefits As the second pillar is a defined-contribution system, future pensions depend on the value of assets accumulated by the individual over the whole career – the total value of contributions and the rate of return from their investment, minus administrative expenses. Thus, the net rate of return is inter alia also influenced by the fees charged by the fund management company. For a person to be eligible for a second pillar pension, he/she must have: • reached pensionable age (the same as in the first pillar); • been granted a first pillar pension; • participated in the second pillar scheme for at least 5 years. In fact, the latter rule has relevance only for those who joined the system voluntarily. Generally, second pillar benefits must be paid out in the form of annuities, i.e., the purchase of annuities is mandatory. However, there are exceptions if the calculated annuity is rather small, i.e., less than ¼ of the rate of national pension (or ¼ of the minimum first pillar benefit) per month. In this case, the person has a right to a programmed withdrawal of assets from the pension fund. In addition, when the calculated annuity exceeds 3 times the rate of national pension, the person has a right to a supplementary programmed withdrawal of the funds that remain after purchasing the compulsory annuity. In the other words, there is a minimum and a ceiling on the mandatory annuity. If the calculated annuity falls below the minimum (currently corresponding to about 16 EUR) or exceeds the ceiling (currently corresponding to about 190 EUR), the person may opt for a programmed withdrawal, which in the former case replaces the annuity and in the latter case supplements it. 97
The benefit rules therefore regard life annuity as the primary option, providing insurance against the uncertainty of remaining lifetime. Programmed withdrawals are limited to very small accumulations, but give also some flexibility for high income persons. To claim an annuity, a person must enter into a contract with a specially-licensed life-insurance company and thereafter submit an application to the Central Register of Securities for redemption of pension units. For a programmed withdrawal, the person must submit an application to the Register. Whereas during the accumulation period pension funds are not allowed to guarantee any rate of return, insurance companies may guarantee interest rates on annuities. However, the interest rate used for calculation of an annuity (i.e., guaranteed interest) may not exceed 3%. The insurance company must apply unisex mortality tables for men and women in calculating annuities. Considering the rather high differences between life-expectancy of men and women in Estonia (see Table 5), the latter feature implies considerable redistribution from men to women. In a way, this feature would somewhat balance the impacts of wage differences between men and women. If the insured person dies before reaching pensionable age, the units of the fund are inherited. However, if the person dies thereafter, only undistributed assets from programmed withdrawals are inherited. Any reserves from life-long annuities are not inherited unless a guarantee period is agreed at the time of purchasing the annuity. The payment of benefits from the second pillar will start from 2009. Second pillar pensions constitute taxable income in principle. However, the higher non-taxable allowance for pensions also applies for second pillar pensions. Only the part of the sum of first and second pillar pensions, which exceeds the higher non-taxable threshold, is taxed.87 Therefore, at least for the initial years (and for persons with lower earnings in later years), there would be no effective tax on second pillar pensions for many persons, creating an additional incentive to join the system.
87
98
In 2004, the general non-taxable allowance was 16,800 EEK per year. For pension income, the non-taxable allowance is further increased by 36,000 EEK. According to amendments to the Income Tax Act adopted in 2003, the general non-taxable allowance is increased to 24,000 EEK a year by 2006. Therefore, pension income would be taxed only if the sum of the first and second pillar pension exceeds 5,000 EEK and the tax would apply only to the part exceeding 5,000 EEK. Considering the average old-age pension (currently 2,200 EEK) and projections on the development of first and second pillar pensions, it seems very likely that at least in the initial years (when the second pillar benefits are likely to be rather small due to short collection period) the effective tax on second pillar pensions will be very small, if any.
4.7. INTERNATIONAL INFLUENCES The Estonian pension reform package was clearly not developed in isolation from outside influences. However, direct foreign inputs (for example, drafting of legislation by foreign experts) were very limited. Instead, various international influences were internalized by different national actors. Lindeman (2004:12) characterized the situation as follows: “Though key policy advisors in Estonia closely monitored pension reform innovations in the region and debates in the literature, the eventual three-pillar reform was a home grown product /…/”. Indeed, there are clear indications that developments in the other Central and Eastern European countries were monitored. For example, the 1997 Conceptual framework of pension reform included an annex describing pension reform developments in Czech Republic, Hungary, Latvia, Slovenia and Poland. Although the Estonian reform does not copy any other, the experiences of earlier reformers were taken into account. Besides this regional learning, a number of international agencies had some influence on policy developments in Estonia. World Bank Readers of the 1997 Conceptual framework of pension reform would obviously notice some clear influences from the 1994 World Bank’s Averting the Old Age Crises. This includes not only the Bank’s general three-pillar framework but also argumentation on why the reform was needed. However, a closer look also reveals a number of differences from the World Bank approach. For example, for the first pillar, the Social Security Reform Commission suggested a gradual replacement of the earlier length-of-service pensions with a German-type point system, rather than shifting towards the means-tested or flat-rate pensions advocated by the Bank at that time. While the Government held regular consultations with the World Bank on economic policy and some social policy issues, direct involvement of the Bank in the area of pension policy was rather limited. One Bank-commissioned expert gave advice on draft legislation for the third pillar in 1998, and one seminar to discuss the policy preferences of the Social Security Reform Commission concerning the second pillar was held in 1999 (see also Lindeman 2004). This limited involvement relates partly to the fact that the Estonian Government did not seek to take any structural adjustment loans from the Bank to reform the pension system. However, it must also be said that by the time Estonia entered into serious pension reform debates, the Bank’s approach had shifted towards a more flexible and country-specific one (see e.g., Holzmann 1999) compared to the policy prescription in the 1994 Averting the Old Age Crises.
99
Council of Europe Estonia became a Member State of the Council of Europe in 1993, and this membership was widely considered an important step in integration into the European structures. One of the key instruments of the Council of Europe – the European Social Charter – had a clear influence on Estonian pension policy objectives. The 1997 Conceptual framework for pension reform identified compliance with the European minimum standards on social security as one of the main objectives of the reform. These standards are prescribed by the European Code of Social Security, which in turn is modeled on ILO Convention 102, Social Security Minimum Standards. In May 1998, the Government signed the revised, 1996 version of the European Social Charter. The Charter was ratified by Estonia in September 2000, including Article 12 on the right to social security, thus committing it to observe the standards of the Code. For old-age pensions, the Code requires that the benefit of a person with an insurance period of 30 years shall correspond to at least 40 percent of the wage of an ordinary adult male laborer. Analysis of the Ministry of Social Affairs at the time that the Government was preparing to ratify the Charter indicated that Estonia barely exceeded this minimum standard: the old-age pension for the standard beneficiary amounted to just 41.4% of the average net wage of a male production worker in 1998. However, in political terms, it was inconceivable that the Government would abstain from ratification of this article of the Charter, as such an action could clearly turn the pensioner electorate against it. Recognizing the political realities, the Government signed the Code in January 2000; and the instrument was ratified in May 2004, including the part on old-age pensions. Nevertheless, in 2001 the average replacement rate for a person with 30 years insurance record dropped below the standards of the Code – to 37.4% – as the wages of production workers had increased substantially. However, additional pension increases in 2002 again raised the replacement rate above the minimum level, to 41.6%. The three-party political coalition (Res Publica, Reform Party and People’s Union), which took seats in Government in 2003, promised in its coalition agreement to maintain the level of pensions above the minimum standards of the Code, thus reaffirming the Government’s earlier commitment. The same position has been upheld by a renewed coalition, where Res Publica was replaced by the Center Party in 2005. Thus, the 40% replacement rate stipulated by the Code and Charter played an important role in Estonian pension policy debates. It has served as a key benchmark against which the adequacy of the system has been measured; and it has been widely accepted that this standard will be observed in the future. Furthermore, through ratification of the Code, this is now an international commitment of the Estonian Government.
100
European Union Estonia submitted an application to join the European Union in November 1995 and, from that time, EU accession was high on the political agenda of every successive Government. Although the practical implications of the EU aquis communautaire for the Estonian pension system were rather limited, the role played by the EU in the Estonian reform was more substantial than generally recognized. As to the impacts of the EU aquis, two issues can be outlined. Firstly, the EU social security co-ordination rules for migrant workers (Regulations 1408/71 and 574/72) implied an obligation to export pensions to other EU Member States upon accession. Until accession, the payment of pensions was limited to persons residing in Estonia, export of pensions being possible only under bilateral agreements.88 This new obligation would increase pension expenditures, but only slightly due to limited number of persons concerned. A second relevant aspect of the EU legislation concerned equal treatment of men and women. Different qualifying conditions for men and women inherited from the Soviet pension system were already largely equalized with the 1993 State Allowances Act, with the main remaining exception being pensionable ages for men and women. However, derogations in respect of pensionable age in statutory systems were allowed by the relevant EU law (Directive 79/7). Nevertheless, the EU equal treatment principles were used by the Government to justify the plan on gradual equalization of pensionable ages for men and women. Considerably more important than these limited impacts of EU legislation was the placing of pension reform among the economic criteria for accession by the EU Council and Commission. This was to some extent surprising as at that time, the EU did not have any common policy on pensions.89 Indeed, since according to the EU basic treaties the regulation of pension systems is within the competence of the member states, the grounds for putting forward conditions related to the pension system were unclear. However, in the EU Commission’s opinion Agenda 2000 (published in July 1997) on Estonia’s economic progress as a candidate country, the Commission expressed concern that, “reform of the pension system has not yet started.” (European Commission 1997:39). The descriptive and analytic part of the opinion made no further reference to this subject. For example, it did not state what problems were observed regarding the existing system or what was to be the expected outcome of reform. Notably, as the opinion was delivered after the Estonian Government had adopted the Conceptual framework for pension reform in May 1997, the opinion could be interpreted as a silent approval of reform principles coupled with a concern that the reform had not yet been started. This position became explicit in the later EU documents.
88 89
Estonia had such agreements with Latvia, Lithuania, Finland and Ukraine. The EU common pension policy objectives were developed only in 2001 and 2002 as a part of the Lisbon process.
101
In the 1998 Accession Partnership between the EU and Estonia, adoption of key legislation linked to pension reform was explicitly listed under short-term priorities for 1998.90 With the 1999 amendments to the Accession Partnership, completion of the necessary legislation for pension reform was regarded as a short-term priority for 2000, while completion of the whole reform was viewed as a medium-term priority.91 In fact, Estonia became formally committed to meeting these objectives since, according to the Accession Partnership agreement, pre-accession financial assistance under the PHARE, ISPA and SAPARD programmers was conditional on fulfillment of obligations taken under the Accession Partnership. The EU Commission closely monitored the reform process and reflected upon developments in its annual progress reports. For example, in the progress report published in November 2000 (i.e., after entry into force of the new first pillar legislation), the Commission noted (European Commission 2000): “Progress on pension and health-care reform has been steady. With the entry into force of the State Pension Insurance Act on 1 April 2000, the long term financial sustainability of the first pillar (pay-as-you-go element) has been reinforced. (p.25) /…/ Nevertheless, /…/ not enough progress in containing government expenditures has been made, especially in the pension and health care reform area. (p.31) /…/ further measures need to be taken in order to complete the pension reform programmed. Therefore, this priority has been only partially met. (p.89)” The Commission expressed satisfaction with the new State Pension Insurance Act, but at the same time considered that this was not enough, and Government should pursue other elements of the earlier envisaged reform. In the 2001, progress report published in November 2001 (i.e., after adoption of the second pillar legislation) the Commission stated (European Commission 2001): “The pension system is being gradually reformed to establish a 3-pillar model. /…/ The law establishing the mandatory funded scheme (second pillar) has been adopted by parliament and will come into force in 2002. (p.29) /…/ The pension reform has been adopted. The legal, institutional, and regulatory framework is in place and enforcement is largely adequate. (p.33) /…/ The necessary legislation for pension reform has now been adopted. /…/ Therefore, this priority has been met to a large extent. (p.94)” The Commission thus notes with satisfaction that the second pillar was implemented and regards the short-term pension reform priority as having been met. To conclude, the European Commission acted as a guarantor of implementation of the reform. Though its assessments were sometimes vaguely worded, the Commission’s position was clearly in support of the Government’s reform plan. As quick accession to the EU was a consensual priority for all major political parties represented in parliament, commitments taken in the Accession Partnership helped to sustain the reform timetable (Leppik 2003). 90
91
Council Decision 98/264/EC of 30 March 1998 on the principles, priorities, intermediate objectives and conditions contained in the accession partnership with the Republic of Estonia. Official Journal of the European Union, L121 of 23 April 1998, p.26–30. Council Decision 1999/855/EC of 6 December 1999 on the principles, priorities, intermediate objectives and conditions contained in the Accession Partnership with the Republic of Estonia. Official Journal of the European Union, L335 of 28 December 1999, p.35–40.
102
International Monetary Fund From the early 1990s, after introduction of the Estonian Kroon based on a currency board arrangement, the Government held regular consultations with the IMF on monetary policy within the framework of so-called stand-by arrangements.92 IMF staff reports on Estonia and its memoranda on economic policies make pension policy a relatively common theme, placing it among the structural policies to support general economic policy. The IMF was known for its support for private pension arrangements. Following consultations in the fall of 1997 (i.e., after the Government had approved the Conceptual framework for pension reform), the IMF Board of Directors (IMF Press Information Notice No 97/41, 24 December 1997): “/…/ urged the authorities to press ahead with privatization of the large infrastructure enterprises and reform of the pension system /…/”.93 A shift in the IMF position occurred in 2000. The IMF mission, which visited Estonia in summer 2000, expressed rather skeptical views on the second pillar and advised the Government to reconsider its reform plan or at least to limit the size of the second pillar to keep transitional financing costs under control. Following the mission, the IMF Board of Directors noted that (IMF Public Information Notice No 00/49, 11 July 2000): “On pension reform, Directors agreed that a second, fully funded, defined contribution pillar has certain advantages, but that it would not, by itself, solve issues arising from the adverse demographic shift. /…/ Directors stressed that care would need to be taken that the transition costs associated with a second pillar are constrained to avoid budget pressures or an excessive increase in public debt.”94 Similar concerns were expressed later in the IMF paper on pension reform in the Baltic countries (Schiff et al 2001). These events placed the Estonian Government in the curious situation where it had to justify the feasibility of the second pillar to the formerly enthusiastic IMF. In sum, international influences on Estonian pension reform derived mainly from these four international agencies. The role of other organizations (e.g., ILO, OECD) was more modest. Besides these big international players, bilateral assistance from the United Kingdom also deserves a mention. The UK Know How Fund financed assistance to the Ministry of Finance in developing pension projections. Assistance with modeling was provided by Callund Consulting Ltd.
92
93 94
Stand-by credits were promised by the IMF in case of unexpected balance of payments needs. In reality, the need never arose and Estonia never used such credits. http://www.imf.org/external/np/sec/pn/1997/pn9741.htm http://www.imf.org/external/np/sec/pn/2000/pn0049.htm
103
4.8. EARLY POST-MULTI-PILLAR REFORM EXPERIENCE 4.8.1. Results of first pillar reforms Changes in the number of pensioners Changes in the state pension system affected the total number of pensioners only slightly. From 2000 (pre-reform) to 2001 (post-reform), the number of pensioners declined by about 6,000 persons, or less than 1.5%, mainly due to some former pension recipients (disabled children) being transferred to state benefits financed from general revenues. More considerable were the influences of structural changes internal to the state pension system. As a result of the transition from invalidity pensions to work incapacity pensions, the total number of pensioners in this category declined by 23,000. At the same time, the number of old-age pensioners increased by 13,000 persons. This reflected largely the transfer to old-age pensions of invalidity pensioners who were already over the pensionable age. Due to the introduction of a qualification period for work incapacity and survivors’ pensions, some of the former recipients of these types of pensions were transferred to the national pension, contributing to a rapid increase in the number of recipients of this benefit (Table 13). Table 13. The number of pensioners, 2000–2006 (beginning of years)
Old-age pension Superannuated pension Work-incapacity pension Survivors' pension National pension Total
Pre-reform Post-reform 2000 2001 2002 2003 2004 2005 2006 284327 297363 298490 296836 294063 294736 292970 3240 3369 3386 2839 2820 2821 2848 66814
43394
47140
51339
55480
59174
61921
15318 1655 371354
15712 14017 8183 7924 9312 9766 6816 7481 11391 11012 9438 8110 366654 370514 370588 371299 375481 375615
Source: Ministry of Social Affairs, National Social Insurance Board
In the wake of the reform, the total number of pensioners stabilized at around 370,000 persons. The number of old-age pensioners has slightly declined. After the sudden reform-induced decrease, the number of work-incapacity pensioners has shown a steady increase and approaches the pre-reform level. At the same time, further restrictions introduced on eligibility for survivors’ pensions in 2002 have decreased the number of recipients of this benefit, but produced an increase in the number of national pensions on grounds of survivorship. The total number of pensioners in 2006 exceeds the pre-reform level, being comparable to the number of pensioners in mid 1990s. 104
While the increase of the pensionable age is causing a decline in the total number of old-age pensioners, the possibility of early retirement – introduced by the State Pension Insurance Act from 2000 – has been used relatively broadly. In 2000-2005, early retirement pensions accounted for about one-fifth of all newly granted old-age pensions, although their share has slightly declined (Table 14). Table 14. Newly granted early retirement pensions in 2000–2005 Year 2000 2001 2002 2003 2004 2005
The number of newly granted early retirement pensions 2170 2363 1723 1564 1867 1414
Share from all newly granted old-age pensions 24 % 21 % 21 % 20 % 19 % 18 %
Source: National Social Insurance Board, calculations by author
At the same time, the possibility of deferring receipt of an old-age pension and thereby receiving a larger benefit, introduced in 2002 to counterbalance the option of early retirement and to allow for more flexibility individual retirement decisions, has been used rather seldom in spite of strong financial incentives. In 2002, a deferred old-age pension was granted to only 87 persons, followed by 79 new cases in 2003, 92 cases in 2004 and 82 cases in 2005. As Fultz (2006) indicates, older persons in Estonia apparently apply a high discount rate in evaluating the new increments: they prefer the simultaneous receipt of wage income and a pension today to larger pension later, at a time when they will have no other income. A closer look at the recipients of early retirement pensions indicate that about 80% of them were out of a job at the time the pension was granted (Võrk and Uudeküll 2002; Tiit et al 2004). Unemployment has made early retirement a practical necessity for many older persons, despite attractive financial incentives to delay drawing their pensions (Fultz 2006). Looking further at the interaction of the pension system and the labor market, it can be observed that following a decline in the employment rate of older workers (aged 55-64) in the first half of 1990s, there was a recovery in the second half of the decade and the employment rates have continued to increase in early new century (Leetmaa et al 2004).
105
Men
100 90 80 70 60
50-54
50
55-59
40
60-64 65-69
30 20
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
0
1989
10
Women 100 90 80 70
50-54
60
55-59
50
60-64
40
65-69
30 20
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
0
1989
10
Source: Estonian Labor Force Surveys, Leetmaa et al 2004, updates by author
Chart 19. Employment rates (%) in age groups 50–64, 1989–2005
These developments have been attributed to legislative changes in the pension system: the increase of the pensionable age, which commenced in 1994, and the payment of full pensions to working pensioners, beginning in 1996 (Tiit et al 2004). The increase of employment rates has been most significant for women in age 106
group 55–59, i.e. the age group, which is most strongly influenced by increase of pension age. However, there is also another supply side factor, which keeps older persons in the labor market – the relatively low real value of pensions. Nevertheless, the increase of the pensionable age, accompanied with the increase of employment rate, has had a positive impact on the labor market and thus contributed to the sustainability of the pension system. Further evidence in this respect is the increase of the average exit age from the labor market, which in 2002 reached 61.6 years, exceeding the EU average (Leetmaa et al 2004). Nevertheless, as the analysis of early retirement pensions indicates, there appears to be segregation in the labor market situation of older workers – whereas on average the employment rate of older persons (55–64) has increased, some long-term unemployed are unable to find a job and enter the pension system before the normal pensionable age. While the payment of a full pension to working pensioners has contributed to increasing the activity rates of older persons, there is another side of the coin, as the measure also provides an incentive to draw pension as early as possible.95 As a result, the average effective pensionable age is below the average exit age from the labor market. A study by the PRAXIS Center for Policy Studies showed that the average effective pensionable age in 2001 for both genders was about 2 years below the statutory pensionable age.96 About one-third of women and one-half of men actually retired before the normal pensionable age, i.e., were pension recipients at least a year before reaching the pensionable age (Tiit et al 2004). At the same time, among population in pension age, the coverage of the state pension system is practically universal (Table 15). The difference from 100% is explained by the number of pensioners receiving pensions from other states97. Table 15. Coverage of the state pension insurance in 2003
Men Women Total
Population in pension age 84276 200766 285042
Pensioners in pension age 81737 198704 280441
Share of pension recipients in population over pension age 97.0 % 99.0 % 98.4 %
Sources: Social Insurance Board, Estonian Statistical Office, PRAXIS
95
96
97
It should be noted that early retirement pensions are not paid in case of working. However, old-age pensions on favorable conditions and superannuated pensions, which are also granted before the normal pensionable age, can be combined with earnings from work. The pensionable age in 2001 was 58 for women and 63 for men. By 2004, the pensionable age of women had increased to 59. Mainly ex-military receiving pension from Russia.
107
Changes in the real value of pensions and replacement rates Since politicians declined to provide automatic indexation of pensions in the 1998 State Pension Insurance Act, the increase of pensions in 2000 and 2001 was still dependent on political decisions. Due to the fact that pensions had been increased by 20% in 1999, which caused exhaustion of earlier reserves by 2000, there was no pension increase in 2000, as previously explained. In 2001, the average old-age pension was increased by about 3.3% against an inflation rate of 5.8% and an increase of the average wage of about 11% (Table 17). As a result, the real value of pensions and the replacement rate declined in both 2000 and 2001 (see also Tables 10 and 18). However, over the 5 year period 2001–2005, the real value of pensions has increased by 40% (Table 16). Table 16. Average old-age pension (annual average, EEK), 2000–2005 Average old-age pension Average old-age pension in real terms (2000=100)
2000 1532 100
2001 1583 98
2002 1758 105
2003 1985 117
2004 2244 128
2005 2558 140
Source: National Social Insurance Board, Ministry of Social Affairs
Starting in 2002, regular indexation was introduced, using an index that gives equal weights to the consumer price index and the increase in social tax revenues. The latter factor, in principle, equals the increase of the total wage bill, depending thus on changes in the number of insured persons and changes in the average wage to which contributions apply. Pensions are indexed once a year on 1 April. In practical terms, in 2002 the CPI had increased by 5.8%, while social tax revenues increased by 11.0%.98 According to the index formula, pensions were increased by 8.4%, i.e. 11% plus 5.8% divided by 2 (Table 17). The value of pension index decreased from 8.4% in 2002 to 6.3% in 2004 due to the decline in CPI. At the same time, the annual increase of social tax revenues remained at the level of 11%. As the number of pensioners remained stable in 2002–2004 around 370 thousand persons (see Table 13), increase of pensions with such index did not use up all revenues from social tax and created surplus in the pension insurance budget. In this situation, the coalition, which entered into Government in 2002, initiated additional pension increases on top of the regular indexation from July 2002.
98
To calculate the pension index, the CPI change and increase of social tax revenues from the previous year is taken, i.e. it is an ex post facto indexation. Accordingly, in Table 17, the CPI change which is used for determining the 2002 pension index is from 2001, etc.
108
Table 17. Adjustments in pensions (the pension index and its components, plus ad hoc increases), 2002–2006 Increase by pension index CPI change Increase of social tax revenues Additional ad hoc pension increase 99
2002 8.4 % 5.8 % 11.0 % 3.7 %
2003 7.4 % 3.6 % 11.1 % 5.2 %
2004 6.3 % 1.3 % 11.3 % 4.8 %
2005 2006 6.7 % 9.7 % 3.0 % 4.1 % 10.4 % 15.3 % 11.5% 4.6%
Source: National Social Insurance Board, calculations by author
The same approach was also continued in 2003 and 2004 by the next Government, which came to power after 2003 elections. Rather than changing the pension index, the Government (in fact another new cabinet after 2005 changes in coalition) implemented additional pension increases also in 2005 and 2006. The average net replacement rate of old-age pension, which jumped to over 50% in 1999 (see Chart 16), declined by 2001 again to the level of 43% – the same level as observed in 1995–1998. The indexation and additional pension increases in 2002– 2005 have maintained the average replacement rate largely constant (Table 18), at the level comparable to mid 1990s. Table 18. Development of the replacement rate of the average old-age pension, 2000–2005 2000 Average gross earnings 4193 (EEK)100 Average net earnings (EEK) 3311 Gross replacement rate of 36.5% average old-age pension Net replacement rate of 46.3% average old-age pension
2001 4658
2002 5247
2003 5824
2004 6397
2005 7208
3707 34.0%
4104 33.5%
4527 34.1%
5050 35.1%
5831 35.5%
42.7%
42.8%
43.8%
44.4%
43.9%
Source: National Social Insurance Board, calculations by author
It is clear however, that without additional ad hoc pension increases, using only indexation, the replacement rate would have declined. This relates to the character of the pension index – in a situation where wages increase more than prices and the number of insured persons increases (or at least remains constant), the index results in declining replacement rate even though the real value of pensions increases. This aspect was not debated publicly at the time that indexation was introduced. The 99
100
In fact, different methods have been used for ad hoc pension increases. In July 2002, the value of a service year was increased by boosting only the employment-related part of the formula. In July 2003, only the flat-rate base amount was increased. In April 2004, July 2005 and April 2006 both the base amount and the value of a service year were increased. The percentage indicated in Table 17 refers to the increase in the average old-age pension. The average earnings upon which social tax has been paid, includes wages of employees and taxable income of self-employed persons.
109
arguments used by the Government coalition in 2001 (when indexation was legislated) characterized indexation as a desirable tool to put an end to political manipulation of pension increases before elections. Nevertheless, the current method of combining indexation with ad hoc increases does not provide a full solution to this problem either. Tiit et al (2004) have suggested that in order to avoid a substantial decline in the replacement rate, the portion of social tax revenues counted in the first pillar pension index could be increased to 2/3 (from the current 1/2). Although this would increase the deficit of the first pillar in the next decade, it would not jeopardize long-term financial sustainability. While the average replacement rate measures the relative income position of the average pensioners to the average worker, this does not give a clear picture on changes in income position of those cohorts who currently retire. To analyze this, individual replacement rates have been calculated for persons who retired in 2001 based on micro data of the pension insurance register (Table 19)101. Table 19. Frequency distribution of old age pensions granted in 2001 based on individual gross replacement rates Replacement rate bracket 0–20% 20–40% 40–60% 60–80% 80–100% 100–200% 200–300% 300–400% over 400% No earnings Total Median replacement rate
Share of total old age pensions granted in 2001 6.6 22.2 14.6 9.5 6.1 9.7 2.1 1.3 3.9 24.1 100.0 51.4%
Sources: Social Insurance Board, PRAXIS Note: Median replacement rate is calculated only for cases, where the person had some earnings
These data indicate that nearly one quarter of persons who retired in 2001 had no earnings at all in previous year (i.e. 2000). For them, granting of pension provided a stable income. For another 23%, net replacement rate was over 100% (gross replacement rate over 80%), meaning that their disposable incomes increased 101
The analyses covered 22934 persons who retired in 2001. Individual replacement rate was calculated by dividing the amount of granted pension to average gross earnings over the previous calendar year upon which social tax was paid. Replacement rate was not calculated in case the person had no earnings over the previous year.
110
compared to pre-retirement earnings. The median gross replacement rate was 51.4%, corresponding to median net replacement rate of nearly 65%102. This is primarily due to lower wages (compared to the national average) of older workers. These findings indicate that for many persons who retire, the actual decline in incomes is not as high as the average replacement rate would suggest. However, this situation is likely to change, considering the distributional effects and incentive structure of the reformed first pillar. As shown by Tiit et al (2004), the current distribution of wages in Estonia could in 40 years time lead to a situation where about 17% of old-age pensioners receive a minimum pension.103 This group would consist mainly of workers receiving the minimum wage and those who have substantial gaps in their careers. At the same time, Lindell (2001) has pointed out that, with the current minimum pension guarantee, the incentives for minimum wage earners to contribute are low. According to her calculations, the individual gross replacement rates under the new first pillar formula would range from 50% for minimum wage earners to 27% for those earning 3-times the average wage.104 Changes in financing Against the background of high economic growth rates (see Table 6) and related increase in real wages coupled with recovery in employment situation (Chart 7), social tax revenues continued to increase steadily over the period of 2000-2005 with annual increase rates over 10%, reaching 15% in 2005 (see also Table 17). Table 20. Revenues, expenditures and reserves of the state pension system (in million EEK), 2000–2005 Social tax revenues Second pillar contributions State budget allocations Other revenues Total revenues Total expenditures Cash reserves at year end Annual change in reserves
2000 6297 254 3 6554 6504 20 -196
2001 6988 242 6 7236 6648 608 +588
2002 7754 -55 346 2 8048 7325 1332 +723
2003 8625 -530 375 2 8473 8198 1607 +275
2004 9520 -883 388 2 9027 9158 1475 -131
2005 10975 -1167 408 3 10218 10501 1192 -283
Source: National Social Insurance Board, calculations by author 102
103
104
Median replacement rate was calculated only for those persons who had at least some earnings over the year before retirement. This projection was based on the analysis of actual pension insurance coefficients of all insured persons from 1999 to 2003. Whitehouse (2004) has argued that the Estonian first pillar provides a linear individual replacement rate. While his paper is methodologically interesting, his calculations are unfortunately not correct due to mistakes in representing the pension formula. He has taken into account neither the flat-rate base amount nor the minimum pension guarantee.
111
It is noteworthy that the surplus of the state pension budget continued to increase in the first years after introduction of the second pillar (2002–2003), which redirected part of social tax to private pension accounts of participants. At the same time the average replacement in the first pillar remained stable (see Table 18). Further increases in the second pillar participation rate (see Chart 21) have caused a current deficit of the first pillar (expenditures exceeding the revenues for the same year) in 2004 and 2005, which was financed from reserves. Nevertheless, by the end of 2005 the state pension system retained reserves amounting to 1.2 billion EEK or about While pension expenditures as a share of GDP declined in 2000-2001 in aftermath of rapid increase in 1999, from 2002 the expenditure level has stabilized and slightly increased in 2003–2005 (Chart 20). The expenditure data thus do not uphold arguments that the burden of transition costs associated with the 2002 introduction of the second pillar was placed of shoulders of current pensioners. If that were the case, a relative decline in pension expenditures as a share of GDP should be observed. 8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0 Pensions in GDP (%)
2000
2001
2002
2003
2004
2005
7.0
6.4
6.3
6.4
6.5
6.4
Sources: Ministry of Social Affairs, calculations by author Note: 2005 – preliminary data
Chart 20. Pension expenditures in GDP (%), 2000–2005
4.8.2. Transition to the mixed system The second pillar became operational on 1 July 2002, from which date contributions to second pillar pension funds could be made. Applications to join the second pillar were accepted beginning on 4 May 2002, the first deadline set for 31 May 2002. 112
Number of participants by phases During the first phase of the open season, which lasted only one month, about 37,000 persons (6% of insured persons) joined the second pillar in order to begin accumulating their contributions on 1 July 2002. While the number of those who signed up in this phase fell below what was predicted, the next phase brought a real boom in joining the multi-pillar system. By the end of the second phase, 31 October 2002, some 170,000 new participants had signed up. They started to accumulate contributions on 1 January 2003.105 The total number of second pillar participants thus reached over 207,000 persons. This period was the last chance to join the second pillar for persons born during 1942– 1956 (i.e., aged 46–60 in 2002). Over the third phase of the open season, which lasted from 1 November 2002 to 31 October 2003, nearly 144,000 more persons joined the system, the total number of participants thus reached over 351,000. Over the fourth period of the open season about 72,000 more persons joined, taking the number of participants over 423,000 persons by 1 November 2004. By 1 November 2005, the end of the fifth period of the open season the number of participants reached 478,000. The participation rate thus reached 62% of the total population in the age bracket 18–60 (i.e., the age group that was eligible to join the second pillar) and 80% of the labor force from 16 to pensionable age. With the number of participants exceeding 485,000 persons by March 2006, future increases in participation rate will come mainly from new entrants to the labor market whose participation is compulsory. However, some increase may be expected also from younger age cohorts who still have the right to join voluntarily. In earlier forecasts, the Government had predicted that under optimistic scenario participation in the second pillar would attract 50% of those insured with the first pillar in course of 4–5 years (see Oorn 2004). More conservative assessments suggested that about 50 thousand persons would join in the first round in 2002 and the total number of participants would reach 100–200 thousand persons over a couple of years106. These forecasts were based on earlier public opinion polls. According to a poll carried out by ES Turu-uuringute AS (2002) in April 2002, i.e., immediately before introduction of the second pillar), 6% of working age respondents stated that they intended to join the mixed system in 2002. An additional 20% of respondents expressed their intention to join the system in the future, 33% were undecided, and 41% stated that they either probably or certainly were not going to join the second pillar.
105
106
Applications are accepted until 31 October each year in which case collection of contributions starts from 1st of January of the following year. The first year of reform – 2002 – was an exception since the system was started from July. See expert interviews in Põiklik (2005).
113
From the other side, by the time of implementing the Estonian reform, information on actual participation rates in the Polish and Hungarian second pillars was already available. It was also known that in those countries the number of persons who joined the system voluntarily far exceeded earlier forecasts. However, the Government still considered that in Estonia the rate of those joining the system voluntary would probably stay below what was observed in Poland and Hungary, because of the additional contribution required in the Estonia second pillar. The actual developments showed that the latter discouragement was overshadowed by the advertisement campaign and sales work of pension fund managers and the public awareness campaign initiated by the Government. It seems that the awareness and advertisement campaigns were successful primarily in convincing those persons who were undecided at the time the reform was launched. 500000 450000 400000 350000 300000 250000 200000 150000 100000 50000
1.03.06
1.01.06
1.11.05
1.09.05
1.07.05
1.05.05
1.03.05
1.01.05
1.11.04
1.09.04
1.07.04
1.05.04
1.03.04
1.01.04
1.11.03
1.09.03
1.07.03
1.05.03
1.03.03
1.01.03
1.11.02
1.09.02
1.07.02
1.05.02
0
Source: Estonian Central Depository for Securities Note: The chart reflects the dates of submission of applications to join the second pillar. Actual participation starts from the date the first contribution is made
Chart 21. Cumulative increase in second pillar participation, 2002–2006
Chart 21 illustrates the cumulative increase in second pillar participants.107 The five deadlines for submitting applications appear as waves – many people joined during the last week or even on the very last day before a deadline. The phenomenon can be termed as the ‘deadline effect’. 107
The term “participants” here refers to persons who have submitted applications to join the system voluntarily or who have joined the system on a compulsory basis and for whom a pension account has been opened by the ECDS. The term includes persons who are temporarily inactive or unemployed. In the other words, the number of active contributors is smaller than the number of participants. Although contributions to the second pillar can be made only by persons who are at least 18 years old, some persons who are 16 or 17 have already chosen a fund and therefore also appear as participants.
114
Switching behavior: distribution of second pillar participants by age and gender Although the employment rate for men is higher than that for women in Estonia, women outnumber men amongst second pillar participants (Chart 22).108 Men 46%
Women 54%
Source: Estonian Central Depository for Securities
Chart 22. Participants of the second pillar by gender (as of 1 November 2005)
Chart 23 shows second pillar participation rates in different age cohorts in 2003, representing participants as a share of the total population in one-year age groups. The chart reflects another interesting aspect of the deadline effect. While participation rates among 20–41 year olds are in the range of 40–50%, participation rates in the age group 42–46 are over 10 percentage points higher. This is because, for the latter age group, 31 October 2003 was the last date to join the second pillar and many persons did so, so as not to miss the last chance. 100%
Share of population
90% 80%
Men
70%
Women
60% 50% 40% 30% 20% 10% 0% 16 18
20 22
24 26 28
30 32 34
36 38
40 42 44
46 48 50
52 54
56 58 60
Age
Source: Estonian Central Depository for Securities, Estonian Statistical Office, Leppik and Võrk 2006
Chart 23. Second pillar participants as percent of age group by gender (as of 1 November 2003) 108
In 2003, the employment rate for men was 66.7% and for women, 58.8% (15–64 age group).
115
Due to the shorter deadlines for older cohorts (see Table 11), older persons had to make their decision in a situation of limited information, while younger cohorts had the option to adopt a ‘wait-and-see’ strategy, first observing how the system was evolving and how the funds performed, and then make a more informed decision at a later stage. In fact, since persons over 46 years of age could join only in 2002 (i.e., the first year of reform), when information about the actual performance of the system was limited, the participation rates in older cohorts are relatively low. By the end of 2003, there was already information on the experience of the first year, and in this situation, the final deadline apparently increased participation rates among 42–46 year olds. Participation rate in the 18–20 age group is over 90%. Although for this age group participation is compulsory, there is no automatic enrolment of all residents. There are two ways to become a participant. The major channel is through submitting applications – this applies also for compulsory participants as in the application the person indicates the choice of a pension fund. Persons whose participation is compulsory by age but who have not submitted applications enter the system when they receive their first wage payment that is subject to social tax. At this point, a pension account is opened and the person is assigned at random to a fund. The lottery includes only conservative funds. However, 18–20 year olds who have not submitted applications and who have not yet received income, which is subject to social tax, do not show up as participants in the system until they receive such income. Fund choice: distribution of participants by fund manager and risk level Considering that second pillar participants have to bear the entire investment risk, it may be somewhat surprising that the majority – by 1 November 2005 over three quarters, 75.4% − have joined higher-risk funds, which invest up to 50% of assets in equities. Obviously, this relates to their hopes for higher returns from their investments and perhaps to a lesser degree of risk adversity than has been observed among workers in countries with longer experience with stock market volatility. Only 8.2% have chosen low-risk funds, while the remaining 16.3% joined mediumrisk funds. However, there is a clear age-gender pattern, with older persons and women opting more frequently for low-risk funds. In the 45–60 age group 38% joined a low-risk fund and 34%, a medium-risk fund (2004 data).
116
78.7% 72.6%
80% 70% 60% 50%
Men Women
40% 30% 14.1%
20%
18.2% 7.2%
9.1%
10% 0% Higher risk
Medium risk
Lower risk
Source: Estonian Central Depository for Securities
Chart 24. Distribution of second pillar participants by pension fund risk level (as of 1 November 2005)
Hansa Asset Management – the market leader operating the biggest pension funds – recommends higher-risk funds to persons under 45 years of age, suggesting that they switch to medium-risk funds beginning at the age of 45 and to lower-risk funds, at the age of 55 (Hansapank 2004). Factors explaining the high participation rate in the second pillar The unexpectedly high number of persons choosing to join the second pillar on a voluntary basis leads to a question, what are the main factors behind it? Success of the reform can be attributed to several factors (see also Leppik 2004). Attractive design of the reform can be seen as one of the factors. The designers and promoters of the reform succeeded in changing its ‘optics’. What could be perceived as an increase in the total contribution rate was largely perceived as a bargain offered by the state. The main slogan promoted by the Government and pension fund managers was: “You pay 2%, the state pays 4%”. As another crucial factor, the reform debates reached the grass-roots level. The Government information campaign succeeded in bringing the issue onto public agenda109. Obviously, the advertisement campaign of fund management companies and direct sales work of bank tellers and fund managers cannot be underestimated. However, it seemed that often ordinary people sold the idea to other people and joining the second pillar became a social-psychological phenomenon, creating a 109
See Veel (2004) and Põiklik (2005) and for analysis of Governments’ communication campaign.
117
certain snowball effect – once a critical mass of participants was achieved, other persons joined just as their friends and family members had done so. Transparency can also be pinpointed as an important factor. In an era when Internet is a major channel of communication and information, many Estonians could take all the needed actions from the convenience of their own home or office: it is possible to join the second pillar, to choose a pension fund, to check the balance of the pension account, to view the investment portfolios of pension funds, to compare the performance of different funds etc. A special web site – www.pensionikeskus.ee – established by the Estonian Central Depository for Securities – serves as the main clearing house for information on the second pillar. Finally yet importantly was the role of efficient implementing bodies and infrastructure. The tasks of co-coordinating the logistical side of the second pillar were delegated to the Estonian Central Depository for Securities – a private company in charge of the whole infrastructure for securities market in Estonia – turned to be motivated and efficient in setting up user-friendly procedures for providing information to fund participants and solving the daily problems that have arisen in the implementation process110. This was also noted by Lindeman (2004:15): “Estonia’s implementation encountered few of the glitches that have plagued second pillar implementations in the general region. This was because it could build upon two already up-and-running agencies [the Tax Office and Central Depository for Securities] that were fully up to advanced system standards.” 4.8.3. Transition costs Implementation of the second pillar changed the allocation of the social tax. For participants in the mixed pension system, 4 percentage points of social tax were redirected to their individual accounts, causing the so-called transition cost of the reform. More specifically, transition costs can be divided into: • gross transition costs, expressed as the amount of social tax which is transferred to the second pillar on behalf of persons who have joined the new scheme; • net transition costs, expressed as the difference between post-reform revenues and expenditures of the first pillar. When drafting the state budgets for 2003 and 2004, the Government envisaged covering the transition costs by reducing the reserves of the first pillar, i.e., using the surplus developed in previous years. However, in reality, despite of transferring parts of social tax to the second pillar pension funds and increases of state pensions in 2002–2004, the surplus of the first pillar increased. In other words, in spite of gross transition costs, there were no net transition costs to the pension reform in 2002–2004. On the contrary, reserves of the first pillar increased as revenues from the remaining social tax exceeded expenditures on state pensions (Table 20). 110
118
In contrast, in Poland, the Social Security Institution (ZUS) encountered numerous difficulties in the first years of reform.
This unexpected situation was attributable to a combination of factors. Increasing employment and real wages, boosted social tax revenues (see Tables 17 and 20) while due to the ex post facto pension indexation, the increase of pensions did not exhaust all social tax revenues. Given the decline in the number of old-age pensioners (see Table 13), even with the additional pension increases, which were implemented to maintain the average replacement rate (see Table 18), all social tax revenues were not exhausted. In 2004, the Government established an additional reserve for the first pillar, transferring 532 million EEK (or 0.4% of GDP) from the higher-than-expected general tax revenues of 2003. The total first pillar reserves thus amount to 1.6% of GDP. Additional reserves were further increased in 2005. However, with the continuing increase in second pillar participation and the additional ad hoc pension increases implemented in 2004, 2005 and 2006, the first pillar reserves will be exhausted in the next few years and other revenues have to be sought to cover the first pillar deficit. The issue of transition costs has led to political debates between different parties concerning the method to cover the gap between social tax revenues and expenditures. The Social Security Reform Commission suggested the use of the stabilization reserve in the short run, while for the longer run it proposed transfers from the state budget and possibly the issuance of Government bonds. The SSRC also called for intergenerational equity in covering the transition costs, noting that different instruments place the burden of transition costs on different generations, e.g., the drawing down of existing reserves places the burden on the cohorts that participated in creating them, whereas the use of loans places the burden on future taxpayers. The Government has announced that as the first source they intend to use the existing first pillar reserves (amounting currently to 1.6% of GDP) and subsequently, the stabilization reserve (amounting currently to 3.5% of GDP). The Center Party and the People’s Union have opposed the use of existing reserves and argued that resources have to be found in the general state budget, i.e., from other tax revenues. There have also been accusations on placing a disproportionate burden on current pensioners by limiting the increase of state pensions. However, at least for the time being the developments of total pension expenditures and net replacement rate do not indicate that pensioners have suffered because of launching the second pillar. However, the ambitious plan of the coalition for reducing the income tax rate from 26% in 2004 to 20% by 2009 reduces public revenues and may thereby necessitate cuts in public expenditures111. This adds difficulties to finding resources from the general budget to cover the transition costs of the pension reform.
111
Unless the decline of tax rate is balanced by increase in revenues due to higher tax compliance and/or high economic growth.
119
4.8.4. The third pillar Participation in the voluntary third pillar can take two different forms: • pension insurance policies offered by licensed private insurance companies; • units of pension funds managed by private fund managers. In case of pension insurance policies, the pensionable age is a matter of contract between the person and the insurance company. However, the minimum age at which tax privileges apply is 55. In the case of voluntary pension funds, the participating person decides the time that he/she will withdraw the savings; but again, tax advantages apply only from 55 years of age. Third pillar pensions are thus available up to 8 years before the general pensionable age. Pension savings may also be withdrawn in the event of total and permanent work incapacity. The following taxation rules apply to the third pillar: • contributions (premiums paid on the bases of a pension insurance policy or sums paid for purchasing the units of a voluntary pension fund) are deductible from taxable income up to the ceiling of 15% of annual income; • benefits paid on the basis of a private pension insurance policy or from redemption of the units of a pension fund are taxable at a lower rate (10%), instead of the normal income tax rate; and • periodic, lifelong benefits paid from a defined-benefit type of insurance policy in equal or increasing amounts are not taxable. The tax treatment is thus very favorable. In the case of life-long annuities, neither contributions nor benefits are taxable. This was not the original plan of the Social Security Reform Commission but rather the result of successful lobbying by insurance companies when the third pillar legislation was discussed in Parliament. Five life insurance companies (ERGO Life Insurance, Sampo Life Insurance, Seesam Life Insurance, Ühispanga Life Insurance and Hansapanga Life Insurance) have been issued licenses to sell pension insurance policies under favorable tax treatment. Four fund managers (Hansa Asset Management, Ühispank Asset Management, Sampo Asset Management, LHV Asset Management) currently operate 7 voluntary pension funds. Third pillar pension funds are by their nature similar to second pillar funds, except investment rules are more relaxed, allowing higher degree of investments in equities. This, in turn implies higher risk for participants. Contrary to the earlier fears of insurance companies, the introduction of the second pillar has in fact increased participation in the third pillar, which was relatively low until 2001 despite the very favorable tax treatment. This probably relates to the general higher public awareness of pension issues achieved by the second pillar campaigns. By the end of 2005, over 75,000 persons (ca 12.5% of all employed persons) had concluded a pension contract with a life insurance company, while the number of participants in voluntary pension funds was about 12,500 (ca 3% of employed 120
persons). Insurance reserves under pension insurance policies amounted to 1.3 billion EEK (or 0.9% of GDP), while total assets of voluntary pension funds approached reached 500 million EEK (0.4% of GDP). Insurance companies have dominated the third pillar market, mainly due to their more preferential tax treatment compared to voluntary pension funds. However, in recent years, the rate of increase of the number of participants in voluntary pension funds has been higher than for pension insurance policies, indicating that many persons prefer the greater flexibility of voluntary pension funds compared to higher security of insurance products112.
112
Insurance companies offer pension products with guaranteed rate of return.
121
5. CONCEPTUALIZATION AND THEORIZING OF TRANSFORMATION OF THE ESTONIAN PENSION SYSTEM In this concluding chapter, I shall first summarize the reform stages, the basic policy choices, and the main earlier policy outcomes. This is followed by an analysis of the pension reform process in the light of theoretical approaches presented in the second chapter. 5.1. OVERVIEW OF REFORM STAGES AND POLICY OUTCOMES Transformation of the Estonian pension system occurred in two major waves. The first set of paradigmatic changes took place in 1990–1993 in circumstances of a turbulent period, when both the financing and benefit sides of the Estonian pension system were separated from the Soviet system. This was followed by a period of relative stability when the pension system operated under a transitional arrangement. Over this period, the pension system retained the new paradigm chosen in early 1990s, albeit some important parametric reforms were implemented. The second wave of transformation took place in 1998–2002, when the new three-pillar pension system was introduced. Table 21. Stages of transformation of the Estonian pension system Stages
Dates January 1990
First wave of transformation 1990–1993
May 1991 February 1992 April 1993
Transition period 1993–1998/2000
July 1994 September 1996 July 1998
Second wave of transformation 1998–2002
January 1999 April 2000 April 2002 July 2002
Source: own compilation
122
Main policy changes Financial separation from the Soviet pension system Benefit side separated from the Soviet pension system Flat rate pensions introduced linked to the minimum wage Old age pensions differentiated on the basis of length of service; pension age increased Pensions decoupled from the minimum wage Full pensions paid to working pensioners Tax advantages for voluntary pension schemes (third pillar) Individualized accounting of social tax; determination of new pension rights on the basis of social tax payments Contribution-related element in the pension formula; equalization of pension age of men and women Indexation of pensions Introduction of compulsory funded pension scheme (second pillar)
The first wave of reform in the early 1990s, separating the Estonian pension system from the Soviet one was implemented in different stages, first by shifting pension financing from the general state budget to an earmarked social tax, i.e., converting the system from a non-contributory to a contributory one. New rules were also adopted in an attempt to improve benefits, extend coverage (to include all residents), and the level of protection. However, high social expectations of the population collided with the turbulent conditions during the first years of independence. The introduction of flat rate pensions (linked to the minimum wage) in 1992 is to be regarded as a temporary rescue measure rather than a purposeful shift towards egalitarian principles. At the same time, a considerable retrenchment of the pension system took place in a situation of hyperinflation. In the period that followed, 1993–2000, the application of a macro-level definedcontribution approach limited state pension expenditures to the revenues available from the social tax. The calculation of pensions based on a flat-rate base supplemented by a length-of-service component, resulted in a relatively flat benefit structure. The challenge for policy makers was to satisfy the public desire for a stronger link between benefits and previous earnings, while at the same time keeping all beneficiaries above the poverty level. Considering the relatively low average replacement rate, the question was in essence, how much differentiation may be introduced without increasing the poverty rate of those at the lower end of the pension benefit scale. Table 22. Changes in basic principles of the state pension scheme Pension act 1956 Soviet Pension Law 1991 Pension Law 1992 Decree on Living Allowances 1993 Law on State Living Allowances 2000 State Pension Insurance Act
Basic principle DB DB
Basis of pension rights employmentrelated residence-based
DB
residence-based
MDC
residence-based
old age pension: DB (with elements of NDC) national pension: DB
employmentrelated
residence-based
Calculation principle earningsrelated flat-rate + earningsrelated flat-rate
Financing source general taxes earmarked social tax
flat-rate + length-ofservice flat-rate + length-ofservice + earningsrelated flat-rate
earmarked social tax
earmarked social tax
earmarked social tax + general taxes general taxes
Source: own compilation
123
The second wave of pension transformation – initiated in 1997 with the approval of the Conceptual Framework for Pension Reform – introduced significant parametric changes to the state pension scheme (e.g., equalisation of the pensionable age for men and women, linking pensions to contributions), while even more importantly, setting the goal of establishing a multi-pillar system by adding two supplementary pillars based on pre-funding. The parametric reform of the first pillar addressed the long-term sustainability of the system and attempted to find a better consistency with the dominant public perception of distributional fairness. The state pension system has managed to keep the majority of pensioners above the poverty level, but the average net replacement rate – about 40% – is rather low in the European context, leaving the majority of pensioners in the lower-middle range of the income continuum. Introduction of the pre-funded second pillar based on savings in individual accounts marks a paradigmatic shift in pension policy. The paradigm shift is not merely related to the technical introduction of a multi-pillar system. The paradigm shift impacts the focus of pension debates and the way the topic of old age security is addressed. Reformers inter alia hoped that the increase of the total contribution rate for the second pillar would dampen future pressures to increase social tax for the first pillar. Second, they hoped that the second pillar contribution would enable individuals to visualise the cost element of pensions, which had by far been partially hidden from them. Third, they wanted participation in the second pillar and payment of individual contributions for private pensions to give people a more realistic assessment and expectations concerning future state pensions. It is a matter of time as to whether these expectations will materialise. Estonia was not an international trailblazer in the field of multi-pillar reforms, but the fifth country in Central and Eastern Europe – after Hungary, Poland, Latvia, and Croatia – to implement a three-pillar pension system with a fully funded second pillar. On a broad scale, the Estonian pension reform bears certain similarities with other pension reforms in CEECs, namely the three-pillar framework with a second pillar consisting of privately-managed individual savings accounts paying definedcontribution benefits. However, several parameters of the reform including the configuration of the first pillar as well as the conditions for implementing the second pillar are different from the other countries, making the Estonian case unique. In particular, Estonia has been the only country in Central and Eastern Europe to increase the total contribution rate when introducing the second pillar, using both so-called “top-up” and “carve-out” methods simultaneously. The study has also identified some other features, which make the Estonian pension policy design unique. Even though participation in the second pillar requires additional contributions from employees – 2% of wage – the participation rate is one of the highest among CEECs who have undertaken a similar reform, exceeding 60% of the population of active age in 2005. 124
The popularity of the reform can be attributed to several factors. Firstly, the low replacement rate of state pensions created a favorable context for private supplements. Secondly, the reform design looked attractive. The reform slogan “You pay 2%, the state pays 4%” created an impression of a bargain offered by the state rather than an increase of the total contribution rate. Thirdly, a ‘snowball effect’ can be pointed out. Joining the second pillar became a socio-psychological phenomenon – once a critical mass had joined, others joined because their family members and friends had done so. Fourthly, government awareness campaigns and marketing by fund managers had influences. Finally, the use of the Internet provided ease in both joining the second pillar and monitoring its performance. Efficient implementing bodies and infrastructure appear to be an additional crucial determinant of reform success. The relevant Estonian administrative bodies (in particular the Tax Office and the Estonian Central Depository for Securities) appeared to be efficient in setting up the necessary procedures, in providing information to fund participants, and in solving the daily problems that arose during the implementation process. Last, but not least, timing of the reform appears important. By 2002 when the reform was implemented, the financial sector had consolidated and addressed the problems of the early years of transition. Had the reform been implemented just a few years earlier, e.g., before the 1998 stock market crash in Estonia or 1999 Russian financial crises, probabilities of failure would have been much higher. Introduction of the second pillar has raised controversies concerning transition costs. The main issue is that of equity – whether transition costs fall disproportionately on the shoulders of elderly persons. However, post-reform pension expenditures in GDP have not declined, nor has the average replacement rate. It is noteworthy that against the background of numerous policy changes the main policy outcome indicators have been relatively stable over the period of 1993–2005 (i.e. after the turbulent take-off of the Estonian pension system). The average net replacement rate has been in the range of 40–45%. The total expenditures of pension have been in the range of 6–7% f GDP, while the total number of pensioners has been around 375 thousand persons. This inter alia entails that there were no significant changes in the relative income position of pensioners over the period of 1993–2005. Obviously, this does not imply that the reforms have had minor significance, as the relative stability of outcome indicators is largely due to the reforms, not in spite of them. Furthermore, the stability of macro-level indicators hides important distributional effects of reforms and implications on system sustainability.
125
5.2. ANALYSIS OF THE REFORM PROCESS IN ESTONIA AGAINST EXISTING THEORIES OF PENSION REFORM 5.2.1. Re-experimentation and re-calibration in the Estonian case The conceptual framework elaborated by Ferrera et al (2000), in particular the constructs of re-experimentation and re-calibration, prove useful in capturing the essence of Estonian pension reforms. While Estonian reforms were clearly influenced by elements of policy learning (including, in particular, lesson-drawing), there are several elements of policy innovation that introduce some innovative policy design features into the world of pension reforms. Re-experimentation in the Estonian case – the ‘2+4’ formula for introducing the second pillar – was an essential element of forging political compromise acceptable for three parties of the politically diverse coalition in 1999–2001, and with employers and trade unions. Secondly, it was one of the keys to success of the reform in terms of attracting participants. However, beyond these practical aspects, this innovative policy design has broader merits. Thus far, Estonia is the only country in Central and Eastern Europe that increased the total contribution rate when introducing the second pillar. Other CEECs have used only the carve-out method, i.e., redirected a portion of the former first pillar contribution to the second pillar, while some countries (e.g. Poland) even lowered the total contribution rate. The combined use of top-up and carve-out methods for financing the second pillar is unique. The lesson from this is that similar institutional designs (e.g., fully funded defined-contribution second pillar schemes) may be introduced to fulfil rather different political and policy objectives. In some CEECs, the second pillar was established with an implicit agenda to downsize the compulsory pension system and reduce pension expenditures, whereas in the Estonian case the second pillar extended the compulsory system and required overtly that joiners pay a higher price for a higher pension. Besides re-experimentation, the Estonian pension reform also bears clear characteristics of re-calibration. The transformation process has been conditioned by a set of constraints, resulting from a complex interaction of domestic challenges (e.g. labor market restructuring with resulting decline in employment, decline of birth rates, increase of life expectancy) and external pressures (competitive forces, EU requirements, social standards of the Council of Europe etc.). In response to these constraints, the pension system has been recalibrated along different dimensions. Firstly, there has been a functional re-calibration. This includes redefinition of the contingency, implying a substantial increase of pension age, but also possibilities to combine pension and work income113. Functional re-calibration can also be related to the fact that the reform package included both additions and subtractions. For example, as additions in the first pillar count flexibility around pension age through 113
126
I.e. the new definition of contingency no longer assumes suspension of earnings from work.
the options of early and deferred retirement, while subtractions include tighter rules for acquisition of pension rights on the basis of social tax payments, tightened conditions for survivors’ and work-incapacity pensions, etc. The reform has also entailed substantial distributive re-calibration, albeit, at different stages of transformation the distributive re-calibration has worked in different directions. In the early stage of reform (first wave of reform), interpersonal redistribution was increased while intergenerational redistribution declined. After the first shocks, the level of intergenerational redistribution somewhat increased, although not to the same level as at the outset of the transformation process. With gradual strengthening of the contribution-benefit link, the degree of intragenerational redistribution is declining. Nevertheless, as the history of changing the relative weight of flat-rate versus employment-related elements of the state pension indicates, the appropriate degree of intragenerational solidarity in the state pension system is still contestable. The Estonian reform also entails important elements of normative re-calibration − deliberate shift of weight and emphasis among the instruments and objectives of the system. The former, instrumental aspect is illustrated by the downscaling of the state pension insurance through reduction of the first pillar contribution rate from 20 to 16%, with a subsequent effect on acquisition of future pension rights from the first pillar. The new second pillar based on individual savings approach, serves as another example of shifting objectives. 5.2.2. Explanations along historical institutionalism When testing the Estonian case of pension reform against existing theoretical frameworks on pension reform, let us start with historical institutionalism. According to the logic of path dependency, one could have expected persistence of the Bismarckian features of the former Soviet pension system, perhaps, only in a somewhat modified form. If the defined-benefit logic was to be maintained, one could expect at least some attempts to balance the first economic pressures of transition by increasing contribution (social tax) rates or subsidies from general taxes (see Palier 2001). Indeed, this was the initial strategy in many other Central and Eastern European countries (Hungary, Poland, the Czech Republic, the Slovak Republic, Slovenia) where substantial reform of the benefit side was not undertaken in the early phase of transition, but postponed until the second half of 1990s (see also Guardiancich 2004). In Estonia and other Baltic countries, first – and quite radical – reforms of the pension system were undertaken shortly after the regaining of independence in the 1990s. Even though these reforms did not alter the pay-as-you-go paradigm, they did discontinue the earlier path, inter alia the link between pensions and former earnings, i.e. the earlier defined-benefit character of the system. In Estonia, the approach of increasing contribution rates (social tax) was never used. The rate of social tax allocated for pension insurance has remained unchanged from the time it was introduced in 1991. In spite of pressures from pensioners 127
associations, the option of increasing social tax was virtually excluded from the list of policy alternatives by all major political parties. Besides ideological considerations against tax increases, the vicious cycle of contribution increases observed in some Western European countries was referred to – high contribution rates increase the cost of labor, thereby increasing unemployment, which in turn would necessitate further contribution increases. These self-imposed constraints resulted in a particular design, labeled here as macro-level defined-contribution: the level of pensions depended on available revenues from social tax. Such a policy design can be operational under conditions where revenues from social tax increase rapidly while the total number of pensioners remains largely constant. If social tax revenues decline or if the number of pensioners increase at a higher rate compared to the increase of social tax revenues, this strategy could lead to a reduction of benefits. However, due to favorable circumstances the MDC design was never tested against these negative scenarios. The first wave of transformation in the early 1990s thus constituted a significant pathshift. From the perspective of path-dependency, these reforms should clearly be considered as radical since they discontinued the previous system. However, there is a clear ‘critical juncture’ point, which can explain this path departure – regaining of independence and separation of social systems from those of the former Soviet Union. The path-departure was also necessitated and conditioned by quite extraordinary circumstances (economic crises and related hyperinflation) of first years of economic transition. Nevertheless, the first reforms, although radical in terms of changes to the financing and benefit side, retained the pension system in the public domain. A far more difficult question is what triggered the second, multi-pillar reform. While one of the declared factors was the need to cope with an ageing population, it was mainly a future challenge, not an immediate concern. Therefore demographic developments cannot be considered as a critical juncture point. Clearly, the first reforms left the policy window of opportunity open, as these did not aim to establish a final system, but rather a temporary solution for a transition period. Therefore, the first wave reforms did not set a firm new path for the pension system. However, this indicates only toward the possibility for another pathdeparture, but cannot explain why the path-departure was undertaken. This does not render historical institutionalism completely useless in explaining Estonian reforms. It can be pointed out that parametric reforms of the first pillar, undertaken within the framework of the multi-pillar reform, followed the same path as chosen by the 1993 State Allowances Act: while introducing an earnings-related element, it built upon the former pension formula and entailed a very gradual shift from old to new rules. Furthermore, the privileged rules of some occupational groups (old-age pensions on favorable conditions and superannuated pensions) have remained largely untouched from the Soviet period, in spite of some attempts to cut them, mainly due to high vested interests of beneficiaries and workers employed in those occupations. Nevertheless, historical institutionalism standing alone fails to explain the reasons for a significant path departure associated with the adoption of the multi-pillar paradigm. However, the multi-pillar reform itself has clearly created new 128
constituencies (e.g. participants of second pillar funds and fund managers), establishes a new path for the pension system while conditioning and restricting further policy alternatives, thus creating a new path-dependency. 5.2.3. Explanations along actor-centered institutionalism Müller (1999) has explained the introduction of fully funded second pillar schemes in Poland and Hungary as resulting from a combination of particular actor constellations and structural-institutional contexts. According to her explanatory model, the financial situation of the pay-as-you-go scheme (deficit or surplus), influences the perceived urgency of radical reform. Another crucial factor outlined by Müller is the degree of external debt and resulting influence of international financial institutions that favor pension privatization as a strategy. In her case study, she showed that these two factors triggered the active involvement of two actors – the Ministry of Finance and the World Bank – resulting in adoption of pension privatization. In Estonia, however, the constellation of factors behind the reform was clearly different from that of both Poland and Hungary. The Estonian case conforms to the finding of Müller (1999) in that the main actors shaping the pension reform were the two Ministries in the cabinet, the Ministry of Social Affairs and the Ministry of Finance, while the secondary actors (trade unions, employers’ organizations, financial institutions etc.) were able to influence only the details of the eventual policy design, but not the basic paradigm choice. However, similar to the situation in Slovenia, as described by Guardiancich (2004), the positions of Ministers of Social Affairs and Minister of Finance towards multipillar pension reform in Estonia did not conform to the standard roles ascribed to them by Müller114. The Ministry of Finance was active, but the key role was played by the Minister of Social Affairs. However, at the same time some other observations of Müller hold in the Estonian case. The institutional arrangement of the Social Security Reform Commission (similar to the respective arrangement in Poland) and involvement of different actors from early on in the reform process, helped to find consensus between actors and circumvent opposition to reform. Nevertheless, the Estonian case clearly outlines some of the limitations of explanatory power of actor-centered institutionalism. While the ‘benefit of crises’ argument seems to fit for explaining the first wave reforms – discontinuation of the Soviet legacy and introduction of flat-rate pensions – the economic context where the paradigmatic multi-pillar reform was undertaken does not uphold the ‘benefitof-crises’ hypothesis. In Estonia, there was no fiscal imbalance of the pay-as-you-go pension system at the time the reform was planned, legislated, and implemented. It was the opposite, the public pension system was in surplus, and furthermore, the general fiscal position of the government was rather good. In particular, reserves were created 114
Nevertheless, we shall keep in mind that in the Slovenian case, the sum of the equation was rejection of the multi-pillar reform, whereas in the Estonian case, the choice was pro-reform.
129
prior to reform with reference to the need to shoulder transition costs. It therefore could be argued that with Estonia, it was not a poor fiscal situation that led to reform, but rather its good fiscal position that enabled the undertaking of the multipillar reform. Broadly, this observation conforms with the conclusion of Guardiancich (2004), in that neither fiscal nor demographic factors alone can explain the final choice of reform design. In analyzing the political dimension, it is remarkable that in spite of the changes in the Government, the 1997 concept paper served as the road map on pension reform for successive cabinets until all three pillars were in place. After lengthy deliberations on the reform design, there was substantial continuity in implementation. In contrast, from 1994 to 1997, there were many competing proposals, which primarily centered on the pay-as-you-go paradigm; however, none of which where successful. The short life of governments appeared then to be a factor inhibiting reforms115. The Estonian case of pension reform upholds the conclusion of Grimmeisen (2004), that the composition of government has no substantial explanatory power in the context of post-communist pension reforms. This is contrary to theories that have viewed party politics as a crucial determinant of welfare state developments. For example, Schmidt (1996) has argued that the composition of government is the major determinant of variation in policy choices. In addition, Kitschelt (2001:274) has argued that one should expect that governmental participation of ''parties with a track-record of liberal market policy advocacy is a facilitator of shifting the social policy development'' towards cost containment and institutional revision. However, Grimmeisen (2004) analyzing the cases of Poland, Hungary, and Czech Republic has demonstrated that in those countries the crucial role in adopting pension reform was rather played by the centre-left parties. This finding is also confirmed in the Estonian case. The key role in adopting the multi-pillar reform in Estonia was played by the social democratic party and the trade unions. Nevertheless, this does not mean that the role of liberal and conservative parties was inferior. Both sides of the political spectrum had important influence on the final reform design. Kitschelt (2001) is obviously correct in linking neo-liberalism to cost containment and institutional revision. For the Estonian Reform Party, cost containment and privatization were clearly the preferred alternatives. However, considering the ‘double payment’ problem – the need to finance transition costs – the cost containment argument paradoxically led the Reform Party at some point of time to prefer the no-reform option, as the multi-pillar reform seemed to put at risk their plans for tax cuts. In this perspective, cost containment was a higher value for the Reform Party when compared to partial privatization of the pension system in the framework of multi-pillar reform. While party politics had to do with the choice of a particular reform design, it appears to be a weak explaining factor for the overall choice of paradigmatic reform. This argument is supported by the fact that piecemeal sequencing of the 115
In 1994, there was also a (rather primitive) proposal for radical privatization of the pension system.
130
reform over a period of 5 years (1998–2002) did not derail the move towards multipillar reform, reflecting cross-party consensus. The coalition that finally implemented the second pillar in 2002 was the third coalition following the one that adopted the reform plan. As coalition governments have been made up of 2–3 political parties, in fact 7 different political parties were directly involved either with preparation or with implementation of the pension reform. At the same time, in many other areas of public policy there was no consensus between parties resulting in no reform, as was the case with efforts to introduce an administrative-territorial reform (see Põiklik 2005). Although pension reform plan did create controversies, these were mainly rhetorical rather than serious political conflicts. Considerable analysis has been devoted to the role of international organizations on choice of reform design. Some authors, e.g. Casey (2004), Kulu and Reiljan (2004) apply rather straightforward arguments subscribing the Estonian reform to the influence of the World Bank. However, as shown in this volume, external influences on the Estonian reform have clearly been more diverse. Direct involvement of the World Bank in the Estonian reform was rather limited. The Bank provided only one advisor on issues related to the third pillar and sponsored one national seminar on the second pillar, but these activities occurred after the government had adopted the Conceptual Framework of Pension Reform in 1997. In contrast to other CEECs, opting for multi-pillar reform – Hungary, Poland, and Latvia – there was no long-term co-operation project with the Bank on pensions or loan-seeking to restructure the pension system. As documented by Tavits (2003), this was a deliberate choice of the reform commission not to involve the Bank directly in elaboration of the reform package. As noted by Fultz (2006), the Bank’s limited visibility in Estonia is consistent with its growing caution concerning pension privatization (see Holzmann 1999; World Bank 2006). This growing caution was echoed by the IMF – another former supporter of radical pension privatization. This, of course, does not preclude possibilities for indirect influence. Toots (2003), investigating the role of international organizations in the shaping of the Estonian pension reform, assumes that it is the fit between international and national visions, which is important for applying a particular policy. Clearly, the agenda of pension reform in the whole region of Central and Eastern Europe was influenced by the World Bank. As Estonia was not among the first reformers, the regional learning aspect was present. This could also be seen as a second-step indirect influence of the World Bank: the Bank influenced reforms in Hungary, Poland, and Latvia, which in turn influenced the reform decision in Estonia. Obviously, it is a matter of speculation, whether Estonia had undertaken the reform, if there were no other similar reforms undertaken elsewhere; this is simply not possible to determine. However, the role of the World Bank shall not be overemphasized. The actual reform design in Estonia in crucial elements substantially departed from the policy advocated by the Bank. This alone indicates that the Bank was not the only source of information for Estonian policy designers. 131
By contrast, as demonstrated in this volume, the European Commission cited the need for the three-pillar model in several pre-accession reports. In its evaluations, the Commission clearly indicated support to the government’s multi-pillar reform plan. At the same time, frequent mentioning of pension reform among short-term priorities in pre-accession reports helped the government to sustain the reform time-table. To summarize, while the basic proposition of actor-centered institutionalism also holds in the Estonian case – in that it is the particular constellation of actors in a particular institutional setting that produced pension reform – the common caveat of this theoretical framework is to ascribe certain policy preferences to particular actors and overemphasize the relevance of international actors and certain contextual factors. 5.2.4. Ideational explanations In particular, the absence of a ‘critical juncture’ point, i.e. no immediate crises, leads to a search for alternative explanations of preference for a multi-pillar pension system in the Estonian case. As Andersen (2001) showed, radical reform does not require ‘real’ challenges, but these challenges could be constructed. Following Pfau-Effinger (2004), one could argue that an important cause of the reform was the increase of a basic contradiction – between the individualized, autonomous person who integrates into society through the labor market, and the construction of an elderly person solely dependent on the state. This developmental contradiction had (and still has) many facets in Estonia. From one side, there is a contradiction between demand and supply of public welfare, in particular (but not limited to) pensions. In this context, high preference for second pillar – reflected by the high number of switchers – conforms with the observation of Andersen (2001) in that when the level of current public welfare support is perceived insufficient (low replacement rate of state pensions) or its future uncertain, demand for private welfare will increase. There is also an intergenerational conflict – claims on the level of public pensions are higher than the willingness of the active population to pay for them. These ideational factors may not only explain the later popularity of the reform, but could also serve as explanations as to why the reform was undertaken in Estonia. The four explanatory factors identified by Bönker (2004) in studying the spread of multi-pillar systems – the exhaustion of the old paradigm, the compatibility of the new paradigm, the presence of role models and the support by major interest groups – seem all to apply to the Estonian case. Müller later (2003, 2006) recognized that the ideational setting can play an important role in enabling paradigm change in old-age security. However, she subscribes this only to the advocates of multi-pillar systems, not to the broader public. It appears to be a common caveat in the theoretical frameworks used to 132
explain pension reforms that people (including those who make the decision to switch to the new system, i.e. active participants in the reform) are regarded as objects rather than subjects of the reform. Andersen (2001) has correctly pointed out that the question of whether it is possible to maintain the old welfare state arrangement in face of new challenges should not be conflated with the question of whether the old arrangements are likely to survive. Whereas, at least in principle, it was possible to maintain the single-pillar pay-asyou-go pension system – obviously subject to some (possibly extensive) parametric reforms – this scenario was highly unlikely given the basic contradiction. However, it would be erroneous to regard ideational aspects as the sole explanatory factors behind the multi-pillar reform. Clearly, had the economic and demographic context factors or the pre-reform institutional setting been different, the set of policy alternatives would had been different. 5.3. CONCLUSIONS This study examined pension policy development and policy outcomes in Estonia over the period of 1990–2005 in the context of societal transition. Starting from the legacy of the Soviet pension system, transformation of the Estonian pension system has occurred in two major waves of paradigmatic changes. For the first wave of transformation in early 1990s, the critical independent variable appears to be rejection of the former Soviet pension system, which is by essence an ideational factor. Obviously, the turbulent circumstances that followed – economic decline coupled with rapid decline in employment – further conditioned the policy choices, in particular the adoption of flat rate benefit rules for the period of economic transition. Nevertheless, it would be erroneous to interpret the introduction of flat rate pensions as a deliberate shift towards egalitarian distribution principles. Rather, this was a temporary rescue measure, while the aim was to reintroduce earningsrelated pensions as soon as economic as well as administrative circumstances permit. Key actors in the first phase were single government politicians, while expert knowledge on pensions was scarce. While the first wave of reforms was motivated by a desire of government politicians to improve the level of social protection through separation from the Soviet pension system, in reality the real value and replacement rate of pensions declined, constituting a significant benefit retrenchment. The transitional arrangement applied in 1993–2000, in particular the macro-level defined-contribution design, whereby the level of pensions was determined by revenues from social tax, provided relative stability for the state pension system both in terms of benefit levels and financing. The period of 1993–2000 may be 133
regarded as a transitional period between two waves of transformation as the pension system remained within a single paradigm. At the same time, a number of important parametric reforms were initiated, including gradual increase of the pension age, linked to the change of economic and demographic context factors, in particular the increase of system dependency ratio. Payment of full pension to working pensioners from 1996 in essence constituted a redefinition of the social risk insured by the pension system – the old age pension became a subjective right of persons attaining pension age, rather than a benefit replacing lost earnings. The political desire to reduce egalitarism in the benefit distribution formula was reflected in the decline of the relative share of the flat rate element of old age pension. Key actors behind evolution of the pension policy in the period of 1993–1997 were officials of the Social Insurance Board and single coalition politicians. The key policy dilemmas of this period were linked to two major contradictions. From one side, there was a contradiction between the public demands to increase the replacement rate and the political agreement between leading parties not to raise the rate of social tax. From the other side, the public desire for a stronger link between benefits and previous earnings contradicted the need to keep as many as possible pensioners above the poverty level, because in circumstances of the closedbudget approach, higher differentiation of pensions would had entailed increasing the poverty rate of those at the lower end of the pension benefit scale. In brief, these contradictions could not be solved within the framework of existing (and partly politically imposed) constraints. For the second wave of transformation – introduction of the three-pillar pension system – ideational factors again appear crucial. Although the motivation for a multi-pillar reform was largely linked to the ageing of population, this future challenge can not explain the reform decision alone. An important ideational factor at play appears to be that the basic contradictions related to the former pension system could not be solved within the (partly self-imposed) constraints of the former paradigm. A key role in developing the outlines of the multi-pillar reform and co-ordinating its implementation was played by the Social Security Reform Commission, which was initially an expert commission, but after reorganization in 1999 combined politicians and experts. Considering the magnitude of the multi-pillar reform was implemented in different stages over the period of 1998–2002. It is remarkable that in spite of several changes in government coalitions over this period, successive governments proceeded with preparation and implementation of the multi-pillar reform, reflecting an underlying political consensus about the reform. Parametric reform of the first pillar, undertaken within the framework of the multipillar reform equalized the pension age of men and women and introduced the contribution-benefit link of state pensions, addressing thus simultaneously sustainability and distributional concerns related to the previous scheme. At the same time, the new 134
pension formula built upon the previous formula, reflecting a gradualist approach to changing distributional principles. Introduction of pension indexation, motivated by a desire to depoliticize the mechanism of pension increases, simultaneously marked an end to the former macro-level defined-contribution principle. Introduction of the pre-funded second pillar based on individual savings marks a paradigmatic shift in pension policy. Estonia was the fifth country in Central and Eastern Europe – after Hungary, Poland, Latvia, and Croatia – to implement a three-pillar pension system with a fully funded second pillar. While the regional learning aspect was present, it is argued here that the Estonian pension reform was a home-grown product, entailing a number of policy innovations. In particular, Estonia has been the only country in Central and Eastern Europe to increase the total contribution rate when introducing the second pillar, using both so-called “topup” and “carve-out” methods simultaneously. The study has also identified some other features, which make the Estonian pension policy design unique. In studies of pension reforms in Central and Eastern Europe, the introduction of mandatory fully funded pension scheme is often attributed to the influence of the World Bank. However, as indicated in the current study, the role of the World Bank in shaping the pension reform in Estonia was quite modest. By contrast, the role of the European Commission was more significant, as the implementation of a threepillar pension reform was cited as one of the economic conditions for accession to the European Union. Nevertheless, it would be erroneous to ascribe the adoption of the multi-pillar reform merely to the influence of international organizations. As demonstrated, the reform was internalized by main political forces. Finally, the high number of voluntary switchers to the second pillar – in spite of an additional individual contribution – indicates that the new pension paradigm broadly fitted with prevalent perceptions about the role of collective versus individual responsibility about old age security. The new paradigm does not entail a complete shift from collective responsibility to an individual one, but a shift in the balance between different policy principles and a shift in the public-private mix of pension provision. Nevertheless, even under the multi-pillar arrangement the role of the state pension scheme remains dominant. Against the background of numerous policy changes, the main policy outcome indicators have been relatively stable over the period of 1993–2005 (i.e. after the turbulent take-off of the Estonian pension system, which entailed a benefit retrenchment). The average net replacement rate has been in the range of 40–45%. The total expenditures of pension have been in the range of 6–7% f GDP, while the total number of pensioners has been around 375 thousand persons. This does not imply that the reforms have had minor significance. Vice versa, the relative stability of outcome indicators is largely due to the reforms, not in spite of them. Furthermore, the stability of macro-level indicators hides important distributional effects of reforms and implications on system sustainability. Deductive analysis of testing the Estonian case of pension system transformation against three theoretical frameworks – historical institutionalism, actor-centered institutionalism, and ideational approach – indicates that all three frameworks have some explanatory power for understanding the choices of pension policy and 135
pension reform process in Estonia. At the same time, all three theoretical frameworks have some limitations. The combination of these three frameworks – ideational approach combined with historical and actor-centered institutionalism – appears to provide the best understanding of the transformation process. In particular, I claim that paradigmatic pension reforms not only entail a paradigm shift, but these reforms are largely (although not solely) conditioned by an underlying paradigmatic value shift. In a way, it appears almost as stating the obvious that in explaining paradigmatic pension reforms the ideational explanatory model has an important role. Obviously, the year 2005 does not mark an end of the pension reform process in Estonia. As economic, demographic and social circumstances or political and public preferences change, further adjustments in the pension system may be expected. Whether these future changes will be parametric or paradigmatic, will depend on a combination of factors, including both objective and ideational.
136
EESTI PENSIONISÜSTEEMI TRANSFORMATSIOON: POLIITIKA VALIKUD JA TULEMID Kokkuvõte Käesolev väitekiri vaatleb Eesti pensionisüsteemi arenguid aastatel 1990–2005. Pensionisüsteemid on paljudes riikides – ka Eestis – suurim sotsiaalpoliitiliste siirete süsteem. Pensionid mõjutavad tuluallikana suure osa elanike heaolu. Pensionisüsteemi finantseerimises osaleb sisuliselt kogu majanduslikult aktiivne elanikkond. Nii praktikas kui ka analüüsi seisukohalt muudab pensionisüsteemi üheks keerulisemaks sotsiaalpoliitiliseks süsteemiks asjaolu, et selles põimuvad väga erinevad sotsiaalsed, demograafilised, majanduslikud, õiguslikud ja poliitilised küsimused. Käesoleva töö eesmärkideks on: - analüüsida Eesti pensionipoliitika muutusi ajateljel 1990–2005, neid muutusi mõjutanud tegureid, reformi kujundamisel osalenud osapooli ning poliitika tulemeid; - kontseptualiseerida pensionisüsteemi transformatsiooni protsessi; - pakkuda seletusmudeleid kesksetele poliitika valikutele, sh otsusele rakendada kolmesambalist pensionisüsteemi. Uurimisstrateegialt on käesoleva väitekirja puhul tegemist selgitava juhtumiuurimusega (Yin 1994), juhtumiülese võrdleva analüüsi (Huberman, Miles 1998) elementidega. Juhtumiuurimused ja riikidevahelised võrdlused on heaoluriigi transformatsiooni, sh pensionireformide analüüsimisel domineerivad uurimisstrateegiad. Arvestades, et viimastel aastatel on ilmunud mitmeid Kesk- ja Ida-Euroopa riikide pensionireforme käsitlevaid võrdlevaid uurimusi (s.h Müller 1999; Fultz 2002a, 2002b, 2006; Schmähl, Horstmann 2002; Chłon-Dominczak, Mora 2003; Schmähl 2003; Casey 2004; Grimmeisen 2004; Whitehouse 2004; Schubert 2005), kasutab töö autor oma suhtelist eelist ning keskendub Eesti juhtumi sügavamale analüüsile. Tulenevalt juhtumiuurimuse strateegiast, analüüsitakse juhtumit – Eesti pensionisüsteemi transformatsiooni – tema reaalses kontekstis, s.t samaaegsete majanduslike, sotsiaalsete, poliitiliste ja demograafiliste muutuste kontekstis. Sellest tuleneb uurimuse multi-dimensionaalne lähenemine. Samas iseloomustab uurimust n-ö mitmetsentriline lähenemine (vt Siegel 2003), kus pensionisüsteemi transformatsiooni analüüsimisel vaadeldakse ühelt poolt pensionisüsteemi parameetrites toimunud muutusi ning teisalt poliitika tulemeid iseloomustavaid keskseid indikaatoreid. Juhtumiuurimuse keskseks analüüsiühikuks on pensionipoliitika, hõlmates nii poliitika valikuid kui poliitika järelmeid. Teisisõnu, analüüsitakse pensionipoliitika kujundamisel tehtud sotsiaalpoliitilisi valikuid, vaadeldes samas ka nende valikute mõju.
137
Uurimuse alusteesiks, mis suunab uurimisküsimuste analüüsi, on varasemates pensionireforme puudutavates uurimustes (nt Aidukaite 2004) kinnitust leidnud väide, et pensionisüsteemi transformatsiooni protsessi mõjutavad mitmed erinevad tegurid (sh majanduslikud, demograafilised, sotsiaalsed, kultuurilised, poliitilised ja ajaloolised) ja osapooled nii siseriiklikult kui rahvusvaheliselt. Töö teoreetilises osas esitatakse esmalt kontseptuaalne raamistik pensionisüsteemide ja -reformide kirjeldamiseks. Seejuures käsitletakse pensionisüsteeme kirjeldavaid keskseid kontsepte, sh pensioniskeemide põhiparameetreid e. pensionipoliitika keskseid instrumente ja pensioniskeemide põhitüüpe, ning pensionireforme kirjeldavaid põhikontsepte. Viimasel juhul tuginetakse muu hulgas Halli (1993) poliitika muutuste kontseptualiseeringule ning sellega seotud pensionireformide liigitamisele parameetrilisteks ja paradigmaatilisteks (vt nt Holzmann et al 2003). Pensionireformide kontseptualiseerimisele on olulise panuse andnud ka Ferrera et al (2000), kes võtsid kasutusele reeksperimenteerimise ja rekalibreerimise mõisted, iseloomustamaks vastavalt reformi protsessi ja reformide suunda. Teooriapeatükis esitatakse ka ulatuslik kirjandusülevaade kolmest teoreetilisest raamistikust, mida on viimastel aastatel enim kasutatud pensionireformide (sh eriti Kesk- ja Ida-Euroopa riikide reformide) teoreetilisel seletamisel – need on ajalooline institutsionalism, tegelaskeskne institutsionalism ja ideeline seletusmudel. Eesti pensionisüsteemi transformatsiooni analüüsimisel kasutatakse töös kesksete meetoditena pensionipoliitika protsessi ajaloolist rekonstruktsiooni kombineeritult pensionipoliitika tulemite kirjeldava statistikaga. Pensionisüsteemi reformimisel tehtud poliitikavalikute analüütilisel kirjeldamisel vaadeldakse kesksete parameetritena pensionide kvalifikatsioonitingimusi, pensionide arvutamise printsiipe, pensionisüsteemi finantseerimis- ja institutsionaalset korraldust. Uurimismaterjaliks on seejuures pensionipoliitikat mõjutanud õigusaktid (üle 50 seaduse koos mitmete rakendusaktidega), riigieelarved ja nende täitmise aruanded. Pensionipoliitika protsessi rekonstruktsioonis pööratakse muu hulgas tähelepanu poliitika kujundamises osalenud osapoolte – nii siseriiklike kui protsessi mõjutanud rahvusvaheliste organisatsioonide – seisukohtadele. Vastava analüüsi allikateks on Vabariigi Valitsuse poolt heaks kiidetud kontseptsioonid, koalitsioonilepingud, sotsiaalkindlustusreformi komisjoni töömaterjalid, aga ka muud kirjalikud materjalid, sh Euroopa Liidu Komisjoni aruanded, IMFi pressiteated jms. Töö autor oli aastatel 1997–2001 Vabariigi Valitsuse sotsiaalkindlustusreformi komisjoni liige ning seega nimetatud perioodil ise vahetult seotud pensionipoliitika kujundamisega. Pensionipoliitika tulemite analüüsis kasutatakse kirjeldavat statistikat. Põhiindikaatoriteks on seejuures pensionäride ja pensionikindlustatud isikute arv, süsteemisõltuvusmäär, pensionisüsteemi tulud ja kulud, pensionide nominaal- ja reaalväärtused ning asendusmäär. Makroindikaatorite puhul kasutatakse peamiselt Statistikaameti, Sotsiaalkindlustusameti ja Sotsiaalministeeriumi agregaatandmete sekundaaranalüüsi. Poliitika tulemite mõju analüüs mikrotasandil tugineb pensionisaajate ja pensionikindlustatud isikute individuaalandmetele administratiivsetest andmebaasidest – Sotsiaalkindlustusameti riikliku pensionikindlustuse registri, Eesti Väärtpaberite Keskregistri ja Finantsinspektsiooni andmetele. Indikaatoritena kasutatakse seejuures erinevate pensioniskeemide hõlmatuse määrasid ja individuaalseid asendusmäärasid. 138
Eesti pensionisüsteem on kujunenud läbi kahe transformatsioonilaine. Esimesed reformid aastatel 1990–1992 eraldasid Eesti pensionisüsteemi Nõukogude pensionikorraldusest. Sellele järgnes 1993–1999 üleminekuperioodi pensionikorraldus stabiilse finantseerimissüsteemi ja pensionide maksmise reeglitega. Üleminekuperioodi lahendus ei oleks siiski olnud pikemaajaliselt jätkusuutlik, arvestades töötajate ja pensionäride suhtarvu halvenemist. Teine transformatsioonilaine sai alguse 1997. aasta kontseptsioonist, mis seadis eesmärgiks kolmesambalise pensionisüsteemi rakendamise. Uus kolme samba süsteem rakendus etapiliselt aastatel 1998–2002. Paljude poliitikamuutuste taustal on kesksed poliitika tulemeid kirjeldavad indikaatorid püsinud perioodil 1993–2005 suhteliselt stabiilsed. Keskmine netoasendusmäär on püsinud vahemikus 40–45%, pensionikulude osakaal SKP-s vahemikus 6– 7%. Makroindikaatorite suhteline stabiilsus ei peegelda siiski reformide vähest tähtsust, vaid vastupidi, on suures osas mitmete läbiviidud reformide tulemus. Riikliku pensioni saajate arvu stabiliseerumine 375 000–380 000 inimese tasemel on suuresti pensioniea tõstmise tagajärg. Samuti on reformidel olnud oluline roll pensionisüsteemi jaotusmõju ja pensionisüsteemi jätkusuutlikkuse suhtes. Empiirilise materjali analüüsile tuginedes testitakse töö viimases osas Eesti pensionisüsteemi transformatsiooni juhtumit kolme teoreetilise raamistiku – ajaloolise institutsionalismi, tegelaskeskse institutsionalismi ja ideelise lähenemisviisi – poolt väljapakutud seletusmudelite suhtes. Deduktiivne analüüs näitab, et kuigi igal eelnimetatud seletusmudelil on teatav seletusjõud Eesti pensionisüsteemi transformatsiooni selgitamisel, ei suuda ajaloolise ja tegelaskeskse institutsionalismi seletusmudelid eraldiseisvatena adekvaatselt selgitada, miks valiti Eestis paradigmaatiline pensionireform üleminekuga mitmesambalisele pensionisüsteemile. Kokkuvõttes, töös esitatakse tervikkäsitlus Eesti pensionisüsteemi muutustest perioodil 1990–2005, kontseptualiseeritakse Eesti pensionisüsteemi transformatsiooni protsessi ning näidatakse Eesti pensionireformi disaini ainulaadseid jooni. Teoreetilises aspektis täiendatakse kontseptuaalset raamistikku pensionisüsteemide kirjeldamiseks, kontranäidete kaudu näidatakse ajaloolise ja tegelaskeskse institutsionalismi teooriate piiratust paradigmaatiliste pensionireformide seletamisel ning väidetakse, et paradigmaatiliste pensionireformide selgitamisel on ajaloolise ja tegelaskeskse institutsionalismi seletusmudeli kõrval oluline roll ideelisel seletusmudelil.
139
REFERENCES AIDUKAITE, J. (2004). The Emergence of the Post-Socialist Welfare State – the Case of the Baltic States: Estonia, Latvia and Lithuania. Södertörn Doctoral Dissertations: 1. Stockholm. AMENTA, E. (2003). What we know about the development of social policy: comparative and historical research in comparative and historical perspective. – J. Mahoney and D. Rueschemeyer (Eds.) Comparative Historical Analysis in the Social Sciences, Cambridge: Cambridge University Press, 91–130. ANDERSEN, J.G. (2001). Change without Challenge? Welfare States, Social Construction of Challenge and Dynamics of Path Dependency. – J. Clasen (Ed.) What Future for Social Security? Debates and Reforms in National and Cross-national Perspectives. The Hague: Kluwer Law International, 121–138. BARR, N. (1998). The Economics of the Welfare State, 3rd ed., New York: Oxford University Press. BARR, N. (2000). Reforming Pensions: Myths, Truths and Policy Choices. IMF Working Paper WP/00/139. BÉLAND, D. (2005). Ideas and Social Policy: An Institutionalist Perspective. – Social Policy and Administration. Vol. 39, No. 1, 1–18. BONOLI, G. (1997). Classifying Welfare States: a Two-Dimension Approach. – Journal of Social Policy. Vol. 26, 351–372. BRYMAN, A. (2001). Social Research Methods. New York: Oxford University Press. BÖNKER, F. (2004). Accounting for Differences in the Spread of the Multipillar Paradigm: A Comparative Analysis of Pension Reform in Five EU Social Insurance Countries. Paper presented at the WRAMSOC-Workshop, Berlin, 23–25 April 2004. BÖNKER, F., MÜLLER, K., PICKEL, A. (Eds.) (2002). Postcommunist Transformation and the Social Sciences: Cross-Disciplinary Approaches. Lanham: Rowman & Littlefield. BÖRSCH-SUPAN, A. (2003). What are NDC Pension Systems? What Do They Bring to Reform Strategies? Paper for the World Bank – RFV Conference on NDC Pensions. Sandhamn, Sweden, 29–30 September 2003. CALDWELL, B. (2003). A Strategic View of Efforts to Lead the Transformation of Schools. Paper presented at the 6th World Convention of the International Confederation of Principals, Edinburgh, 13–16 July 2003. CASEY, B.H. (2004). Pension Reform in the Baltic States: Convergence with 'Europe' or with 'the World'? – International Social Security Review. Vol. 57, No. 1, 19–45. CERAMI, A. (2005). Social Policy in the Central and Eastern Europe: an Emergence of a New European Model of Solidarity? Doctoral Dissertation. Erfurt University. CHŁON, A., GÓRA, M., RUTKOWSKI, M. (1999). Shaping Pension Reform in Poland: Security Through Diversity. World Bank Social Protection Discussion Paper Series No. 9923. Washington DC. CHŁOŃ-DOMIŃCZAK, A. (2002). The Polish pension reform of 1999. – E. Fultz (Ed.) Pension Reform in Central and Eastern Europe: Volume 1. Restructuring with Privatization: Case Studies of Hungary and Poland. Budapest: ILO-CEET, 95–205.
140
CHŁON-DOMINCZAK, A., MORA, M. (2003). Commitment and Consensus in Pension Reform – R. Holzmann, M. Orenstein and M. Rutkowski (Eds.) Pension Reform in Europe: Process and Progress. Washington, D.C.: The World Bank, 131–155. CLASEN, J. (2005). Reforming European Welfare States: Germany and the United Kingdom Compared. Oxford–New York: Oxford University Press. CLASEN, J., SIEGEL, N. (2005). Exploring the Dynamics of Reform: the Dependent Variable Problem in Comparative Welfare State Analysis. University of Stirling. CREMER, H., PESTIEAU, P. (2003). Social Insurance Competition Between Bismarck and Beveridge. IDEI Working Paper No 141. DEACON, B., HULSE, M. (1997). The Making of Post-Communist Social Policy: The Role of International Agencies. – Journal of Social Policy. Vol.26, No.1, 43–62. DEEG, R. (2001). Institutional Change and the Uses and Limits of Path Dependency: The Case of German Finance. Max-Planck-Institut für Gesellschaftsforschung Discussion Paper 01/6. DE LA PORTE, C. (2000). Which Social Policy Ethos for Estonia? Paper for the “Ethos of Welfare: Metamorphoses and Variations of Governmentality” symposium, 31 August – 2 September 2000. DENZIN, N.K. (1970). Research Art in Sociology. Chicago: Aldine. DRAZEN, A., GRILLI, V. (1993). The Benefit of Crises for Economic Reforms. – American Economic Review, 83(3), 598–607. EAMETS, R. (2001). Reallocation of Labor During Transition. Disequilibrium and Policy Issues: The Case of Estonia. Dissertationes Rerum Oeconomicarum Universitatis Tartuensis. Tartu. EAMETS, R. (2002). Eesti makromajanduslik areng. [Macroeconomic development of Estonia]. – A. Arrak (Ed.) Majanduse ABC [ABC of Economy]. Tallinn: Audentes, 305–320. EESTI VABARIIK (2005). Uuendatud konvergentsiprogramm 2005. [Renewed Convergence Program 2005]. Tallinn. ESPING-ANDERSEN, G. (1990). The Three Worlds of Welfare Capitalism. Cambridge: Polity Press. ESPING-ANDERSEN, G. (2002). Towards the Good Society, Once Again? – G. EspingAndersen, D. Gallie, A. Hemerijk and J. Myles. Why We Need a New Welfare State. Oxford: Oxford University Press, 1–25. ESTONIAN STATISTICAL OFFICE (1998). Social Trends. Tallinn. ESTONIAN STATISTICAL OFFICE (2001). Social Trends 2. Tallinn. ESTONIAN STATISTICAL OFFICE (2003). Population 2002. Tallinn. ESTONIAN STATISTICAL OFFICE (2004). Social Trends 3. Tallinn. ES TURU-UURINGUTE AS (2002). Ülevaade kohustusliku kogumispensioni teemalisest avaliku arvamuse uuringust. [Overview of the public opinion poll on compulsory funded pensions]. Tallinn. http://www.sm.ee/est/HtmlPages/pension/$file/pension.pdf EUROPEAN COMMISSION (1997). Agenda 2000 - Commission Opinion on Estonia’s Application for Membership of the European Union. Brussels. EUROPEAN COMMISSION (2000). 2000 Regular Report from the Commission on Estonia’s Progress Towards Accession. Brussels.
141
EUROPEAN COMMISSION (2001). 2001 Regular Report from the Commission on Estonia’s Progress Towards Accession. Brussels. FERRERA, M., HEMERIJK, A., RHODES, M. (2000). The Future of Social Europe: Recasting Work and Welfare in the New Economy. Oeiras: Celta. FINANCIAL ADVISORY LTD. (2003). Investment of pension funds in CEE countries: Research Report. Budapest. FOX, L., PALMER, E. (2001). New Approaches to Multipillar Pension Systems: What in the World is Going On? – R. Holzmann and J.E. Stiglitz (Eds.) New Ideas about Old Age Security. Washington DC: The World Bank, 90–132. FULTZ, E. (Ed.) (2002a). Pension Reform in Central and Eastern Europe - Volume 1. Restructuring with Privatization: Case Studies of Hungary and Poland. Budapest: ILO-CEET. FULTZ, E. (Ed.) (2002b). Pension Reform in Central and Eastern Europe - Volume 1. Restructuring of Public Pension Schemes: Case Studies of the Czech Republic and Slovenia. Budapest: ILO-CEET. FULTZ, E. (2006). Pension Reform in the Baltics: Expectations and Early Experience. – E. Fultz (Ed.) Pension Reform in the Baltic States: Estonia, Latvia, Lithuania. Budapest: ILO, 353–377. FUNKE, M., STRULIK, H. (2003). Taxation, Growth and Welfare: Dynamic Effects of Estonia’s 2000 Income Tax Act. BOFIT Discussion Papers No.10. Helsinki: Bank of Finland Institute for Economies in Transition. GESELLSHAFT FÜR VERSICHERUNGSVISSENSHAFT UND –GESTALTUNG e.V. (2003). Social Protection in the Candidate Countries: Country Studies Estonia, Latvia, Lithuania. Berlin: AKA. GILBERT, N. (2002). Transformation of the Welfare State: The Silent Surrender of Public Responsibility. New York: Oxford University Press. GILLION, C., TURNER, J., BAILEY, C., LATULIPPE, D. (Eds.) (2000). Social Security Pensions: Development and Reform. Geneva: ILO. GOERTZ, G., STARR, H. (2003). Necessary conditions: theory, methodology and applications. Lanham-Boulder-New York-Oxford: Rowman & Littlefield Publishers. GREEN-PEDERSEN, C. (2004). The Dependent Variable Problem within the Study of Welfare-State Retrenchment: Defining the problem and looking for solutions. – Journal of Comparative Policy Analysis, Vol. 6, No. 1, 3–14. GRIMMEISEN, S. (2004). Path Dependence and Path Departure: Analysing the first decade of post-communist pension policy in Hungary, Poland and the Czech Republic. ZeS Working Paper No.1/2004. GUARDIANCICH, I. (2004). Welfare State Retrenchment in Central and Eastern Europe: the Case of Pension Reforms in Poland and Slovenia. – Managing Global Transitions, No. 2(1), 41–64. HALL, P. (1992). The movement from Keynesianism to monetarism: Institutional analysis and British economic policy in the 1970s. – S. Steinmo, K. Thelen and F. Longstreth (Eds.) Structuring Politics: Historical Institutionalism in Comparative Analysis. New York: Cambridge University Press, 90–113. HALL, P. (1993). Policy Paradigms, Social Learning, and the State. Comparative Politics, 25, 275–296.
142
HANSAPANK (2004). Pensioni ABC: juhend elu õnnelikumaks muutmisel. [ABC of pension: how to make one’s life happier]. Tallinn. HANSSON, A. (1997). Stabilization in the Baltic States. – M. Blejer and M. Škreb (Eds.) Macroeconomic Stabilization in Transition Economies, Cambridge University Press, 256–280. HEMERIJK, A. (2002). The Self-Transformation of the European Social Model(s) – G. Esping-Andersen, D. Gallie, A. Hemerijk and J. Myles. Why We Need a New Welfare State. Oxford: Oxford University Press, 173–213. HINRICHS, K. (2003). The Politics of Pension Reform in Germany. Paper presented at the Conference “Pension Reform in Europe: Shared Problems, Sharing Solutions?”, London School of Economics, Hellenic Observatory/The European Institute, London, 5 December 2003. HOLZMANN, R. (1999). The World Bank Approach to Pension Reform. Social Protection Discussion Paper Series No. 9807. Washington DC: The World Bank. HOLZMANN, R., PALACIOS, R. (2001). Individual Accounts as Social Insurance: A World Bank Perspective. Social Protection Discussion Paper Series No. 0114. Washington DC: The World Bank. HOLZMANN, R., MACKELLAR, L., RUTKOWSKI, M. (2003). Accelerating the European Pension Reform Agenda: Need, Progress, and Conceptual Underpinnings. – R. Holzmann, M. Orenstein and M. Rutkowski (Eds.) Pension Reform in Europe: Process and Progress. Washington, D.C.: The World Bank, 1–45. HUBER, E., STEPHENS, J.D. (2001). Development and Crisis of the Welfare State: Parties and Policies in Global Markets. Chicago: University of Chicago Press. HUBERMAN, M.A., MILES, M.B. (1998). Data Management and Analysis Methods. – N.K. Denzin and Y.S. Lincoln (Eds.) Collecting and Interpreting Qualitative Materials Thousand Oaks-London-New Delhi: SAGE Publications, 179–210. IMMERGUT, E.M. (1998). The Theoretical Core of the New Institutionalism. – Politics and Society, 26: 5–34. IYER, S. (1999). Actuarial Mathematics of Social Security Pensions. Geneva: ILO. JOCHEM, C. (2005). Contemporary Welfare State Reforms: Beyond Institutional Explanations? Paper presented at the Workshop “Exploring the Dynamics of Reform: The Dependent Variable Problem in Comparative Welfare State Analysis”, Stirling, 13–14 May 2005. KATUS, K. (1998). Population Developments. – Estonian Statistical Office. Social Trends. Tallinn. KATUS, K., PUUR, A., PÕLDMA, A., SAKKEUS, L. (1999). Rahvastikuvananemine Eestis. [Population ageing in Estonia]. Tallinn: EKDK. KATUS, K., PUUR, A., PÕLDMA, A. (2002). Eesti põlvkondlik rahvastikuareng. [Cohort Population Development in Estonia]. Tallinn: EKDK. KATUS, K., PUUR, A., PÕLDMA, A. (2004). Population-related Policies in Estonia in the 20th century: Stages and Turning Points – Yearbook of Population Research in Finland 40. Helsinki: The Population Research Institute, 73–103. KINGDON, J.W. (1984). Agendas, Alternatives and Public Policies. New York: Harper Collins.
143
KIRCH, M. (2000). Pensionireformist läbi avaliku arvamuse prisma: valikuid piirab nii info kui raha puudus. [On Pension Reform through a Public Opinion Prism: Choices are Restricted by Lack of Information and Money]. – Riigikogu Toimetised [Journal of Estonian Parliament] No. 2, 194–195. KITSCHELT, H. (2001). Partisan Competition and Welfare State Retrenchment - When do Politicians Choose Unpopular Policies? – P. Pierson (Ed.), The New Politics of the Welfare State. Oxford: Oxford University Press, 265–302. KORHONEN, I. (2003). Some aspects of EU membership on Baltic monetary and exchange rate policies. – V. Pettai and J. Zielonka (Eds.) The road to the European Union. Volume 2: Estonia, Latvia and Lithuania. Manchester University Press, 255–281. KUDDO, A., LEETMAA, R., LEPPIK, L., LUUK, M., VÕRK, A. (2002). Sotsiaaltoetuste efektiivsus ja mõju tööjõupakkumisele. [Social benefits in Estonia: efficiency and impact on work incentives]. Tallinn: PRAXIS. KULU, L., REILJAN, J. (2004). Old-age pension reform in Estonia on the bases of the World Bank’s multi-pillar approach. Tartu: Tartu University. KUTSAR, D., TRUMM, A. (Eds.) (1999). Poverty Reduction in Estonia: Background and Guidelines. Tartu. LEETMAA, R., VÕRK, A., KALLASTE, E. (2004). Vanemaealine tööjõud tööturul ja tööelus. [Older workforce in labor market and employment]. Tallinn: PRAXIS. http://www.praxis.ee/data/Eakad_final_15_06_2004.pdf LEPPIK, L. (1998). Eesti pensionireform: kolm sammast [Estonian pension reform: three pillars]. Talllinn: Jaan Tõnissoni Instituut. LEPPIK, L. (2002). Disability protection in Estonia. – E. Fultz and M. Ruck (Eds.) Reforming Worker Protections: Disability Pensions in Transformation. Budapest: ILOCEET, 93–145. LEPPIK, L. (2003). Social protection and EU enlargement: the case of Estonia. – V. Pettai and J. Zielonka (Eds.) The road to the European Union. Volume 2: Estonia, Latvia and Lithuania. Manchester University Press, 140–162. LEPPIK, L. (2004). Success factors of the pension reform. Paper presented at the international conference “Economic reforms for Europe: growth opportunities in enlarged EU”, Bratislava 18 March 2004. http://www.ineko.sk/euroreform/menu4_leppik2.pdf LEPPIK, L., MÄNNIK, G. (2002). Transformation of old-age security in Estonia. – W. Schmähl and S. Horstmann (Eds.) Transformation of pension systems in Central and Eastern Europe. Cheltenham – Northampton: Edward Elgar Publishing, 89–124. LEPPIK, L., KRUUDA, R. (2003). Study on the Social Protection Systems in the 13 Applicant Countries: Estonia Country Study. – Gesellshaft für Versicherungsvissenshaft und –gestaltung e.V., Social Protection in the Candidate Countries: Country Studies Estonia, Latvia, Lithuania. Berlin: AKA, 3–140. LEPPIK, L., TIIT, E.-M., VÕRK, A. (2004). “Eesti pensionisüsteem Euroopa Liidu ühiste pensionieesmärkide valguses”. [Estonian pension system in the light of the European Union common pension objectives]. Riigikogu Toimetised [Journal of the Estonian parliament] No.9, 81–90. LEPPIK, L., VÕRK, A. (2006). Pension Reform in Estonia. – E. Fultz (Ed.) Pension Reform in the Baltic States: Estonia, Latvia, Lithuania. Budapest: ILO, 17–141.
144
LINDELL, C. (2001). The Estonian pension system. – L. Leppik and C. Lindell. Neighbours look at each other: Estonian and Finnish Pension Systems. The Central Pension Security Institute Reports 2001:25. Helsinki, 5–45. LINDEMAN, D. (2004). Review of Recent Pension Reforms in the Baltic Region. – Organisation for Economic Co-operation and Development. Pension Reform in the Baltic Countries. Private Pension Series No. 5. Paris: OECD, 7–24. MADRID, R.L. (2003). Retiring the State: The Politics of Pension Privatization in Latin America and Beyond. Stanford: Stanford University Press. MASSO, J., STAEHR, K. (2005). Inflation dynamics and nominal adjustment in the Baltic States. Tartu: University of Tartu. MINISTRY OF SOCIAL AFFAIRS OF ESTONIA (2005). Social sector in figures. Tallinn. MOUTON, P. (1998). State of the Reforms in Central and Eastern Europe. – International Social Security Association. Evaluating Social Security Reforms. Social Security Documentation European Series No.26. Geneva: ISSA, 59–82. MUSGRAVE, R. (1986). Public Finance in a Democratic Society. New York: New York University Press. MÜLLER, K. (1999). The Political Economy of Pension Reform in Central-Eastern Europe. Cheltenham–Northampton: Edward Elgar Publishing. MÜLLER, K. (2002). Between State and Market: Czech and Slovene Pension Reform in Comparison. – E. Fultz (Ed.). Pension Reform in Central and Eastern Europe – Volume 2. Restructuring of Public Pension Schemes: Case Studies of the Czech Republic and Slovenia. Budapest: ILO-CEET, 113–146. MÜLLER, K. (2003). The Making of Pension Privatization in Latin America and Eastern Europe. – R. Holzmann, M. Orenstein and M. Rutkowski (Eds.) Pension Reform in Europe: Process and Progress. Washington, D.C.: The World Bank, 47–78. MÜLLER, K. (2006). The Political Economy of Pension Privatization in the Baltics. – E. Fultz (Ed.) Pension Reform in the Baltic States: Estonia, Latvia, Lithuania. Budapest: ILO, 397–420. MÜLLER, K., RYLL, A., WAGENER, H.-J. (Eds.) (1999). Transformation of Social Security: Pensions in Central-Eastern Europe. Heidelberg: Physica-Verlag. MYLES, J., PIERSON, P. (2001). The Comparative Political Economy of Pension Reform. – Pierson, P. (Ed.) The New Politics of the Welfare State. Oxford: Oxford University Press. MYLES, J. (2002). A New Social Contract for the Elderly? – G. Esping-Andersen, D. Gallie, A. Hemerijk and J. Myles. Why We Need a New Welfare State. Oxford: Oxford University Press, 130–172. NATALI, D. (2001). The transformation of the European Welfare State: pension reforms in France and Italy. Paper presented at the 13th Annual Meeting on Socio-Economics, Amsterdam, 28 June – 1 July 2001, http://www.sase.org/conf2001/papers/natali_david.pdf NICKLESS, J. (2002). Trends and developments in old age pension and health-care financing in Europe during the 1990s. Trends in social cohesion, No.2. Council of Europe Publishing. OORN, K. (2004). Developments in the Estonian Pension System. – Organisation for Economic Co-operation and Development. Pension Reform in the Baltic Countries. Private Pension Series No. 5. Paris: OECD, 59–85.
145
ORENSTEIN, M. (2000). How Politics and Institutions Affect Pension Reform in Three Postcommunist Countries. World Bank Policy Research Working Paper No.2310. ORENSTEIN, M. (2006). Transnational Politics and the New Pension Reforms. Syracuse. (book manuscript). http://www.maxwell.syr.edu/psc/faculty/New%20Pension%20Reform%20CORNELL%20S UBMISSION.pdf ORENSTEIN, M., HAAS, M. (2002). Globalization and the Development of Welfare States in Postcommunist Europe. BCSIA Discussion Paper 2002–02, Kennedy School of Government, Harvard University. ORSZAG, P.R., STIGLITZ, J.E. (2001). Rethinking Pension Reform: Ten Myths about Social Security Systems. – R. Holzmann and J.E. Stiglitz (Eds.) New Ideas about Old Age Security. Washington DC: The World Bank, 17–56. PALIER, B. (2000). Beyond retrenchment: Differences in sequence and in patterns of welfare state changes. Paper presented at the ISSA International Research Conference on Social Security “Social security in the global village”, Helsinki, 25–27 September 2000. PALIER, B. (2001). Comparing Social Insurance Reforms in Continental Europe: Analytical Framework and Hypotheses. Paris: CNRS. PALMER, E. (2000). The Swedish Pension Reform Model: Framework and Issues. Social Protection Discussion Paper Series No. 0012. Washington, DC: The World Bank. PATTON, C.V., SAWICKI, D. S. (1986). Basic Methods of Policy Analysis and Planning. Englewood Cliffs: Prentice-Hall. PFAU-EFFINGER, B. (2004). Culture and Path Dependency of Welfare State Development. Paper presented at the ESPAnet conference 2004 “European Social Policy: Meeting the Needs of a New Europe”, University of Oxford, 9–11 September 2004. PFAU-EFFINGER, B. (2005). Culture and Welfare State Policies: Reflections on a Complex Interrelation. – Journal of Social Policy. Vol. 34, No. 1, 3–20 PIERSON, P. (1996). The New Politics of the Welfare State. – World Politics, No. 48, 143–179. PIERSON, P. (2000). Not Just What, but When: Timing and Sequence in Political Processes. – Studies in American Political Development. Vol. 14, 72–92. PIETERS, D. (1993). Introduction into the Basic Principles of Social Security. Deventer/Boston: Kluwer Law & Taxation. PUUR, A. (2000). Change in Economic Status of Older Population: the Case of Estonia during the 1990s. RU Series B, No.42. Tallinn: EKDK. PÕIKLIK, P. (2005). Valitsuse reformikommunikatsiooni edukuse tegurid. [Success factors of governments’ reform communication]. Master thesis. Tartu: Tartu Ülikool. PÕLD, P. (2002). The fee structure of pension products: the Estonian experience. Presentation at the INPRS Regional Meeting for the Eastern and Central European Countries, Tallinn, 7–8 February 2002. https://www.oecd.org/dataoecd/61/43/2067381.pdf PÜSS, T. (1995). Reform of the Social Protection System. – U. Venesaar and G.A. Hachey (Eds.) Economic and Social Changes in the Baltic States in 1992-1994. Tallinn: Estonian Acaemy of Sciences Institute of Economics, 33–45. RIBHEGGE, H. (1999). The Controversy Between the Pay-As-You-Go System and the Fully-Funded System in Old-Age Security. – K. Müller, A. Ryll and H.-J. Wagener (Eds.) Transformation of Social Security: Pensions in Central-Eastern Europe. Heidelberg: Physica-Verlag, 61–77.
146
ROLAND, G. (2000). Transition and Economics. Cambridge, MA: MIT Press. SCHARPF, F. (1997). Games Real Actors Play: Actor-Centered Institutionalism in Policy Research. Boulder, CO/Oxford: Westview Press. SCHIFF, J., HOBDARI, N., SCHIMMELPFENNIG, A., ZYTEK, R. (2001). Pension Reform in the Baltics: Issues and Prospects. IMF Occasional Paper No. 200 SCHMIDT, M.G. (1996). When Parties Matter: A Review of the Possibilities and Limits of Partisan Influence on Public Policy. – European Journal of Political Research, 30, 155–183. SCHMÄHL, W. (2002). Introduction: Design of the analysis, its methodological approach and basic decisions in designing pension schemes. – W. Schmähl, S. Horstmann (Eds.). Transformation of Pension Systems in central and Eastern Europe. Cheltenham – Northampton: Edward Elgar Publishing, 3–24. SCHMÄHL, W. (2003). Pensions. – European Commission, Social Protection in the 13 Candidate Countries: a Comparative Analysis. Luxembourg: Office for Official Publications of the European Communities, 31–69. SCHMÄHL, W., HORSTMANN, S. (Eds.) (2002). Transformation of Pension Systems in Central and Eastern Europe. Cheltenham – Northampton: Edward Elgar Publishing. SCHUBERT, M. (2005). Reformen der Alterssicherung in Mittel- und Osteuropa Erklärungsgrößen für Reformentscheidungen. Berlin: Logos. SIEGEL, N.A. (2003). Worlds of Comparative Welfare State Analysis: Some Suggestions for Potential Bridge-Over Strategies. Paper presented at the ESPAnet Annual Conference, Kopenhagen, 13–15 November 2003. SILLASTE, J. (1998). Sotsiaaltrendide üldmajanduslik ja riigiehituslik taust [The general economic and statehood-building background of social trends]. – Estonian Statistical Office. Sotsiaaltrendid [Social Trends]. Tallinn, 91–93. STONE, D. (2001). Learning Lessons, Policy Transfer and the International Diffusion of Policy Ideas. CSGR Working Paper No. 69/01. TAVITS, M. (2003). Policy Learning and Uncertainty: The Case of Pension Reform in Estonia and Latvia. – The Policy Studies Journal. Vol.31, No.4, 643–660. THOMPSON, L. (1998). Older and Wiser: The Economics of Public Pensions. Washington DC: The Urban Institute Press. TIIT, E.-M., LEPPIK, L., VÕRK, A., LEETMAA, R. (2004). Euroopa Liidu ühiste pensionieesmärkide mõju Eesti pensionisüsteemile. [The impact of the European Union common pension objectives on the Estonian pension system]. PRAXISe Toimetised 14/2004. http://www.praxis.ee/data/PRAXISe_Toimetised__Pension_27040410.pdf TITMUSS, R. (1974). Social Policy. London: George Allen and Unwin. TOOTS, A. (2003). International and national actors in Estonian pension reform. – Delivering Public Services in CEE countries: Trends and developments. OECD LEED Program. VANDENBROUCKE, F. (2002). Foreword: Sustainable Social Justice and ‘Open Coordination’ in Europe. – G. Esping-Andersen, D. Gallie, A. Hemerijk and J. Myles. Why We Need a New Welfare State. Oxford: Oxford University Press, viii–xxiv. VEEL, L. (2004). Public Sector Marketing: the Case of the Estonian Funded Pension. SSE Riga Working Papers 2004:3 (58). Bachelor thesis. Stockholm School of Economics in Riga.
147
VILLA, P. (2004). Typologie et équivalence des systèmes de retraites. CEPII Document de Travail No 9. Paris. VÕRK, A., UUDEKÜLL, K. (2002). Ennetähtaegne vanaduspension ja selle seos tööturukäitumisega Eestis. [Early retirement pensions and their impact on labor market behavior in Estonia]. Tartu. WAGENER, H.-J. (1999). Social Security – A Second Phase Transformation Phenomenon? – K. Müller, A. Ryll and H.-J. Wagener (Eds.) Transformation of Social Security: Pensions in Central-Eastern Europe. Heidelberg: Physica-Verlag, 13–30. WALKER, A. (1983). Social Policy, Social Administration and the Social Construction of Welfare. – M. Loney, D. Boswell and J. Clarke (Eds.) Social Policy and Social Welfare. Milton Keynes: Open University Press, 127–152. WEBB, E.J., CAMPBELL, D.T., SCHWARTZ, R.D., SECHREST, L. (1966). Unobtrusive Measures: Nonreactive Measures in the Social Sciences. Chicago: Rand McNally. WHITEHOUSE, E. (2004). Comparing the new pension systems of the Baltic States. – Organisation for Economic Co-operation and Development. Pension Reform in the Baltic Countries. Private Pension Series No. 5. Paris: OECD, 25–57. WILENSKY, H. (1975). The Welfare State and Equality: Structural and Ideological Roots of Public Expenditures. Berkeley: University of California Press. WILENSKY, H. (2002). Rich Democracies: Political Economy, Public Policy and Performance. Berkeley: University of California Press. WORLD BANK (1994). Averting the Old Age Crisis: Policies to Protect the Old and Promote Growth. New York: Oxford University Press. WORLD BANK (2006). Pension Reform and the Development of Pension Systems: An Evaluation of World Bank Assistance. Washington, D.C.: The World Bank. YIN, R.K. (1994). Case Study Research: Design and Methods. Applied Social Research Methods Series, Vol.5. London-Thousand Oaks-New Delhi: SAGE Publications.
148
Legal acts 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) 14) 15) 16) 17) 18) 19) 20) 21) 22)
Riiklike pensionide seadus, vastu võetud NSV Liidu Ülemnõukogu poolt 14. juulil 1956. Eesti Vabariigi pensioniseadus (15.04.1991), RT 1991,12,169 Eesti Vabariigi Ülemnõukogu otsus „Pensionide alammäärade ja hooldustasu tõstmisest“ (27.09.1990), RT 1990,11,119 Eesti Vabariigi Ülemnõukogu otsus „Muudatuste ja täienduste tegemise kohta Eesti Vabariigi Ülemnõukogu otsuses «Eesti Vabariigi pensioniseaduse rakendamise kohta»“ (7.11.1991), RT 1991,40,491 Eesti Vabariigi soodustingimustel vanaduspensionide seadus (14.05.1992), RT 1992,21,292 Eesti Vabariigi Ülemnõukogu otsus „Eesti Vabariigi soodustingimustel vanaduspensionide seaduse rakendamise kohta“ (14.05.1992), RT 1992,21,293 Eesti Vabariigi väljateenitud aastate pensionide seadus (14.05.1992), RT 1992,21,294 Eesti Vabariigi Ülemnõukogu otsus „Eesti Vabariigi väljateenitud aastate pensionide seaduse rakendamise kohta“ (14.05.1992), RT 1992,21,295 Eesti Vabariigi Ülemnõukogu otsus „Muudatuste ja täienduste tegemise kohta Eesti Vabariigi Ülemnõukogu otsuses «Eesti Vabariigi pensioniseaduse rakendamise kohta»“ (12.08.1992), RT 1992,33,422 Eesti Vabariigi Ülemnõukogu otsus „Eesti Vabariigi pensioniseaduse taaskehtestamise kohta“ (12.08.1992), RT 1992,33,423 Riigikogu otsus „Eesti Vabariigi Ülemnõukogu 1992.aasta 12.augusti otsuse «Eesti Vabariigi pensioniseaduse taaskehtestamise kohta» muutmine“ (15.12.1992), RT 1992,54,660 Riiklike elatusrahade seadus (17.03.1993), RT 1993,15,256 Eesti Vabariigi soodustingimustel vanaduspensionide seaduse ja Eesti Vabariigi väljateenitud aastate pensionide seaduse osalise muutmise ja täiendamise seadus (16.06.1993), RT I 1993,40,596 Riiklike elatusrahade seaduse muutmise ja täiendamise seadus (23.03.1994 ), RT I 1994,24,397 Soodustingimustel vanaduspensione ja väljateenitud aastate pensione käsitlevate õigusaktide osalise muutmise seadus (23.03.1994), RT I 1994,24,398 Soodustingimustel vanaduspensione ja väljateenitud aastate pensione käsitlevate õigusaktide osalise muutmise seadus (27.06.1994), RT I 1994,51,856 Soodustingimustel vanaduspensionide seaduse ja väljateenitud aastate pensionide seaduse muutmise seadus (15.11.1994), RT I 1994,83,1450 Riiklike elatusrahade seaduse, soodustingimustel vanaduspensionide seaduse ja väljateenitud aastate pensionide seaduse muutmise seadus (28.06.1995), RT I 1995,61,1027 Riiklike elatusrahade seaduse, soodustingimustel vanaduspensionide seaduse ja väljateenitud aastate pensionide seaduse muutmise ja täiendamise seadus (6.12.1995), RT I 1995,95,1631 Kaitseväe juhataja pensioni seadus (13.03.1996), RT I 1996, 22, 435 Riiklike elatusrahade seaduse, soodustingimustel vanaduspensione ja väljateenitud aastate pensione käsitlevate õigusaktide muutmise ja täiendamise seadus (20.03.1996), RT I 1996, 22, 437 Riiklike elatusrahade seaduse, soodustingimustel vanaduspensionide seaduse ja väljateenitud aastate pensionide seaduse muutmise seadus (21.11.1996), RT I 1996, 86, 1539
149
23) Riiklike elatusrahade seaduse, soodustingimustel vanaduspensionide seaduse ja väljateenitud aastate pensionide seaduse muutmise ja täiendamise seadus (5.11.1997), RT I 1997,81,1366 24) Riiklike elatusrahade seaduse, soodustingimustel vanaduspensionide seaduse ja väljateenitud aastate pensionide seaduse muutmise seadus (11.02.1998), RT I 1998,17,266 25) Sotsiaalmaksuseadus (15.04.1998), RT I 1998, 40, 611 26) Pensionifondide seadus (10.06.1998), RT I 1998,61,979 27) Riikliku pensionikindlustuse seadus [terviktekst muudatustega kuni 04.04.2000] (26.06.1998), RT I 2000,36,226 28) Riikliku pensionikindlustuse seadus [terviktekst muudatustega kuni 14.12.2000] (26.06.1998), RT I 2001,9,42 29) Riikliku pensionikindlustuse seadus (26.06.1998), RT I 1998,64/65,1009 30) Riiklike elatusrahade seaduse, soodustingimustel vanaduspensionide seaduse ja väljateenitud aastate pensionide seaduse muutmise seadus (25.11.1998), RT I 1998,107,1767 31) Eesti Vabariigi seaduse «Riigikogu liikmete ametipalga, pensioni ja muude sotsiaalsete garantiide kohta» muutmise seadus (19.01.1999), RT I 1999,10,153 32) Riikliku pensionikindlustuse seaduse, sotsiaalmaksuseaduse ja riiklike elatusrahade seaduse muutmise seadus (8.12.1999), RT I 1999,97,857 33) Riikliku pensionikindlustuse seaduse muutmise seadus (15.03.2000), RT I 2000,25,146 34) Soodustingimustel vanaduspensionide seaduse ja väljateenitud aastate pensionide seaduse muutmise seadus (15.03.2000), RT I 2000,25,147 35) Riikliku pensionikindlustuse seaduse muutmise seadus (4.04.2000), RT I 2000,33,194 36) Parandatud ja täiendatud Euroopa sotsiaalharta ratifitseerimise seadus (31.05.2000), RT II 2000,15,93 37) Eesti Väärtpaberite Keskregistri seadus (14.06.2000), RT I 2000,57,373 38) Riikliku pensionikindlustuse seaduse muutmise seadus (13.12.2000), RT I 2000,102,674 39) Sotsiaalmaksuseadus (13.12.2000), RT I 2000, 102,675 40) Riikliku pensionikindlustuse seaduse § 41 muutmise seadus (14.12.2000), RT I 2000,102,673 41) Eesti Vabariigi seaduse «Riigikogu liikmete ametipalga, pensioni ja muude sotsiaalsete garantiide kohta» muutmise seadus (13.02.2001), RT I 2001,21,117 42) Kogumispensionide seadus (12.09.2001), RT I 2001,79,480 43) Riikliku pensionikindlustuse seadus (5.12.2001), RT I 2001,100,648 44) Tagatisfondi seadus (20. 02. 2002), RT I 2002,23,131 45) Riikliku pensionikindlustuse seaduse muutmise seadus (19.06.2002), RT I 2002,53,338 46) Kogumispensionide seadus (14.04.2004), RT I 2004,37,252 47) Riikliku pensionikindlustuse seaduse muutmise seadus (12.06.2003), RT I 2003,48,343 48) Euroopa sotsiaalkindlustuskoodeksi ratifitseerimise seadus (10.03.2004), RT II 2004,6,17 49) Riikliku pensionikindlustuse seaduse muutmise seadus (18.04.2004), RT I 2004,16,120 50) Riikliku pensionikindlustuse seaduse § 43 muutmise ja täiendamise seadus (15.12.2004), RT I 2004,89,608 51) Riikliku pensionikindlustuse seaduse § 61 muutmise seadus (15.06.2005), RT I 2005,37,283 52) Soodustingimustel vanaduspensionide seaduse ja väljateenitud aastate pensionide seaduse muutmise seadus (15.06.2005), RT I 2005,38,295
150
Other acts 1) 2) 3)
Vabariigi Valitsuse 5. mai 1997. a. korraldus nr. 328-k „Sotsiaalkindlustusreformi komisjoni moodustamine“ (RT I 1997,37/38,579) Vabariigi Valitsuse 11. juuni 1997. a. korraldus nr. 464-k “Valitsuse tegevuse põhieesmärkide kinnitamine 1997. ja 1998. aastaks” (RTI 1997, 46, 766) Vabariigi Valitsuse 4. mai 1999.a korraldus nr 523-k “Vabariigi Valitsuse 5. mai 1997. a. korralduse nr 328-k “Sotsiaalkindlustusreformi komisjoni moodustamine” muutmine” (RTL 1999,78,973)
Other sources Working materials of the Government Social Security Reform Commission 1) 2) 3) 4) 5) 6) 7) 8)
Pensionireformi kontseptuaalsed alused, heaks kiidetud Vabariigi Valitsuse istungil 3. juunil 1997. a. Sotsiaalkindlustusreformi komisjoni koosoleku protokoll, 4. juuni 1999.a. Kohustusliku kogumispensionisüsteemi kava, 1. juuli 1999. a. projekt Kohustusliku kogumispensioni (nn. teise pensionisamba) rakendamise põhiküsimused, Vabariigi Valitsuse 13. juuli 1999. a. kabinetinõupidamise arutelu tulemustele tuginev projekt Seletuskiri kohustusliku kogumispensionisüsteemi rakendamise põhiküsimuste kohta Kogumispensionisüsteemi eesmärgid, põhivalikud ja rakendamise kava: diskussiooni alusdokument, detsember 1999. a. Sotsiaalkindlustusreformi komisjoni ettepanek kogumispensionisüsteemi ehk nn. II pensionisamba käivitamiseks (16+4+2), Vabariigi Valitsuse 27. juuni 2000. a. kabinetinõupidamise materjal Vabariigi Valitsuse otsused kogumispensionisüsteemi (teise pensionisamba) rakendamisest, 23. jaanuar 2001. a.
Other materials 1) 2) 3) 4) 5) 6) 7)
Eesti Rahvusliku Sõltumatuse Partei, “Isamaa” ja “Mõõdukate” koalitsioonilepe (1992) http://www.valitsus.ee/failid/Valitsuse_programm_1990.pdf Eesti Reformierakonna, Isamaaliidu ja Mõõdukate koalitsioonilepe (1999) http://www.valitsus.ee/?id=1118 Eesti Keskerakonna ja Eesti Reformierakonna koalitsioonileping (2002) http://www.valitsus.ee/failid/koalitsioonileping_2002.pdf Ühendus Vabariigi Eest – Res Publica, Eesti Reformierakonna ja Eestimaa Rahvaliidu koalitsioonileping (2003) http://www.respublica.ee/old_site/docs/koalitsioonilepe_lopp_2.pdf Eesti Reformierakonna, Eesti Keskerakonna ja Eestimaa Rahvaliidu koalitsioonilepe (2005) http://www.reform.ee/dokumendid/RE-KE-RL.doc IMF Press Information Notice No 97/41, 24 December 1997, http://www.imf.org/external/np/sec/pn/1997/pn9741.htm IMF Public Information Notice No 00/49, 11 July 2000, http://www.imf.org/external/np/sec/pn/2000/pn0049.htm
151
ELULOOKIRJELDUS Nimi: Lauri Leppik Sünniaeg: 11. detsember 1964 Kodakondsus: Eesti Haridus 1971–1979 1979–1982 1982–1990 1992–1993 2003 2004–2006
Tallinna 32. Keskkool Tallinna 1. Keskkool Tallinna Pedagoogiline Instituut, matemaatika-füüsika eriala, diplom kiitusega Southern Connecticut State University (USA), sotsiaaltöö magistriprogrammi kursused Tallinna Pedagoogikaülikool, sotsiaaltöö magistriprogramm, Magister Artium kraad sotsiaaltöö alal Tallinna Ülikool, sotsiaalteaduste doktorikool, sotsiaaltöö eriala
Teenistuskäik 2001– Poliitikauuringute Keskus PRAXIS, sotsiaalpoliitika analüütik 1999–2001 Sotsiaalministeerium, nõunik 1998–1999 European University Institute, Robert Schuman Centre, teaduslik kaastöötaja 1997–1998 Sotsiaalkindlustusamet, peadirektori asetäitja 1995–1997 Sotsiaalkindlustusamet, nõunik 1991–1995 Tallinna Pedagoogikaülikool, vanemõpetaja 1990–1991 Tartu Teaduspark, grupijuht Osalemine erialaliste organisatsioonide ja komiteede töös Euroopa Sotsiaalõiguste Komitee liige Euroopa Sotsiaalkindlustuse Instituudi liige Eesti pere- ja sündimusuuringu teadusnõukogu liige Statistikaameti sotsiaalstatistika teadusnõukogu liige Vabariigi Valitsuse sotsiaalkindlustusreformi komisjoni liige (1997–2001)
152
CURRICULUM VITAE Name: Lauri Leppik Date of birth: 11. detsember 1964 Citizenship: Estonian Education 1971–1979 1979–1982 1982–1990 1992–1993 2003 2004–2006
Employment 2001– 1999–2001 1998–1999 1997–1998 1995–1997 1991–1995 1990–1991
Tallinn Secondary School No. 32 Tallinna Secondary School No. 1 Tallinna Pedagogical Institute, speciality of mathematica and physics, diploma cum laude Southern Connecticut State University (USA), graduate courses in Master of Social Work program Tallinn Pedagogical University, Magister Artium in social work Tallinn University, Doctoral School of Social Sciences, social work program
Center for Policy Studies PRAXIS, social policy analyst Ministry of Social Affairs, adviser European University Institute, Robert Schuman Centre, research fellow Social Insurance Board, deputy director general Social Insurance Board, adviser Tallinna Pedagogical University, senior teacher Tartu Science Park, research group leader
Membership in professional organisations and committees Member of the European Committee of Social Rights Member of the European Institute of Social Security Member of the Research Council on Estonian Family and Fertility Survey Member of the Research Council on Socal Statistics of the Estonian Statistical Office Member of the Government Social Security Reform Commission (1997–2001)
153
TALLINNA ÜLIKOOL SOTSIAALTEADUSTE DISSERTATSIOONID 1. MARE LEINO. Sotsiaalsed probleemid koolis ja õpetaja toimetulek. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 1. Tallinn: TPÜ kirjastus, 2002. 125 lk. ISSN 1406–4405. ISBN 9985-58-227-6. 2. MAARIS RAUDSEPP. Loodussäästlikkus kui regulatiivne idee: sotsiaal-psühholoogiline analüüs. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 2. Tallinn: TPÜ kirjastus, 2002. 162 lk. ISSN 1406–4405. ISBN 9985-58-231-4. 3. EDA HEINLA. Lapse loova mõtlemise seosed sotsiaalsete ja käitumisteguritega. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 3. Tallinn: TPÜ kirjastus, 2002. 150 lk. ISSN 1406–4405. ISBN 9985-58-240-3. 4. KURMO KONSA. Eestikeelsete trükiste seisundi uuring. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 4. Tallinn: TPÜ kirjastus, 2003. 122 lk. ISSN 1406–4405. ISBN 9985-58-245-2. 5. VELLO PAATSI. Eesti talurahva loodusteadusliku maailmapildi kujunemine rahvakooli kaudu (1803–1918). Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 5. Tallinn: TPÜ kirjastus, 2003. 206 lk. ISSN 1406–4405. ISBN 9985-58-247-0. 6. KATRIN PAADAM. Constructing Residence as Home: Homeowners and Their Housing Histories. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 6. Tallinn: TPÜ kirjastus, 2003. 322 lk. ISSN 1406–4405. ISBN 9985-58-268-3. 7. HELI TOOMAN. Teenindusühiskond, teeninduskultuur ja klienditeenindusõppe konseptuaalsed lähtekohad. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 7. Tallinn: TPÜ kirjastus, 2003. 368 lk. ISSN 1406–4405. ISBN 9985-58-287-X. 8. KATRIN NIGLAS. The Combined Use of Qualitative and Quantitative Metods in Educational Research. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 8. Tallinn: TPÜ kirjastus, 2004. 200 lk. ISSN 1406–4405. ISBN 9985-58-298-5. 9. INNA JÄRVA. Põlvkondlikud muutused Eestimaa vene perekondade kasvatuses: sotsiokultuuriline käsitus. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 9. Tallinn: TPÜ kirjastus, 2004. 202 lk. ISSN 1406–4405. ISBN 9985-58-311-6. 10. MONIKA PULLERITS. Muusikaline draama algõpetuses – kontseptsioon ja rakendusvõimalusi lähtuvalt C. Orffi süsteemist. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 10. Tallinn: TPÜ kirjastus, 2004. 156 lk. ISSN 1406–4405. ISBN 9985-58-309-4. 11. MARJU MEDAR. Ida-Virumaa ja Pärnumaa elanike toimetulek: sotsiaalteenuste vajadus, kasutamine ja korraldus. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 11. Tallinn: TPÜ kirjastus, 2004. 218 lk. ISSN 1406–4405. ISBN 9985-58-320-5. 12. KRISTA LOOGMA. Töökeskkonnas õppimise tähendus töötajate kohanemisel töötingimustega. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 12. Tallinn: TPÜ kirjastus, 2004. 238 lk. ISSN 1406–4405. ISBN 9985-58-326-4. 13. МАЙЯ МУЛДМА. Феномен музыки в формировании диалога культур (сопоставительный анализ мнений учителей музыки школ с эстонским и русским языком
154
обучения). Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 13. Tallinn: TPÜ kirjastus, 2004. 209 lk. ISSN 1406–4405. ISBN 9985-58-330-2. 14. EHA RÜÜTEL. Sociocultural Context of Body Dissatisfaction and Possibilities of Vibroacoustic Therapy in Diminishing Body Dissatisfaction. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 14. Tallinn: TPÜ kirjastus, 2004. 91 lk. ISSN 1406–4405. ISBN 9985-58-352-3. 15. ENDEL PÕDER. Role of Attention in Visual Information Processing. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 15. Tallinn: TPÜ kirjastus, 2004. 88 lk. ISSN 1406–4405. ISBN 9985-58-356-6. 16. MARE MÜÜRSEPP. Lapse tähendus eesti kultuuris 20. sajandil: kasvatusteadus ja lastekirjandus. Tallinna Pedagoogikaülikool. Sotsiaalteaduste dissertatsioonid, 16. Tallinn: TPÜ kirjastus, 2005. 258lk. ISSN 1406–4405. ISBN 9985-58-366-3. 17. АЛЕКСАНДР ВЕЙНГОЛЬД. Прагмадиалектика шахматной игры: основные особенности соотношения формально- и информально-логических эвристик аргументационного дискурса в шахматах. Tallinna Ülikool. Sotsiaalteaduste dissertatsioonid, 17. Tallinn: TLÜ kirjastus, 2005. 74 lk. ISSN 1406–4405. ISBN 9985-58-372-8. 18. OVE SANDER. Jutlus kui argumentatiivne diskursus: informaal-loogiline aspekt. Tallinna Ülikool. Sotsiaalteaduste dissertatsioonid, 18. Tallinn: TLÜ kirjastus, 2005. 110lk. ISSN 1406–4405. ISBN 9985-58-377-9. 19. ANNE UUSEN. Põhikooli I ja II astme õpilaste kirjutamisoskus. Tallinna Ülikool. Sotsiaalteaduste dissertatsioonid, 19. Tallinn: TLÜ kirjastus, 2006. 193lk. ISSN 1736–3632. ISBN 9985-58-423-6. 20. LEIF KALEV. Multiple and European Union Citizenship as Challenges to Estonian Citizenship Policies. Tallinna Ülikool. Sotsiaalteaduste dissertatsioonid, 20. Tallinn: TLÜ kirjastus, 2006. 164 lk. ISSN 1736–3632. ISBN-10 9985-58-436-8. ISBN-13 978-9985-58-436-1
ILMUNUD MONOGRAAFIANA AILE MÖLDRE. Kirjastustegevus ja raamatulevi Eestis. Monograafia. Tallinna Ülikool Tallinn: TLÜ kirjastus, 2005. 407 lk. ISBN 9985-58-347-7. LINNAR PRIIMÄGI. Klassitsism. Inimkeha retoorika klassitsistliku kujutavkunsti kaanonites. I-III. Monograafia. Tallinna Ülikool Tallinn: TLÜ kirjastus, 2005. 1242 lk. ISBN 9985-58-398-1, ISBN 9985-58-405-8, ISBN 9985-58-406-6.
155
