productivity and price performance in the privatised water ... - CiteSeerX

Productivity & Price Performance in the Privatised Water & Sewerage Companies of England & Wales

PRODUCTIVITY AND PRICE PERFORMANCE IN THE PRIVATISED WATER AND SEWERAGE COMPANIES OF ENGLAND AND WALES By David S Saal and David Parker

David Saal & David Parker, Aston Business School, Aston University, Birmingham, B4 7ET, UK

RP0029 November 2000 ISBN No: 1 85449 590 9

Aston Business School Research Institute is the administrative centre for all research activities at Aston Business School. The School comprises more than 70 academic staff organised into thematic research groups along with a Doctoral Programme of more than 50 research students. Research is carried out in all of the major areas of business studies and a number of specialist fields. For further information contact: The Director, Aston Business School Research Institute, Aston University, Birmingham B4 7ET Telephone No: (0121) 359 3611

Fax No: (0121) 333 5620

http://www.abs.aston.ac.uk/

Aston Business School Research Papers are published by the Institute to bring the results of research in progress to a wider audience and to facilitate discussion. They will normally be published in a revised form subsequently and the agreement of the authors should be obtained before referring to its contents in other published works.

0


Abstract After the ten Regional Water Authorities (RWAs) of England and Wales were privatised in November 1989, the successor Water and Sewerage Companies (WASCs) faced a new regulatory régime that was designed to both promote economic efficiency and improve water quality. Moreover, the new regulatory régime was designed to encourage the capital investment needed to improve water quality. The industry's economic regulator, the Office of Water Services (Ofwat), implemented a unique RPI + K pricing system which was designed simultaneously to compensate the WASCs for efficient capital investment programs, encourage gains in operating efficiency, and maintain fair prices for customers. It is now over ten years since privatisation. This paper evaluates how successful privatisation with economic regulation has been in raising productivity. Estimates of labour and total factor productivity (TFP) growth, derived with quality adjusted output indices, suggest that while substantial reductions in labour usage have occurred, TFP growth has not improved relative to the pre-privatisation period. Moreover, a total price performance (TPP) index, attributable to Jorgenson and Griliches (1967), indicates that increases in the quality adjusted output price of water and sewerage services in England and Wales have been substantially greater than increases in input costs. These results are consistent with economic profits having increased considerably since 1989.

Key Words: privatisation, productivity, price performance, utility regulation, water and sewerage industry, England & Wales JEL Classification: D24, L33, L51, L95

1


Introduction1 Governments have argued that state ownership is inferior to private ownership and privatisation is a way of raising corporate efficiency. In the UK over £60bn of industrial assets have been privatised since 1980. Where privatised firms continue to control natural monopoly networks, new regulatory structures have been introduced to prevent monopoly abuse. This paper is concerned with the performance of the water and sewerage industry in England and Wales before and after privatisation in November 1989.2 Performance is measured using indices of labour and total factor productivity (TFP). Estimates of economic profitability are also reported alongside an analysis of total price performance (TPP). The TPP analysis extends the approach adopted by Waters and Tretheway (1999) in their study of Canadian railways. The aim is to distinguish the contributions of productivity and price changes to economic profits. This is important in the case of the water and sewerage industry in England and Wales because criticism has been directed at the high profits earned in the industry since privatisation. The water and sewerage industry in England and Wales is structured as a number of regional or local monopolies. A provision was put in place in 1989, and strengthened in 1992, to allow new suppliers to operate, under licence within the area of an existing supplier (known as an ‘inset appointment’: Ofwat, 1995). Water supplies have also been opened up to limited cross-border competition. In practice, however, the vast bulk of consumers have no choice of supplier. Consequently, to avoid monopoly abuse, the industry is regulated by the Office of Water Services (Ofwat) headed by a Director General (DG). The industry is primarily regulated through a price cap that takes the general form of RPI+K. RPI is the ‘retail price index’ and measures inflation in the economy. The K factor is composed of an efficiency factor, X, which reflects the DG’s assessment of each company’s scope to reduce its unit costs over a five year period, and a Q factor to reflect the higher costs resulting from meeting stricter water quality targets. Quality is policed by the Drinking Water Inspectorate (DWI) and the Environment Agency (EA).3 Both, along with Ofwat, report to the Department of Environment, Transport and Regions (DETR). In recent years Directives of the

1

We would like to thank Emmanuel Thanassoulis, Joshy Easaw, and Jim Love for helpful comments on an earlier draft of this paper. The usual disclaimer applies. 2 This study is concerned only with the industry in England and Wales. Water and sewerage services in Scotland and Northern Ireland remain publicly owned. 3 The National Rivers' Authority (NRA) was initially responsible for monitoring the industry's environmental impact before the establishment of the EA in 1996. 2


European Commission, for example on urban waste water, have had a significant effect on costs in the industry. The industry has had to undertake much capital expenditure (capex) to meet the new water quality requirements and to make up for a large backlog of investment. Under state ownership investment had been inadequate to maintain the serviceability of the system. Since 1989 the K factor in the price cap has been positive leading to price rises. The Q factor, reflecting increases in capex and associated operating expenditure (opex) has more than offset the DG’s expected efficiency savings in the industry, or the X factor. In his latest price determinations for 2000-2005, announced in November 1999, however, the DG has programmed prices to fall. The companies will be expected to reduce their prices by an average of 2.1 percent per annum over the quinquennium (Ofwat, 1999). In recent years there have been a number of empirical studies that have looked at the impact of privatisation on economic performance. Some studies have confirmed performance improvements after privatisation (Galal, Tandon and Vogelsang, 1994; Megginson, Nash and van Randenburgh, 1994; Bhaskar and Khan, 1995; Newbery and Pollitt, 1997; Boubakri and Cosset, 1998). Others have been more sceptical suggesting that performance improvement is by no means guaranteed and that efficiency may be related to product market competition rather than ownership per se (e.g. Vickers and Yarrow, 1991; Bishop and Thompson, 1992; Bishop and Green, 1995; NERA, 1996; Martin and Parker, 1997; Parker and Wu, , 1998; Parker, 1999). There are also theoretical studies that suggest regulated privately-owned firms will not necessarily perform better than state owned enterprises (e.g. Sappington and Stiglitz, 1987; De Fraja, 1993; Laffont and Tirole, 1991; Pint, 1991; Willner, 1996) and regulation is considered to introduce disincentives in terms of the use of factor inputs, notably capital (Averch and Johnson, 1962; Bailey, 1973). The ‘price cap’ method of regulation used in the UK was intended to overcome the efficiency disincentives of rate of return regulation, but the continued operation of price caps over time risks reducing the incentives (Littlechild, 1983, 1986; Vass, 1997). To date there has been very little study of the comparative performance of the water and sewerage industry in England and Wales before and after privatisation. This is surprising given the importance of the industry to the economy and the controversy that surrounded its sell off. But the neglect reflects the difficulty that researchers face in creating a consistent data set for the enterprises over the relevant period and in particular in dealing with changes in water quality. The research reported in this paper involved the creation of a continuous data set and

3


the computation of a quality adjusted output series. This is important because a substantial portion of the additional capital input has been concerned with water quality enhancement, as well as capital maintenance. Base water and sewerage output, by contrast, has been fairly static, with an average annual growth rate of base output of only 0.7 percent since 1990 (Table 1). Turning to the few studies that have been conducted, Hunt and Lynk (1995) use a multi-output cost function to assess performance over the ten Regional Water Authorities (RWAs) in England and Wales between 1979/80 and 1987/88. They report the existence of substantial economies of scope between water supply and environmental activities that became the responsibility of the NRA after privatisation. (also see Lynk, 1993). They argue that these economies were lost when environmental regulation and a number of other activities such as river maintenance became the responsibility of the NRA. Their study suggests that privatisation may have led to a decline in performance because of the mode of restructuring. The study does not, however, test this proposition. By contrast, Shaoul (1997) does test for the effects of privatisation on performance. Based on financial and accounting data drawn from the privatised water industry, the conclusion is that ‘Significant increases in efficiency had occurred prior to privatisation leaving little room to improve efficiency without jeopardising levels of services and future service provision’ (p.479). There is no evidence of lower supply costs after privatisation. This study is more comparable to the research reported in this paper, but there are serious problems with it. Firstly, it only covers the period up to 1995, when a new and tighter price cap may have been expected to provide greater efficiency incentives. Secondly, performance measurement is limited to the use of accounting ratios, with all of the uncertainties associated with the accounting changes over the period discussed. Thirdly, and very importantly, the study does not take into account improvements in water quality. It is therefore heavily biased against finding performance improvements after privatisation. More limited in focus are the studies by O’Connell Davidson (1993) and Ogden (1994). These are concerned with changes in industrial relations in the water industry following privatisation. These changes included sweeping away public sector methods of collective bargaining and instituting union derecognition, localised pay bargaining and new labour contracts. A recent survey commissioned by the Transport and General Workers Union has similarly pointed to subcontracting of operations leading to less favourable terms for water workers after privatisation (PSIRU, 1999). Cox, Harris and Parker (1999) found that water

4


companies had resorted to new procurement methods to reduce operating costs, including outsourcing. These studies imply that privatisation has been associated with significant gains in labour productivity and perhaps in the use of other inputs. Also relevant to the current study is the research published by Bosworth and Stoneman (1998) and Europe Economics (1998). Both studies were undertaken as part of the price review process leading up to Ofwat’s K determinations for the years 2000 to 2005 (Ofwat, 2000). The Bosworth and Stoneman study was sponsored by one of the water and sewerage companies and the Europe Economics study by Ofwat. Bosworth and Stoneman look at data over two periods, 1979-89 and 1989-95 and employ value-added based measures of output. They report that labour productivity in water and sewerage averaged 2.2 percent per annum in the first period and only 0.03 percent per annum in the second. This suggests that labour productivity growth was lower under private ownership and over both periods productivity was poor compared to comparator industries. Their study only goes up to 1995, however, and it does not take much account of capital inputs. By contrast, the Europe Economics’ study found that in the five years after 1992/93, base-service operating expenditure in the water industry fell by around 3.8 percent per annum. Unfortunately, their study does not allow for a comparison with performance in earlier periods. There are similar difficulties with the international comparisons of water industries found in Smith, Hitchens and Davis (1982), O’Mahony (1998), and Ofwat (1998a). The research reported in this paper is directly concerned with the impact of privatisation on corporate performance and studies the period 1985 to 1999. It also analyses the impact of the price cap tightening that took effect on 1 April 1995. The research contributes to the existing literature on water privatisation and regulation because of the use of alternative measures of output to value-based measures, which are recognised to be problematic in regulated, non-competitive, markets (Bosworth, Stoneman and Thanassoulis, 1996). Also, total factor productivity is calculated, allowing a distinction between operating efficiency including labour productivity and total productive efficiency. This is important when discussing an industry with high capital investment. The paper is organised as follows. In the next section the structure of the water and sewerage industry in England and Wales is briefly described. The section can be omitted by those readers already familiar with the industry. The discussion then turns to the hypotheses tested and the method and data used to assess

5


performance. This is followed by a discussion of the results. The paper ends with the main conclusions.

The Water and Sewerage Industry in England and Wales The water and sewerage industry in England and Wales (henceforth, W&S,EW) developed during the nineteenth century as a mixture of municipal and small private undertakings. A major reorganisation of the industry occurred in 1974. The municipal undertakings were merged into ten RWAs and each became responsible for the whole water cycle within their catchment area. The RWAs covered around 75 percent of the country with twenty-nine ‘statutory water only companies’ (WoCs) remaining in private ownership. In the areas where water supplies were provided by the WoCs, sewerage services were the responsibility of the local RWA. In 1989 the RWAs were privatised, but most of the existing structure of the W&S,EW was retained. The ten RWAs became publicly quoted water and sewerage companies (WASCs)4 and the twenty-nine WoCs were re-established as normal public limited liability companies, thereby removing what had been severe constraints on their rate or return and/or dividends. The biggest change involved the hiving off of environmental regulation and operational activities to do with environmental services from the WASCs to a new regulatory body, the NRA. It was felt inappropriate for the water companies to remain responsible for regulating their own water quality once privately owned. Also, the government reconstructed the balance sheets of the RWAs prior to the sale. This included writing off of the industry’s debts of £4.95bn and giving a further £1.5bn cash injection, known as the ‘green dowry’, towards the cost of future investment. The W&S,EW serves around 22 million water and 21 million sewerage consumers with households accounting for over 90 percent of customers and 70 percent of the water delivered. The WASCs and the WoCs have a combined annual revenue of almost £7bn and invest over £3bn a year, of which between 40 percent and 50 percent is to meet environmental requirements set by government or the EU. Operating costs account for around 40 percent of revenue and capital charges for a further 30 percent with the return on capital accounting for the remainder (Ofwat, 1999, p.21). This study is concerned with the performance of the WASCs and not the WoCs, The WOCs have been subject to numerous mergers and takeovers since 1989 (see Table 1) and this creates serious data

4

The WASCs are identified in the tables below. 6


problems when tracking their performance over time. Their exclusion from the research, however, does not significantly detract from the results. This is because the WASCs dominate the W&S,EW. They provide water supply, water treatment and water distribution for 78 percent of all connected properties in England and Wales (Waterfacts 2000), alongside all sewerage treatment, collection and disposal. They also account for 90 percent of the industry’s investment. Moreover, only the WASCs were privatised in 1989. The industry’s economic regulator, Ofwat, is mainly concerned with prices, profits and the quality of services. The industry is regulated through Ofwat by a price cap. Initially this was set laxly in 1989, to ensure the success of the privatisation flotation and to allow price increases to finance new investment. The cap was reset and tightened from 1 April 1995 and, again, from 1 April 2000. With K set as a positive figure, customers saw sharp rises in their water and sewerage bills during the 1990s, the largest rises occurring before 1995 (Ofwat, 1999, p.24). Increases varied from company to company and across measured and unmeasured water and sewerage charges but averaged over 40 percent in real terms in the ten years after privatisation (ibid., p.76). The result was a large growth in profits and some spectacular returns to investors (Parker, 1997). The latest price determinations for 2000 to 2005 are intended to allow for some of this profit to be returned to consumers through lower bills. As the industry's profits have been justified on grounds that the WASCs outperformed the efficiency savings factored into their price caps, part of the research reported below involved decomposing the impact of productivity and price changes on performance. The industry invested around £33 bn. (in May 1999 prices) in the ten years after privatisation (Ofwat, 1999, p.21). This has had the desired effect on water quality and the quality of services. Ofwat uses a range of key outputs as measures of performance, including customer service standards, leakage, water delivered and interruptions to supply. Some of these indicators are included in Ofwat’s range of nine official ‘DG standards’. These standards are used by Ofwat to compare performance between companies and can have a bearing on the price caps set at the next periodic review. In Ofwat’s publication, Prospects for Prices (1998b), a number of significant quality improvements are mentioned. It is essential, therefore, that alterations in water quality and the quality of service are reflected in the output measure used when computing performance changes in the water industry. Output growth will be underestimated and price growth overestimated if output quality improves but is not fully measured.

7


Hypotheses, Method, and Data Hypotheses The argument as to why privatisation may lead to higher corporate performance has been well rehearsed. Capital market pressures are said to produce a superior principal-agent relationship in the private sector, leading to management incentives that are more consistent with efficient operations than exists under state ownership (Alchian, 1965; Jensen and Meckling, 1976; De Alessi, 1980; Böss, 1991; Holmstrom and Tirole, 1993) . Public ownership is also associated with political and economic goals that may conflict with the efficient use of factor inputs (Tullock, 1976; Buchanan, 1978; Boycko, Shleifer and Vishny, 1986). That privatisation results in improved performance is repeated in the UK Government’s White Paper on the privatisation of water: ‘Privatisation should lead to improved standards, greater efficiency, and a better allocation of resources within the water industry’ (DOE, 1986, p.13). At the same time, however, the extent of performance improvement resulting from privatisation depends, at least in part, upon shareholders’ ability to monitor management effort in the pursuit of efficiency gains (Grossman and Hart, 1980; Vickers and Yarrow, 1988). In addition, in the W&S, EW takeovers have been restricted to ensure that the regulator maintains sufficient firms to undertake his comparative assessment of their performance at each price review. In particular, no mergers have been allowed between the WASCs. The takeover threat is usually considered to be an important part of the capital market constraint on management behaviour. It should also be born in mind that in the period under study, 1985-99, there was negligible product market competition in the water industry. Some studies have suggested that after privatisation performance is likely to be most disappointing where there is limited product market competition (e.g. Vickers and Yarrow, 1986). The first hypothesis tested in the research is consistent with the view that privatisation with restricted capital market pressures and without choice for consumers will not necessarily lead to performance gains. Performance is measured in terms of the growth rate of both labour and total factor productivity: H1: privatisation of the water and sewerage industry in 1989 did not lead to a statistically significant improvement in performance at either the industry or individual company levels. Economic regulation is an attempt to introduce pressures on management to manage resources efficiently where there is limited or no product market 8


competition. The RPI + K pricing régime implemented by Ofwat is designed to compensate the privatised WASCs for their capital investment programs, encourage gains in economic efficiency, and maintain fair prices to customers. In addition to studying the impact of privatisation on performance,

this study

evaluates how successful the system of economic regulation has been so far in meeting its objectives. This is achieved by employing traditional estimates of productivity performance, coupled with a total price performance index (TPP) attributable to Jorgenson and Griliches (1967). The paper’s second hypothesis is concerned with the causes of increases in economic profitability: H2: Due to the implementation of RPI+K pricing regulation, increases in the economic profitability of the WASCs can be primarily attributed to improvements in their total factor productivity (TFP) performance, rather than an increase in the price of outputs relative to inputs(TPP). The regulatory pricing régime was tightened during the period of study, in 1995. The third hypothesis is concerned with the impact of tougher price regulation: H3: Regulatory adjustment of the price cap in 1995 led to a statistically significant change in performance at the aggregate industry level and in each of the individual WASCs. Method In this study, performance is assessed by using non-parametric methods to determine labour and total factor productivity (TFP) growth rates. As is standard, labour productivity growth is measured as changes in the volume of output relative to the volume of labour usage. Because of the substantial factor substitution that appears to have occurred in the water and sewage industry, as indicated by substantial labour shedding and capital investment, labour productivity growth will give an upwardly biased estimate of total productivity growth. Nevertheless, as labour inputs are the largest single component of noncapital related WASC operating costs, labour productivity growth is still a useful measure of improvements in WASC performance. In the W&S,EW, a substantial and increasing portion of the industry’s employment costs result from internally performed capital projects.

It is

necessary, therefore, to adapt the standard labour productivity growth measure if we wish to employ it as an accurate index of WASC operational efficiency. If the labour productivity growth measure were based on total employment, downwardly biased estimates of operational efficiency improvements would be generated. An index of non-capitalised labour productivity was therefore created by using data in each WASC annual report to remove employment attributable to 9


internal capital projects.

However, we also calculated a further labour

productivity series based on total employment measures, for comparison. In contrast to labour productivity growth, TFP provides a fuller estimate of productivity performance.

It takes into account other factors of production,

notably capital and materials and fuel usage. Also, despite several well known theoretical limitations, TFP growth is accepted as a standard index of technical progress because it provides a readily estimable index of Hicks neutral technical change (Diewert, 1976). We therefore follow common practice and estimate TFP growth as output growth less a Tornqvist weighted index of input usage growth. The formulation can be specified in logarithmic form as: (1)

TFPt = ln(Yt / Yt −1 ) − ∑ 1 / 2( s j ,t + s j ,t −1 ) ln( X j ,t / X j ,t −1 ) i

where sj,t represents factor Xj,t’s share of total economic costs at time t, and Yt is an output index. As discussed below, the factors employed are labour, capital, and other inputs including fuel and services. By exploiting the direct link that exists between productivity, prices, and financial performance, Waters and Tretheway (1999) demonstrate that it is possible to represent overall financial performance as a function of TFP growth and total price performance (TPP). This index, which is attributable to Jorgenson and Griliches (1967), measures the growth of input prices relative to the growth of output prices. Assuming that a firm's revenues and economic costs are equal in each period, as is the case under perfect competition, Jorgenson and Griliches (1967) demonstrate that TPP and TFP growth must be the same because firms are required by market conditions to pass on all productivity gains and reduced input costs to consumers.

This is not the case under conditions of imperfect

competition, like those found in a regulated monopoly such as the water and sewerage industry.

Therefore, in the absence of perfect competition, a

comparison of TFP and TPP measures will reveal to what extent an increase in economic profits occurred because of improved productivity performance or because increases in output prices were greater than increases in input costs. To better illustrate the link between TFP and TPP, a simple example can be used where production occurs in two time periods. It is assumed that both output and input prices and quantities can be characterised by a single index number that accurately represents multiple output and input prices and quantities. Here Pt is an output price index for time t, Yt is an aggregate output index capturing multiple outputs, Wt is an aggregate input price index, and Xt is an aggregate input usage index capturing multiple input use.

As explained in Waters and Tretheway

(1999, p.211) "the(se) respective price and quantity indices must be dual to one 10


another so that there is computational consistency". This simply implies that if total revenues (Rt) and economic costs (Ct), are respectively expressed as Rt=PtYt and Ct=WtXt, it must be the case that the respective price indices are the dual to the aggregate output and input indices. Therefore, Wt must equal Ct/Xt and Pt must equal Rt/Yt. Given this framework, the relationship between productivity, prices and financial performance can be demonstrated by employing the ratio of revenues to economic costs ( π ), which reflects economic profitability. It is then possible to generate an index which we label ‘economic profitability’ (ψ t ) in period t relative to period t-b.

Rt πt Ct ψt = = π t −b Rt −b C t −b

(2)

In this notation b represents the number of years between the index base year and year t. By substituting for R and C and rearranging terms, ψ t is expressed as:

 Yt  Y t −b ψt =   X t X t −b 

(3)

Considering

this

expression

reveals

     that

 Wt  W t −b   Pt  Pt −b the

    

first

bracketed

term,

 Yt  X   Y   t X  is simply an index of TFP in period t relative to period tt −b   t −b   b.

5

W The second bracketed term,  t 

  Pt  is identified as a total Wt −b   Pt −b 

price performance index ( TPPIndex ) by Waters and Tretheway (1999). This is an index of aggregate input prices in period t relative to t-b divided by an index of aggregate output prices in period t relative to t-b.

It therefore measures the

growth in input prices relative to the growth of output prices and is, in general, an indicator of a firm's ability to pass increases in costs on to customers. This is often identified as "price cost recovery" in the management literature. In our specific application, the inclusion of capital and capital costs in C t , X t , and Wt

5

For expositional ease it is assumed that the aggregate input usage index is X t . In practice, the presence of multiple inputs requires TFP to be calculated as the index equivalent of Equation (1). 11


permits employing the TPPIndex to evaluate whether increases in quality-adjusted output prices have been justified by the increase in economic costs that can be attributed to increased capital investment by the WASCs. In sum, consideration of the ψ t index reveals that changes in financial performance can be attributed to improved productivity performance (TFP growth) or improvements in price performance (measured as a decline in the TPP index). Therefore, ψ t can be most simply expressed as:

ψt =

(4)

TFPIndex,t TPPIndex,t

By taking log differences of Equation 4, we can also generate the following approximation of the growth rate of ψ t . This expression reveals that the growth rate of ψ t can be decomposed as the growth rate of TFPIndex less the growth rate of TPPIndex : (5)

(

) (

ln(ψ t ) − ln(ψ t −1 ) = ln(TFPIndex,t ) − ln(TFPIndex,t −1 ) − ln(TPPIndex,t ) − ln(TPPIndex,t −1 )

)

This makes it possible to decompose the proportion of changes in financial performance that can be attributed to productivity and price performance. This formulation is particularly useful in analysing the performance of the WASCs.

It allows us to evaluate whether the RPI+K price cap method of

regulation has been associated with increased economic profits arising from gains in efficiency, as designed, or because it has resulted in overly generous increases in quality-adjusted output prices relative to aggregate input costs. Moreover, the decomposition of changes in financial performance demonstrated in Equation (5) will indicate the relative importance of productivity and price performance in the overall performance of the WASCs. Output Data The privatised water and sewerage companies are multi-output producers. Therefore, the calculation of labour productivity, TFP, and TPP indices requires the construction of an aggregated output index which captures the growth of both water and sewerage treatment services. Moreover, the physical output of the water and sewerage industry has not increased substantially in the past 15 years, while there have been substantial increases in both drinking water and sewerage treatment quality. This means that a quality adjusted output index should be used

12


in order to measure the output growth of the WASCs.

The following

formulation for quality-adjusted output has been employed: (6) Yt = SWater ,tWatert QualityWater ,t + S Sewerage,t Sewerage t Quality Sewerage,t where SWater ,t and S Sewerage,t are the respective shares of water and sewerage services in total turnover at time t, Watert and Sewerage t are indices of the base level of water and sewerage treatment demand indexed to equal 100 in 1990, and

QualityWater ,t and Quality Sewerage,t are indices of the quality of drinking water and sewerage treatment indexed to equal 1 in 1990.

Watert is proxied by the resident water supply population served by each WASC, available from the publication CIPFA (Services and Costs) for the years ending 1985-88 and the companies’ July regulatory returns to Ofwat. The missing years ending 1989 and 1990 were estimated by interpolation.

While a physical

measure of water usage could be argued to be an appropriate measure of changes in water demand, figures for "water delivered" are not available for the preprivatisation period . "Water supplied" figures are available, but they suffer a downward bias because of the WASCs’ efforts to reduce leakage, particularly after Ofwat set leakage targets following the 1995 drought.6

Sewerage t could be proxied by equivalent sewerage treatment population for each WASC, as reported in CIPFA (Services and Costs) and the companies’ July returns to Ofwat. This is because equivalent sewage treatment population is calculated as the sum of the sewerage treatment population plus trade effluent, cesspit, and septic tank load converted to a population equivalent. It is therefore simultaneously a population based measure and a physical output measure. Nonetheless, we have chosen not to report results based on estimated equivalent sewerage treatment population.

This is primarily because the available data

series for several of the WASCs are characterised by substantial, dramatic, and erratic changes in estimated equivalent sewerage treatment population of up to 73 percent in one year, strongly suggesting that these estimates are not reliable. We have, instead, chosen to employ more robust estimates of the population connected to sewerage treatment works, utilising data drawn from CIPFA (Services and Costs), Waterfacts, and the companies’ July returns.7 6

The population based measure will not properly measure variation attributable to water usage by industry. Nevertheless, Ofwat estimate an over 0.9 correlation between population and water delivered and have used population in their own performance modeling (Ofwat, 2000b). 7 The data on the population connected to sewerage treatment plants exhibits an average annual growth rate of 0.2 percent for the 1985-90 period and 0.5 percent for the 1990-99 period, while the equivalent 13


The water quality index , QualityWater ,t is defined as the ratio of the average percentage of each WASC’s water supply zones that are compliant with key water quality parameters, relative to the average compliance percentage in the 1990. The data were drawn from the DWI’s annual report on drinking water quality for the years ending 1991-99.8 Water supply zones are areas designated by the water companies, by reference to a source of supply, in which not more than 50,000 people reside. As these water quality parameters have not changed substantially since their creation at privatisation, and a new EC drinking water directive did not come into force until after the sample period, this data provides a consistent measure of water quality for the entire post-privatisation period. However, as numerical standards were not applied to drinking water quality before the creation of the DWI, consistent quality data is not available for the period 1985-1990. After consultation with the DWI, it is assumed that drinking water quality in the years 1985-90 was the same as in the year 1991, based on the belief that drinking water quality did not change significantly during this period.9 While the specific key water quality parameters included in the quality index will influence its trend, experimenting with different formulations showed little variation in the growth of water quality. This was so regardless of whether all of the parameters were included; or there was a focus on the nine parameters identified by Ofwat as being important for aesthetic reasons, health reasons, and costs reasons (Ofwat, 1996) or a focus on the six parameters that the DWI has identified as being indicative of how well treatment works and distribution systems are operated and maintained (DWI, 1999).10 It was decided to report the results based on the Ofwat nine water quality measure, because these parameters were specifically chosen by the economic regulator as being indicative of parameters that are valued by consumers, and/or have been relatively costly to improve. Several alternative measures of sewage treatment quality were considered for use in the construction of Quality Sewerage,t . A measure of the percentage of sewage treatment plants in compliance with discharge consent orders is available from population series results in estimated annual growth of 1.4 percent for 1985-90 and 0.4 percent for 1990-99. This suggests that if the data on population connected to sewerage treatment works is a biased estimate of sewerage output, it will overestimate productivity growth in the post-privatisation period. 8 Following Hunt & Lynk (1995), the quality indices reported in this paper, have been generated by dividing the value of the relevant environmental indicator in a given year by its value in 1990. 9 This assumption would tend to create a bias favouring greater productivity growth in the post privatisation period. However, alternative modelling, assuming a backward trend in drinking water quality for the period before 1991, did not significantly alter the conclusions. 14


the Digest of Environmental Statistics beginning with the year ending 1987, when the RWAs were first required to report this data. After discussion with the EA and Ofwat, however, it was decided not to base the results on this measure because numerical discharge consent orders have grown progressively more restrictive over time. This could cause a downward bias in estimated sewerage treatment quality growth.11 Instead, a weighted average of river quality and bathing water quality for each WASC was used. Improvements in these two environmental quality indicators are among the primary goals of sewerage treatment quality improvements.

A

weighted average is necessary because the relative importance of each quality measure varies significantly, as exemplified by the fact that Severn Trent has no identified bathing waters, while South West Water has 31 percent of all identified bathing waters in England and Wales. Therefore, it has been assumed that: (7)

 S River ,t Quality Sewerage ,t =   S River ,t + S Bathing ,t 

  S Bathing ,t  River  Quality ,t +   S   River ,t + S Bathing ,t

  Bathing Quality , t  

where S River ,t and S Bathing ,t are respectively a WASC’s share of all classified river length and its share of all identified bathing waters in England & Wales. The actual data used to derive

Bathing Quality ,t

was based on the percentage of

bathing waters in compliance with Directive 76/160/EEC, which regulates the allowable level of coliforms in bathing waters. These data are readily available from the NRA and later EA annual reports, the Digest of Environmental Statistics, and Ofwat’s annual level of service reports (e.g. Ofwat, 1996). As annual compliance has followed an upward but bumpy trend, the quality measure was constructed as the ratio of the three-year moving average of bathing water compliance, relative to the three-year average centred on 1990. Moreover, as the number of identified bathing waters has increased substantially over time, the quality measure has been adjusted proportionally, to reflect the greater effort required to be in compliance. Constructing

RiverQuality ,t

was somewhat more problematic because a consistent

data series on river quality is not available for the entire 1985-99 period. However, the 1985-91 data on chemical river quality, which is based on an earlier classification system, demonstrates virtually no improvement in river quality.

10

Details of the results of these different formulations may be obtained from the authors. As these numerical consent orders are plant specific it was not possible to obtain data on the average increase in compliance standards for each WASC/EA Region, data which could have been employed to correct this bias. 11

15


RiverQuality ,t

was therefore based on data drawn from the NRA and EA Annual

Reports on the net improvement in river quality between the fiscal year ending 1991 and the relevant year for which the index is derived. Net improvement is defined by the EA as the percentage of river length that has been upgraded less the percentage that has been downgraded, which has been used by the EA as an indicator of quality improvement since privatisation (EA, 1999). The river quality index is therefore defined as 1 in 1991 plus the net improvement in river quality relative to this base, with the net improvement assumed equal to zero for the years ending 1985-90.

An alternative methodology based on merging the

available data for 1985-91 with data for 1991-99 based on the new General Quality Assessment (GQA) Chemical Quality system, was rejected because substantial differences in these river classification systems made it difficult to link these series reliably. However, as this alternative methodology suggests a smaller improvement in river quality since privatisation than the series employed, these estimates, which are available from the authors on request, generate more conservative estimates of post-privatisation productivity growth than those presented here. Unfortunately, one limitation that plagues all the river quality indices results from the NRA’s decision to aggregate the former Northumbrian and Yorkshire regions into the North East region and the former Southwest and Wessex regions into a new Southwest Region, when reporting river quality. Hence, river quality data is not available separately for the four WASCs which are contained in the EA’s new North East and South West regions. To overcome this difficulty, a sewerage quality index for each of these companies was calculated by assuming that river quality in their respective EA region is an accurate representation of the actual river quality in the company’s area of operation. As this assumption may not be valid, estimates of the economic performance of the aggregated North East and South West EA regions are also reported below. Intput Data Data on the average number of full time equivalent (FTE) employees is available from each of the RWA's and later company's annual reports for the accounting years ending 31 March.12 At privatisation some employees were transferred to the new NRA. Data on the number of employees transferred and those remaining in the privatised companies is available from the companies' flotation

16


prospectuses.

Consequently, a consistent average full-time equivalent

employment series for the privatised functions of the RWAs was created by adjusting the pre-1990 data. The resulting employment series was then adjusted with data from the New Earnings Survey on average weekly hours of work in the water industry. Next, the nominal unit cost of labour in each accounting year was used to create an index of relative unit employment costs. This was used to calculate the total real cost of employment in each accounting year.

As

mentioned already, a substantial portion of employment costs in the water industry are attributed to capital projects. Therefore, an index of non-capitalised employment was generated to avoid the double counting of labour inputs and costs in the labour productivity, TFP and TPP calculations. The basis of the capital index is the modern equivalent asset (MEA) estimation of the replacement cost of net tangible fixed assets, as provided in each WASC's regulatory accounts, for the years 1990 to 1999. The 1991 regulatory accounts indicate a systematic re-valuation of the replacement costs of the net fixed asset values held by all the WASCs in 1990. This resulted in the tripling of the estimated replacement value of these assets.

This re-valuation generates an

inconsistency between the current cost net fixed assets reported in the privatisation prospectus for 1985-89, relative to those available for 1990-99. Nevertheless, as the 1990 MEA re-valuations were based on the previously reported current cost net fixed asset valuation for 1990, which are consistent with those reported in the privatisation prospectus, it proved possible to produce estimates of the replacement cost of the 1985-89 values that are consistent with the 1990-99 values. This was accomplished by adding the RPI inflation-adjusted value of the 1990 MEA revaluation to the available current cost net fixed asset data for 1985-89. As it is unlikely that any additions to fixed assets which were made between 1985 and 1989 were substantially re-valued in the 1990 MEA revaluation, this methodology should provide a relatively accurate and necessary backward extension of the current cost net fixed asset series. The adoption of MEA accounting policies by Ofwat creates an additional difficulty as the WASCs make periodic revisions (usually upward) of the current replacement costs of their fixed assets. While the inclusion of such periodic revisions may increase the accuracy of the estimated current replacement cost of a company’s fixed assets at any given time, it also has the effect of generating jumps in a WASC’s capital index. 12

This will result in a downward bias in

For Northumbrian Water the 1998/99 year data employed is based on the company's 1998 calendar year report. This was necessitated by the company's switch in reporting period after its acquisition by 17


estimated TFP growth. The reported current cost net fixed asset values were therefore adjusted by backing out the RPI adjusted value of all MEA revaluations made in the 1991-99 period. This methodology not only removes the arbitrary jumps in capital values that were created by these periodic MEA revaluations, but also generates a capital series that is consistent with the perpetual inventory method. An additional issue is the relevant price index to be used in adjusting the nominal replacement cost of each WASC’s net fixed assets. While the RPI index is employed to adjust the current cost replacement values reported in the companies’ regulatory accounts, the industry argues that a water industry specific Capital Cost Index (CCI), prepared for the industry by London Economics is an accurate estimate of actual movements in capital prices (Waterfacts 2000).13 As the CCI index indicates that during the early 1990s capital costs in the industry rose at a substantially lower rate than retail price inflation, the RPI based current cost fixed asset values, reported in the regulatory accounts, will overestimate the replacement cost, and hence the opportunity cost, of the fixed capital stock during this period.

The nominal value of the current cost capital stock was therefore

recalculated by employing the CCI index instead of the RPI index. While the resulting capital input indices and hence the TFP measures that are generated differ little from those based on the RPI based series, the substantial difference in the nominal cost of capital does substantially improve the measurement of TPP in the early to mid 1990s. This is due to the inclusion of a different and better estimate of the replacement cost of capital, than that based on the RPI deflation of capital. The imputed nominal opportunity cost of capital was calculated as current cost depreciation and infrastructure renewal costs, plus an inflation adjusted rate of return on the replacement cost of the fixed asset base sufficient to provide a 6 percent post-tax rate of return in real terms.14 An index for the cost of capital was

Lyonnaise des Eaux SA. 13 Economists at Ofwat argue that the DETR’s Construction Price Index (COPI) is reflective of actual capital costs in the industry, but the CCI index is superior because it is water industry specific. As the CCI index is only available for the years ending 1989-99, it was extended backward to 1985 using the COPI index. As the COPI index indicates very limited increases in construction costs after 1990, alternative estimates based on a COPI based valuation of the capital stock generate moderately lower TFP growth rates, and a substantially larger decline in the TPP measure in the early 1990s. Therefore, such estimates provide a less favourable estimation of the industry’s performance after privatisation. 14 In the water industry, infrastructure assets (mainly below ground assets) are not depreciated when computing profits. Instead, an ‘infrastructure renewals charge’ is made and debited in the P&L account in lieu of depreciation. Other assets (e.g. treatment plants) are depreciated in the normal way. The choice of a 6% real rate of return is based on Ofwat’s determination during the 1994 price review that 5-6% was an appropriate post-tax rate of return for the WASCs. Interestingly, during the 1999 price review, when Ofwat reduced its assumption of the appropriate post-tax return to 4.25-5.25%, the 18


then derived.15 Other input costs were determined as current cost operating profits less current cost depreciation, infrastructure renewal expenditures, and non-capitalised manpower costs. A real index of other costs was then created by deflating these costs with the DETR’s purchasing price index for materials and services purchased by the construction materials industry. Finally, the aggregate input index for the TFP calculations was constructed as a cost share weighted Tornqvist index of the non-capitalised labour, capital, and other costs index with cost shares based on average two period expenditure shares expressed in consistent prices.

An aggregate input price index for the TPP

calculations is then constructed as a dual to the aggregate input index. Thus, total aggregate input costs (including the imputed opportunity cost of capital) were simply divided by the input price index. A quality-adjusted price index for WASC output was similarly calculated as turnover in nominal terms divided by the quality adjusted output index.

The Performance Results: Aggregate Level In order to evaluate the performance of the privatised WASCs, several distinct periods were identified. Privatisation took place on 22 November 1989, which was almost eight months into the 1989-90 accounting year. The years ending 1985-90 were considered, therefore, to be part of the transitional/pre-privatisation period. The post-privatisation period was defined as 1990-99 or accounting years ending 1991-99. Because the RPI+K pricing régime was more lenient to the companies from privatisation up until the first periodic review by the regulator in 1994/95, the post privatisation period has also been divided into two sub-periods, 1990-95 and 1995-99. As the K limits were announced by Ofwat a few months before their implementation, it could be argued that anticipation effects will result in changes in WASC performance in the year before the actual implementation of new K price limits. Moreover, given that actual privatisation took place during the 1989-90 accounting year and managers would have begun implementing changes

companies argued for a higher return on the grounds that historic returns had been in the 5.7-6.6% range. (Ofwat, 1999) Thus, while our results are not particularly sensitive to this parameter, 6% is well within the range reported by the WASCs. 15 Our assumption that the opportunity cost of capital includes the capital costs of the entire fixed asset base might be subject to challenge. It could be argued that the opportunity cost of many WASC assets is zero because they have no viable alternative uses. Alternatively, it could be argued that capital costs have been overestimated because the replacement cost of the fixed asset base exceeds its economic value. Alternative estimates were generated assuming a lower opportunity cost of capital. The results do not materially alter the conclusions reported below and can be obtained from the authors. 19


in anticipation of privatisation, it could also be argued that the accounting year ending 1989 should be taken as the last under state ownership. The results were checked, therefore, to see whether they were significantly altered by defining the pre-privatisation period as 1985-89, the post-privatisation period as 1989-99, and the two sub-periods of the privatisation period as 1989-94 and 1994-99. Differences from the main results are highlighted, where relevant, in the discussion below, and these alternative estimates are available from the authors upon request. Table 1 provides a summary of the growth rates and levels of the various input, output and performance indices that characterise aggregate WASC performance. Statistically significant differences in growth trends at the 10 percent level are indicated in bold and at the 5 percent level in bold italics. Focusing first on input usage, while the growth rate of the employment and other costs indexes are not significantly different in the pre and post-privatisation period, non-capitalised employment growth was significantly lower and capital growth was significantly greater in the post-privatisation period.

Given the capital intensity of the

WASCs, it is not surprising that aggregate input usage growth was also significantly higher after privatisation. Based on these trends, it would appear that privatisation has resulted in an overall increase in the volume of resources dedicated to water and sewerage services. At the same time, comparison of the 1990-95 and 1995-99 periods reveals that, despite the continued growth of the capital stock at a rate that is not significantly different from that during 1990-95, the 1995-99 growth rate of the aggregate input index is significantly lower than it was in 1990-95. This is because of a statistically significant increase in the rate of labour shedding and a statistically significant reduction in the WASCs' other costs index (mainly hired and contract services, materials and power), which had been increasing. This result suggests that, while the initial regulatory régime had been relatively lax, the 1994/95 price review appears to have been successful in encouraging the WASCs to improve their operational efficiency, while simultaneously maintaining incentives to invest in the capital stock.

20


Table 1: Aggregate WASC Performance Indices: England and Wales Average Annual % Change1

Index 1990=100

2 3 2 85-90 90-95 95-99 90-99

1985 1995 1999

Input Indices Employment

-1.9%

-1.5%

-4.5%

-2.8%

109.7

92.7

77.5

Non-Capitalised Employment

-1.8%

-2.5%

-6.4%

-4.2%

109.2

88.2

68.4

Capital Stock

0.2%

1.2%

1.1%

1.1%

99.0 106.2 110.9

Other Costs Index

4.3%

2.6%

-1.3%

0.9%

80.8 113.9 108.1

Aggregate Input Index

0.4%

1.2%

0.8%

1.0%

98.2 106.1 109.4

Output and Implicit Quality Indices Base Output4

0.2%

0.6%

0.7%

0.7%

98.9 103.1 106.2

Quality Adjusted Output5

2.7%

3.2%

1.8%

2.6%

87.6 117.6 126.5

Implicit Quality Index6

2.4%

2.6%

1.1%

1.9%

88.6 114.1 119.1

Labour Productivity

4.5%

4.8%

6.3%

5.4%

79.8 126.9 163.2

Non-Capitalised Labour Productivity

4.4%

5.8%

8.2%

6.8%

80.2 133.4 185.0

TFP

2.3%

2.1%

1.0%

1.6%

89.2 110.8 115.6

Productivity Indices

Price Related Indices Quality-Adjusted Output Price

6.4%

5.2%

2.6%

4.0%

72.6 129.8 143.7

Aggregate Input Price

8.7%

-3.5% 10.5%

2.7%

64.7

84.1 127.9

TPP

2.3%

-8.7%

7.9%

-1.3%

89.1

64.8

-0.02%

10.7%

-6.9%

2.9%

Economic Profitability Index

100.1 171.0 129.8

Share Decomposition of Changes in the Economic Profitability Index 85-90

90-95

95-99

90-99

TFP Contribution Share

106.28

0.19

0.15

0.55

TPP Contribution Share

-107.28

0.81

-1.15

0.45

89.0

7

1. Growth figures are estimated as log differences. 2. Avg. growth figures in bold are significantly different from those in 1985-90 at 10%, bold italic figures at 5% 3. Avg. growth figures in bold are significantly different from those in 1990-95 at 10%, bold italic figures at 5% 4. Base output is measured as Y Base, t = SWater , t Watert + S Sewerage, t Seweraget 5. Quality-Adjusted Output is measured as Yt = SWater ,t Watert QualityWater ,t + S Sewerage,t Seweraget Quality Sewerage,t 6. The implicit composite quality index is calculated as Yt YBase, t 7. TFP and TPP decomposition shares sum to one if the economic profitability index increases and negative one if it decreases.

The base output, implicit quality indices, and quality-adjusted output data summarised in Table 1 provide a less favourable impression of the postprivatisation period. Despite a statistically significant increase in the average rate of base output growth, from 0.2 percent per annum before privatisation to 0.7 percent afterwards, the average growth rate of both the implicit quality index and

21


quality-adjusted output is lower in the post-privatisation period, although this is not statistically significant.16 Moreover, the annual growth rate of both the implicit quality index and quality-adjusted output for 1995-99 is significantly lower than for 1990-95. This suggests that the marginal returns, measured in increased quality-adjusted output, of the WASCs' capital investment program are declining.17 This result is particularly striking, because, as discussed above, the quality and output indices that have been built into these estimates are, if anything, biased towards finding higher quality growth during the postprivatisation period. Table 1 also provides a summary of aggregated WASC productivity trends. While estimates of total labour productivity are provided for comparative purposes, non-capitalised labour productivity is a more accurate estimate of actual WASC labour productivity trends, for reasons already discussed. Given the large capital expenditure of the WASCs, it is not surprising that, while both labour productivity and non-capitalised labour productivity exhibit higher average growth rates in the post-privatisation period, only the non-capitalised labour productivity exhibits a statistically significant higher growth rate. Similarly, while both labour productivity indices exhibit higher growth rates over 1995-99, relative to those achieved for 1990-95, only non-capitalised labour productivity exhibits a statistically significant increase. As non-capitalised labour productivity is an effective index of improvements in the WASCs' day to day operational efficiency, the statistically significant increase in this measure's average growth rate during the post-privatisation period suggests that some improvement in the operation of the existing water and sewerage network can be attributed to the post-privatisation régime.18 However, 16

The increase in estimated base output may occur because the base sewerage output measure, population connected to sewerage treatment plants, captures some improvements in quality. This results because the percentage of connected population served by any sewerage treatment plant has increased somewhat since privatisation. Nevertheless, despite this potential bias toward higher measured quality-adjusted output growth in the post-privatisation period, the estimates do not generate a statistically significant increase in quality-adjusted output growth 17 A similar decline in the growth rates of the implicit quality index and quality-adjusted output is also evident in the alternative time periods, as mentioned above. However, in contrast to the results reported in the main text, the difference in growth rates between the 1989-94 and 1994-99 period is not statistically significant. 18 This conclusion is somewhat tempered by the alternative estimates of average non-capitalised labour productivity growth. The alternative time period estimates exhibit an identical trend in average noncapitalised labour productivity growth rates for the pre and post-privatisation periods, as well as for the two sub-periods of the post-privatisation era. In contrast to the results reported in the main text, however, the difference in growth rates between the periods 1989-94 and 1994-99 is almost always statistically significant, while the difference between the 1985-89 period and the 1989-99 period is usually not. Thus, while it can be concluded that non-capitalised labour productivity improved in the second period of privatisation, it cannot be unequivocally concluded that the non-capitalised labour productivity growth rate has been higher during the entire post-privatisation period. 22


no statistically significant increase in average non-capitalised labour productivity growth occurred between 1985-90 and 1990-95, while a statistically significant increase did occur between 1990-95 and 1995-99. This strongly suggests that the more rigorous economic regulation embodied in the 1994/95 price review was the primary stimulus for

operational efficiency gains and not the change of

ownership. Although there has been no statistically significant change in quality-adjusted TFP growth relative to the pre-privatisation period, the overall trend has been downward.19 Thus, while average quality-adjusted TFP growth during 1985-90 was 2.3 percent per annum, it was only 1.6 percent during the post-privatisation era. Also, average quality-adjusted TFP growth was 2.1 percent during the first period of privatisation but only 1.0 percent per annum during 1995-99 (Table 1). In other words, the results suggest that, while labour productivity has increased, the WASCs’ rate of total productivity improvement has declined. The results suggest, therefore, that high capital investment was undertaken relative to the actual increase in quality-adjusted output in the post-privatisation period. This suggestion is reinforced when it is recalled that if the output and quality indices built into the results are biased, they are most likely to suffer from an upward bias over time. The total price performance (TPP) and economic profitability indices of the aggregated WASCs are summarised in the next two sections of Table 1 and are illustrated in Figures 1 and 2. As Table 1 and Figure 1 show, the growth rate of the quality-adjusted output price has followed a fairly stable upward trend, although it is significantly lower during the post-privatisation period. In contrast, the aggregate input price index is very volatile. Much of this volatility results because the opportunity cost of capital is by far the largest single determinant of total economic costs, in what is a capital intensive industry.20 Since the base rental price of capital is defined as the nominal rate of return necessary to provide a 6 percent real post-tax rate of return, trends in the opportunity cost of capital, and hence the aggregate input price index, are substantially influenced by trends in inflation. In other words, the capital intensity of the WASCs ensures that an

19

None of the various estimates of quality-adjusted TFP growth calculated suggested that TFP growth trends have increased since privatisation. In fact, several suggest that average TFP growth during the post-privatisation period has been statistically significantly lower. Even estimates with a reduced opportunity cost of capital, which tend to downplay the role of increases in the capital stock on TFP growth, do not generate any significant changes in the rate of TFP growth after privatisation. 20 Estimates of WASC performance after assuming that a substantial portion of the capital stock has an opportunity cost of zero reduce the share of capital in total economic costs. The volatility of the TPP index as well as the profitability index is considerably reduced. Nevertheless, the general conclusions drawn from these alternative estimates are the same as those reported above. 23


extremely large portion of their nominal economic costs are determined by macroeconomic conditions that are outside their control.

Nonetheless, the

substantial increase in the aggregate input price index since 1994, cannot be attributed to inflation effects. RPI inflation has been fairly stable, at approximately 2.7 percent per annum since 1994. Instead, the increase of the aggregate input price index during this period is attributed to two other factors. The first is a sharp increase in the water industry’s capital cost index (CCI), which increased the replacement cost of capital by 30.5 percent between 1994 Figure 1: Aggregate WASC Output and Input Price Indices

150 140 130

1990=100

120 QualityAdjusted Output Price

110 100 90

Aggregate Input Price

80 70 60 50 40 1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

1990=100

Figure 2: Aggregate WASC TFP ,TPP, and TFP/TPP

180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 1985

TFP

TPP

Profitability Index

1986

1987

1988

1989

1990

1991

and 1999 (Waterfacts, 2000).21

1992

1993

1994

1995

1996

1997

1998

1999

The second is the increasing burden of

corporation taxes faced by the WASCs after 1995. Earlier in the privatisation period the companies benefited from tax allowances that sharply reduced their tax

24


bills. This implies a higher nominal base rental rate of capital to provide a 6 percent real post tax rate of return. As trends in the TPP index are almost entirely determined by trends in the aggregate input price index, neither the WASCs’ management nor the regulator has been primarily responsible for the industry’s total price performance. Nonetheless, the data on TPP presented in Table 1 and Figure 2 are still of interest because they highlight a substantial improvement in price performance during the first period of privatisation (1990-95), and a substantial decline during the second period (1995-1999).22 Moreover, as evidenced by the decomposition of changes in the economic profitability index at the bottom of Table 1, trends in TPP are the primary determinant of trends in the index. During the 1990-95 period, when the economic profitability index increased by an average annual rate of 10.7 percent, TFP growth was only responsible for 19 percent of this change. Improvements in TPP accounted for 81 percent. In contrast, a decline in WASC TPP was responsible for 115 percent of the average annual decline in the economic profitability index during 1995-99. Arguably, given the volatility of the TPP index, it is perhaps appropriate to focus on the entire post-privatisation period in order to gain a more accurate characterisation of the determinants of changes in the aggregate WASC economic profitability index. The estimates suggest that 55 percent of the average annual increase in economic profitability over the 1990-99 period was attributable to improved TFP and 45 percent was attributable to an improvement in the total price performance of the industry. This is an interesting result because, while the research found that

aggregate WASC productivity growth did not increase

significantly after privatisation, TFP is still the largest single determinant of increased profitability in the post-privatisation period. By contrast, slightly under a half of the net 29.8 percent increase in the economic profitability index between 1990 and 1999 can still be attributed to improved price performance. This suggests that the K price limits have not been particularly onerous, taking the post-privatisation period as a whole.

21

The DETR’s COPI index, which is representative of economy wide construction costs increased by 28.8 % over the same period. Thus, the increases in capital investment costs faced by the WASCs are similar to those faced by certain other sectors of the economy at this time. 22 Alternative time period estimates demonstrate the same trends as those reported in the text. However, given the volatility of the aggregate input price index, and hence the TPP index, the improvement in TPP during the first period of privatisation (1989-94) is not statistically significant, although its decline is during the later period. 25


The Performance Results: Company Level Given that the basic source of data for this study has been company specific data, it is straightforward to provide disaggregated estimates for each of the WASCs. As discussed above, the EA’s decision to aggregate Northumbrian and Yorkshire’s operating area into the North East EA Region and South West and Wessex’s operating area into the South West EA Region, does create some difficulties in identifying a disaggregated sewerage treatment quality index for these companies. It is possible to recover an index of bathing water quality for these four WASCs. But it is necessary to construct the total sewerage treatment quality index by assuming that the river quality of each is equivalent to that of the EA region in which it is located. As this assumption may provide a somewhat skewed impression of these companies’ economic performance, alternative estimates of the aggregated economic performance of the North East and South West EA Regions have also been calculated. For comparison purposes, these estimates, as well as company level estimates for all 10 WASCs, are reported in Tables 2, 3, and 4. Close scrutiny of these tables reveals that, as we should expect, the generally disaggregated performance of the WASCs closely mirrors their aggregated performance. There are nonetheless some substantial individual variations. For example, Anglian Water and Dwr Cymru (Welsh Water) achieved particularly high rates of non-capitalised labour productivity growth during 1995-99, results that can be attributed to their exceptionally high level of labour shedding. Also, North West Water stands out, as its 157 percent improvement in non-capitalised labour productivity growth since privatisation is the largest of all the WASCs. At the opposite end of the spectrum, the estimates suggest that, since 1990, South West Water has only achieved a 20.4 percent increase in non-capitalised labour productivity growth. The total factor productivity estimates in Table 2 reinforce the earlier conclusion that trends in labour productivity and TFP growth are not necessarily contemporaneous. While Dwr Cymru (Welsh Water) is estimated to have achieved the second highest total increase in non-capitalised labour productivity growth since privatisation, its average TFP growth for the post-privatisation period is below that of the average WASC. In contrast, Southern Water has the highest rate of TFP growth during the post-privatisation period, although its noncapitalised labour productivity growth was below average. From an economic perspective these results are particularly interesting. They suggest that the WASCs have scope for substantial factor substitution when they choose how to 26


improve their quality-adjusted output. The results also confirm that employing a partial productivity measure, such as labour productivity, gives a biased estimate of the WASCs’ overall productivity performance, in what is a capital intensive industry. The disaggregated input usage data presented in Table 3 reveals that South West Water’s particularly low average non-capitalised labour productivity and TFP growth are not just an artefact of assuming that its river water quality is equivalent to that of the South West EA Region. Its performance can instead be attributed to its relatively high input usage.

Thus, if consideration of

Northumbrian Water is excluded, for which input usage growth figures are skewed because of its 1996 acquisition of the WOC, North East Water, South West Water is the only company not to have reduced its non-capitalised labour employment between 1990 and 1999. More significantly, South West Water also had the highest average rate of capital expansion of all the WASCs except Northumbrian. Given these two factors, South West’s aggregate input usage expanded by 29.4 percent after 1990, while the average WASC’s capital expansion was only 15.2 percent. Nevertheless, as the WASC’s high capital intensity implies that the aggregate input index’s trend is primarily determined by capital usage, these results are interpreted as suggesting that South West Water is a particularly extreme example. It highlights the strong negative relationship between each WASC’s TFP growth and its rate of capital expansion.

The

disaggregated findings further suggest that the rate of WASC capital expansion has not necessarily resulted in a compensating increase in quality-adjusted output. The disaggregated TPP and economic profitability estimates summarised in Table 4 reveal the same overall trends in price performance and economic profitability that were identified for the aggregated WASCs. Analysis of changes across all ten WASCs suggests that a statistically significant improvement in price performance and economic profitability occurred during 1990-95. Subsequently, the tighter economic regulation of the 1995-99 period may have contributed to a statistically significant decline in both price performance and economic profitability. However, despite the decline in economic profitability after 1995, every WASC's economic profitability in 1999 remained higher than it had been in 1990, with the average WASC's economic profitability index remaining 28.9% higher. The estimates of the contribution of TPP and TFP growth to changes in the economic profitability index do however reveal considerable variation in the determinants of economic profitability (Table 4). Focusing on the entire post-

27


privatisation period, the contribution of TFP to changes in economic profitability ranged from a low of –21 percent for South West to a high of 232 percent for North West.

By coincidence, both these companies’ economic profitability

indices increased by 1.2 percent per annum during the post-privatisation period, which is the lowest average increase in economic profitability.

The results

suggest that North West’s relatively low increase in economic profitability occurred because its relatively large gains in TFP offset declines in its TPP. By

Table 2: Productivity Performance Indices: By Company

Non-Capitalised Labour Productivity

Anglian North West Water Severn Trent Southern Thames Welsh North East EA Region Northumbrian Yorkshire South West EA Region South West Wessex Average WASC

Average Annual % Change1 2 95-993 90-992 85-90 90-95

1985

1995

1999

5.7% 3.4% 3.5% 10.2% 0.3% 6.4% 4.7% 5.8% 3.9% 5.2% 5.4% 3.1%

4.4% 11.1% 6.3% 3.7% 7.3% 1.1% 5.5% 2.3% 6.2% 2.7% 0.2% 6.1%

10.3% 9.7% 9.9% 5.1% 4.1% 19.4% 5.5% 4.7% 6.9% 5.4% 4.3% 6.2%

7.0% 10.5% 7.9% 4.3% 5.8% 9.2% 5.5% 3.3% 6.5% 3.9% 2.1% 6.2%

75.2 84.2 83.9 60.1 98.7 72.6 79.0 75.0 82.3 77.2 76.5 85.6

124.7 174.2 136.8 120.5 143.7 105.7 131.4 112.1 136.3 114.3 101.2 135.8

188.2 257.1 203.3 147.6 169.0 229.3 163.9 135.0 179.5 142.1 120.4 174.0

4.8%

4.9%

8.0%

6.0%

79.4

129.1

180.3

Index 1990=100

Total Factor Productivity (TFP) Average Annual % Change1 Anglian North West Water Severn Trent Southern Thames Welsh North East EA Region Northumbrian Yorkshire South West EA Region South West Wessex Average WASC

Index 1990=100

85-90

90-952

95-993

90-992

1985

1995

1999

3.0% 2.4% 0.4% 5.1% 0.7% 1.8% 1.0% 3.6% -0.3% 3.4% 3.0% 2.1%

1.9% 2.0% 2.2% 4.1% 3.5% 0.2% 2.2% 3.5% 1.4% 0.7% -0.8% 2.4%

-0.4% 3.5% 0.9% 2.2% -0.3% 2.5% -0.1% -2.0% 0.4% 0.6% 0.5% 0.3%

0.9% 2.7% 1.7% 3.2% 1.8% 1.3% 1.2% 1.0% 1.0% 0.7% -0.2% 1.5%

86.2 88.9 98.0 77.4 96.7 91.5 95.1 83.5 101.6 84.2 86.1 90.1

109.8 110.5 111.8 122.7 119.2 101.2 111.5 119.0 107.3 103.7 96.0 112.9

108.3 127.2 116.1 133.9 118.0 112.0 111.0 109.9 109.2 106.2 97.8 114.3

2.2%

2.0%

0.8%

1.4%

90.0

111.1

114.7

1. Growth figures are estimated as log differences. 2. Avg. WASC growth figures in bold are significantly different from those in 1985-90 at 10%, bold italic figures at 5% 3. Avg. WASC growth figures in bold are significantly different from those in 1990-95 at 10%, bold italic figures at 5%

28


Table 3: Selected Input Usage Indices: By Company Capital Stock Average Annual % Change1 Anglian North West Water Severn Trent Southern Thames Welsh North East EA Region Northumbrian Yorkshire South West EA Region South West Wessex Average WASC

Index 1990=100

85-90

90-952

95-993

90-992

1985

1995

1999

0.3% 0.2% 0.1% 0.1% 0.2% 0.7% -0.1% -0.9% 0.1% 0.7% 0.8% 0.7%

2.2% 1.4% 1.1% 0.4% 0.7% 1.8% 1.0% 1.4% 1.0% 2.3% 4.0% 1.1%

1.0% 0.7% 1.0% 0.9% 0.1% 1.5% 3.4% 10.1% 1.2% 1.2% 1.8% 0.8%

1.6% 1.1% 1.1% 0.6% 0.4% 1.7% 2.1% 5.3% 1.0% 1.8% 3.0% 1.0%

98.5 98.9 99.7 99.7 99.0 96.7 100.7 104.5 99.6 96.4 96.0 96.8

111.4 107.4 105.8 102.1 103.4 109.6 105.4 107.1 104.9 112.1 122.3 105.7

115.8 110.5 110.3 105.7 104.0 116.5 121.0 160.5 109.9 117.8 131.5 109.1

0.2%

1.5%

1.9%

1.6%

98.9

108.0

117.4

Non-Capitalised Labour Employment Average Annual % Change1

Index 1990=100

85-90

90-952

95-993

90-992

1985

1995

1999

Anglian North West Water Severn Trent Southern Thames Welsh North East EA Region Northumbrian Yorkshire South West EA Region South West Wessex

-1.9% -0.7% -3.4% -4.6% 0.8% -4.0% -3.6% -2.5% -4.0% -0.7% -1.3% 0.1%

-0.3% -8.0% -2.6% 0.8% -3.3% 1.3% -2.2% 2.0% -3.6% 0.6% 3.9% -2.9%

-9.8% -5.4% -8.6% -2.4% -4.3% -15.8% -2.7% 2.7% -5.8% -4.5% -3.6% -5.5%

-4.5% -6.8% -5.3% -0.7% -3.7% -6.3% -2.4% 2.3% -4.6% -1.7% 0.6% -4.0%

109.8 103.3 118.6 125.6 96.0 122.4 119.7 113.3 121.9 103.5 106.9 99.7

98.6 67.0 87.7 103.9 84.9 106.6 89.6 110.4 83.5 102.9 121.3 86.6

66.7 54.0 62.3 94.2 71.4 56.7 80.3 123.1 66.1 86.0 105.2 69.6

Average WASC

-2.1%

-1.3%

-5.8%

-3.1%

111.8

95.0

76.9

Aggregate Input Usage Index Average Annual % Change1 Anglian North West Water Severn Trent Southern Thames Welsh North East EA Region Northumbrian Yorkshire South West EA Region South West Wessex Average WASC

Index 1990=100

85-90

90-952

95-993

90-992

1985

1995

1999

0.9% 0.4% -0.3% 0.5% 0.4% 0.6% 0.1% -0.3% 0.2% 1.1% 1.0% 1.1%

2.3% 1.1% 1.4% 0.4% 0.5% 2.1% 1.1% 0.8% 1.2% 2.5% 4.9% 0.8%

0.9% 0.8% 0.4% 0.5% 0.0% 1.0% 2.9% 9.4% 0.6% 0.4% 0.3% 0.4%

1.6% 1.0% 1.0% 0.4% 0.3% 1.6% 1.9% 4.6% 0.9% 1.6% 2.9% 0.6%

95.8 97.9 101.5 97.6 98.0 97.2 99.4 101.7 98.8 94.9 94.9 94.8

112.0 105.6 107.3 101.9 102.3 111.3 105.5 104.0 106.0 113.5 127.9 104.1

115.9 109.2 109.0 103.8 102.3 116.0 118.5 151.4 108.6 115.2 129.4 105.9

0.4%

1.5%

1.4%

1.5%

97.8

108.2

115.2

1. Growth figures are estimated as log differences. 2. Avg. WASC growth figures in bold are significantly different from those in 1985-90 at 10%, bold italic figures at 5% 3. Avg. WASC growth figures in bold are significantly different from those in 1990-95 at 10%, bold italic figures at 5%

29


Table 4: Price Performance and Economic Profitability: By Company Total Price Performance (TPP) Average Annual % Change1 Anglian North West Water Severn Trent Southern Thames Welsh North East EA Region Northumbrian Yorkshire South West EA Region South West Wessex Average WASC

Index 1990=100

85-90

90-952

95-993

90-992

1985

1995

1999

2.5% 1.1% -0.6% 4.5% 2.7% 1.6% 1.7% 6.5% -0.7% 3.4% 3.4% 1.6%

-8.0% -9.1% -8.5% -4.9% -7.9% -9.2% -7.3% -7.2% -7.7% -9.9% -11.8% -7.2%

5.8% 14.9% 11.1% 5.8% 5.6% 6.8% 3.6% 4.9% 3.3% 7.4% 11.5% 3.3%

-1.9% 1.5% 0.2% -0.2% -1.9% -2.1% -2.5% -1.8% -2.8% -2.2% -1.4% -2.6%

88.1 94.6 103.0 79.9 87.3 92.1 91.7 72.2 103.6 84.6 84.6 92.2

67.0 63.3 65.5 78.3 67.5 63.0 69.4 69.7 68.2 60.9 55.5 69.7

84.5 114.7 102.0 98.6 84.4 82.7 80.0 84.7 77.8 81.9 88.1 79.5

2.3%

-8.1%

7.3%

-0.9%

89.8

66.8

89.7

Economic Profitability Index (TFP/TPP) Average Annual % Change1

Index 1990=100

85-90

90-952

95-993

90-992

1985

1995

1999


0.5% 1.3% 1.0% 0.6% -2.1% 0.1% -0.7% -2.9% 0.4% 0.1% -0.4% 0.5%

9.9% 11.1% 10.7% 9.0% 11.4% 9.5% 9.5% 10.7% 9.1% 10.6% 11.0% 9.6%

-6.2% -11.3% -10.1% -3.6% -5.8% -4.3% -3.7% -6.9% -2.9% -6.8% -11.1% -3.0%

2.8% 1.2% 1.4% 3.4% 3.7% 3.4% 3.6% 2.9% 3.8% 2.9% 1.2% 4.0%

97.8 93.9 95.1 96.9 110.8 99.4 103.7 115.7 98.0 99.6 101.8 97.7

163.9 174.6 170.8 156.7 176.7 160.6 160.8 170.7 157.3 170.1 173.0 162.0

128.1 110.9 113.8 135.8 139.9 135.4 138.7 129.7 140.2 129.6 111.1 143.8

Average WASC

-0.1%

10.2%

-6.5%

2.4%

100.7

166.6

128.9

Share Decomposition of Changes in the Economic Profitability Index4


85-90 6.59 1.88 0.41 8.05 -0.33 13.68 -1.38 -1.24 -0.78 43.99 -8.18 4.45

Average WASC

21.96

TFP Contribution Share 90-95 95-99 90-99 0.19 0.06 0.32 0.18 -0.31 2.32 0.21 -0.09 1.15 0.46 -0.61 0.96 0.31 0.04 0.49 0.03 -0.59 0.37 0.23 0.03 0.32 0.33 0.29 0.36 0.15 -0.15 0.26 0.07 -0.09 0.23 -0.07 -0.04 -0.21 0.25 -0.11 0.37 0.20

0.12

0.60

TPP Contribution Share 85-90 90-95 95-99 90-99 -5.59 0.81 0.94 0.68 -0.88 0.82 1.31 -1.32 0.59 0.79 1.09 -0.15 -7.05 0.54 1.61 0.04 1.33 0.69 0.96 0.51 -12.68 0.97 1.59 0.63 2.38 0.77 0.97 0.68 2.24 0.67 0.71 0.64 1.78 0.85 1.15 0.74 -42.99 0.93 1.09 0.77 9.18 1.07 1.04 1.21 -3.45 0.75 1.11 0.63 -22.96

0.80

-1.12

1. Growth figures are estimated as log differences. 2. Avg. WASC growth figures in bold are significantly different from those in 1985-90 at 10%, bold italic figures at 5% 3. Avg. WASC growth figures in bold are significantly different from those in 1990-95 at 10%, bold italic figures at 5% 4. Decomposition shares sum to one if the economic profitability increases and negative one if it decreases.

30

0.40


contrast, South West’s relatively low increase in economic profitability was driven by declines in TFP which offset improvements in TPP. A further insight into the variation of the determinants of economic profitability performance is gained by considering the median contribution of TFP growth during the entire post-privatisation period. This was 43 percent across the eight EA regions and 37 percent across the ten WASCs.

These figures are

considerably lower than the 55 percent TFP contribution share for the aggregated WASCs, reported in Table 1. This suggests that for many WASCs, gains in TPP have played a larger role in their improved economic profitability than TFP gains. In fact, five out of eight EA regions and seven out of ten WASCs have TFP contribution shares for the entire post-privatisation period that are less than 50 percent. It is therefore reasonable to conclude that, while the majority of gains in aggregated WASC performance since privatisation are attributable to gains in TFP, most of the disaggregated WASCs can primarily attribute gains in their economic profitability to improvements in their price performance.

Conclusions This paper as analysed the impact of privatisation on the economic performance of the water and sewerage companies in England and Wales. It has provided industry and company-level estimates of productivity, price, and financial performance for both the 1985-90 pre-privatisation period and the 1990-99 postprivatisation years. These estimates have been used to test three hypotheses related to the predicted impact of privatisation with economic regulation on the WASCs’ economic performance. The three hypotheses were: H1: privatisation of the water and sewerage industry in 1989 did not lead to a statistically significant improvement in performance at either the industry or individual company levels. Estimates of labour and total factor productivity (TFP) growth, derived with quality adjusted output indices, suggest that while substantial reductions in labour usage have occurred, TFP growth has not improved relative to the preprivatisation period. Labour costs are a substantial component of operational expenditure and the finding of a statistically significant increase in labour productivity growth suggests a marked improvement in the operational efficiency of the WASCs after privatisation. This is consistent with the findings on unit costs of Ofwat (1999a, 2000a) and Europe Economics (1998) and is contrary to Shaoul (1997), who found no evidence of lower supply costs after privatisation. Taking into account capital inputs, however 31

– water is a capital intensive


industry – the performance results are more disappointing. Capital for labour substitution has been occurring during the 1990s, something consistent with the argument that, where economic regulation allows for a rate of return on investment at or above the cost of capital, incentives exist to over invest (Averch and Johnson, 1962). Regulation of the water industry in the 1990s seems to have failed to counteract this tendency. Indeed, new and more stringent drinking water and environmental regulations have been a partial cause of the extra investment. Another cause was a backlog of capital maintenance resulting from underinvestment while in state ownership. In this sense, at least some of the responsibility for the poor TFP results does not lie with the current water industry management. Nevertheless, on the evidence of this study, the hypothesis that privatisation has not resulted in a statistically significant increase in the overall productivity growth of the WASCs cannot be rejected. Hypothesis 2 was concerned more directly with the effects of economic regulation on performance and profitability: H2: Due to the implementation of RPI+K pricing regulation, increases in the economic profitability of the WASCs can be primarily attributed to improvements in their total factor productivity (TFP) performance, rather than an increase in the price of outputs relative to inputs (TPP). The evidence related to this hypothesis is mixed and is dependent on the level of aggregation. If we focus on the entire post-privatisation period studied, 1990-99, at the aggregate WASC level 55

percent of the increase in the economic

profitability index can be attributed to productivity gains; while 45 percent can be attributed to an increase in quality-adjusted output prices that outstripped increases in input costs. In contrast, company level estimates reveal that the median contribution of TFP growth to increased economic profitability was only 37 percent. Seven out of ten WASCs could attribute more than 50 percent of their increased economic profitability to improvements in their price performance. Therefore, the aggregate results are biased by companies, such as Southern Water, where 96 percent of all of the increase in economic profitability can be attributed to changes in productivity. The company level results, in particular, suggest that the regulatory régime has been lax taking the postprivatisation period as a whole, in the sense of allowing a substantial increase in economic profitability that cannot be attributed to productivity gains. Ofwat’s 1994/95 price review tightened the price caps that were originally put into place at privatisation. The third hypothesis was concerned with the effects of this review:

32


H3: Regulatory adjustment of the price cap in 1995 led to a statistically significant change in performance at the aggregate industry level and in each of the individual WASCs. The evidence reported strongly supports this hypothesis. The aggregated and company-level estimates demonstrate that the statistically significant gains in labour productivity growth were mainly attributable to the post 1995 period. Similarly, statistically significant reductions in input usage growth rates suggest that the WASCs were keener to seek efficiencies after 1995, even if this did not cause an increase in estimated TFP growth. Finally, a statistically significant decline in the aggregate and the company-level economic profitability indexes after 1995 confirms that the new price caps that were put into place had their intended effect. The results are consistent with the notion of more rigorous regulation after 1995 to prevent excess profitability growth, compared with the first five years of private ownership. The price caps set in 1989 were established by a government keen to ensure a successful sale of the industry. Only in 1995 did Ofwat act to reverse the lax régime put in place at privatisation.

33


References Alchian, A.A. (1965) ‘Some Economics of Property Rights’, Il Politico, vol.30, pp.816-829. Averch, H. and Johnson, L.L. (1962) ‘Behavior of the Firm under Regulatory Constraint’, American Economic Review, vol.52, Dec., pp.1052-1069. Bailey, E.E. (1973) Economic Theory of Regulatory Constraint, Lexington: D.C. Heath. Bhaskar, V. and Khan, M. (1995) ‘Privatization and Employment: A Study of the Jute Industry in Bangladesh’, American Economic Review, vol.85, no.1, March, pp.267-273. Bishop, M. and Green, M. (1995) ‘Privatisation and Recession – The Miracle Tested’, CRI Discussion Paper 10, London: Centre for the study of Regulated Industries. Bishop, M. and Thompson, D. (1992) ‘Regulatory Reform and Productivity Growth in the UK’s Public Utilities’, Applied Economics, vol.24, pp.1181-1190. Bosworth, D., Stoneman, P. and Thanassoulis, E. (1996) The Measurement of Comparative Total Efficiency in the Sewerage and Water Industry: An Exploratory Study, Birmingham: Office of Water Services (Ofwat). Boubakri, N. and Cosset, J.C. (1998) ‘The Financial and Operating Performance of Newly Privatized Firms: Evidence from Developing Countries’, Journal of Finance, vol.53, no.3, June, pp.1081-1110. Böss, D. (1991) Privatization: a theoretical treatment, Oxford: Clarendon Press. Bosworth, D. and Stoneman, P. (1998) An Efficiency Study for the Water Industry, Final Report, A Report to National Economic Research Associates, London: A Report, London: National Economic Research Associates. Boycko, M., Shleifer, A. and Vishny, R. (1996) ‘A Theory of Privatisation’, Economic Journal, vol.106, March, pp.309-319. Buchanan, J.M. (1978) ‘From Private Preferences to Public Philosophy: The Development of Public Choice’. In J.M. Buchanan, et.al., The Economics of Politics, IEA Readings, London: Institute of Economic Affairs. CIPFA (1989, and earlier) The Water Industry: United Kingdom Services and Costs, 1988, London: The Chartered Institute of Public Finance and Accountancy. Cox, A., Harris, L. and Parker, D. (1999) Privatisation and Supply Chain Management: On the effective alignment of purchasing and supply after privatisation, London: Routledge. CRI (1995, and earlier) The UK Water Industry: Water Services and Costs 1993/94, London: Centre for the Study of Regulated industries.

34


CRI (1997, and earlier) The UK Water Industry: Financial and Operating Review 1995/96, London: Centre for the Study of Regulated industries. CRI (1999) The UK Water Industry: Water Statistics 1998, Board, M, Boulding, P. and Cawthorn, I. (eds.), London: Centre for the Study of Regulated Industries. De Alessi, L. (1980) ‘The Economics of Property Rights: A Review of the Evidence’, Research in Law and Economics, vol.2, pp.1-47. De Fraja, G. (1993) ‘Productive Efficiency in Public and Private Firms’, Journal of Public Economics, vol.30, pp.15-30. Diewert (1976) ‘Exact and Superaltive Index Numbers’, Journal of Econometrics, vol. 11, pp. 5-45. DOE (1986) Privatisation of the Water Authorities in England and Wales, London: Department of the Environment. DWI (1999, and earlier) Drinking Water 1998: A Report by the Chief Drinking Water Inspector, London: Drinking Water Inspectorate, Department of Environment, Transport, and Regions. EA (1999, and earlier) Annual Reports & Accounts: 1998/99, London: Environment Agency, Department of Environment, Transport, and Regions. Europe Economics (1998) Water and Sewerage Industries General Efficiency and Potential for Improvement, Final Report, London: Europe Economics and N. Crafts for Ofwat. Galal, A., Jones, L., Tandon, P. and Vogelsang, I. (1994) Welfare Consequences of Selling Public Enterprises, New York; Oxford University Press. Grossman, S. and Hart, O.D. (1980) ‘Takeover Bids, the Free-rider Problem, and the Theory of the Corporation’, Bell Journal of Economics, vol.94, pp.691-719. Hicks, C (1998) Regulation of the UK Water Industry, London: Centre for the Study of Regulated Industries. Holmstrom, B. and Tirole, J. (1993) ‘Market Liquidity and Performance Monitoring’, Journal of Political Economy, vol.101, pp.678-709. Hunt, L.C. and Lynk, E.L. (1995) ‘Privatisation and Efficiency in the UK Water Industry: an Empirical Analysis’, Oxford Bulletin of Economics and Statistics, vol.57, no.3, pp.371-388. Jensen, M. and Meckling, W.R. (1976) ‘Theory of the Firm: Managerial Behaviour, Agency Costs and Ownership Structure’, Journal of Financial Economics, vol.3, pp.305-360. Jorgenson, D. W. and Griliches, Z. (1967) “The Explanation of Productivity Change’, Review of Economic Studies, vol. 34, pp. 249-83. Laffont, J. and Tirole, J. (1991) ‘Privatization and Incentives’, Journal of Law, Economics and Organization, vol.7, pp.84-105.

35


Littlechild, S.C. (1983) Regulation of British Telecommunications’ Profitability, London: HMSO. Littlechild, S.C. (1986) Economic Regulation of Privatised Water Authorities, London: HMSO. Lynk, E.L. (1993) ‘Privatisation, Joint Production and the Comparative Efficiencies of Private and Public Ownership: the UK Water Industry Case’, Fiscal Studies, vol.14, no.2, pp.98-116. Martin, S. and Parker, D. (1997) The Impact of Privatisation: Ownership and corporate performance in the UK, London: Routledge. Megginson, W.L., Nash, R.C. and van Randenborgh, M. (1994) ‘The Financial and Operating Performance of Newly Privatized Firms: an International Empirical Analysis’, Journal of Finance, vol.49, no.2, pp.403-452. NERA (1996) The Performance of the Privatised Industries: Efficiency, vol.3, A Report for the Centre for Policy Studies, London: National Economic Research Associates/ Centre for Policy Studies. Newbery, D. and Pollitt, M.G. (1997) ‘The restructuring and privatisation of the CEGB -–was it worth it?’, Journal of Industrial Economics, vol.45, no.3, pp.269304. NRA(1996, and earlier) Annual Reports & Accounts: 1998/99, National Rivers Authority. O’Connell Davidson, J. (1993) Privatisation and Employment Relations: The Case of the Water Industry, London: Mansell Publishing. Ofwat (1995) Competition in the Water Industry: Inset Appointments and their Regulation, July, Birmingham: Office of Water Services. Ofwat(1996) Report on levels of service for the water industry in England and Wales: 1995-6 , Birmingham: Office of Water Services. Ofwat (1998a) A benchmarking study of the England and Wales Water Companies and Sydney Water Corporation Ltd for 1996/97’ RD 14/98, Birmingham: Office of Water Services. Ofwat (1998b) Prospects for Prices, Birmingham: Office of Water Services. Ofwat (1999a) Final Determinations: Future water and sewerage charges 200005, Birmingham: Office of Water Services. Ofwat(1999b) 1999 July Returns for the Water Industry in England and Wales, CD-ROM, Birmingham: Office of Water Services. Ofwat (2000) 1998-1999 Report on Water and Sewerage Service Unit Costs and Relative Efficiency, Birmingham: Office of Water Services. Ogden, S. (1994) ‘The Reconstruction of Industrial Relations in the Privatized Water Industry’, British Journal of Industrial Relations, vol.32, pp.67-84.

36


O’Mahony. M. (1998) Britain’s Competitive Performance: an analysis of productivity, London: National Institute for Economic and Social Research. Parker, D. (1997) ‘Price cap regulation, profitability and returns to investors in the UK regulated industries’, Utilities Policy, vol.6, no.4, pp.303-315. Parker, D. (1999) ‘The Performance of BAA Before and After Privatisation: a DEA Study’, Journal of Transport Economics and Policy, vol.33, part 2, pp.133145. Parker, D. and Martin, S. (1997) The Impact of Privatisation: Ownership and Corporate Performance in the UK, London: Routledge. Parker, D. and Wu, H.L. (1998) ‘Privatisation and Performance: a Study of the British Steel Industry under Public and Private Ownership’, Economic Issues, vol.3, part 2, pp.31-50. Pint, E. (1991) ‘Nationalization vs. Regulation of Monopolies: The Effects of Ownership on Efficiency’, Journal of Public Economics, vol.44, no.2, pp.131164. PSIRU (1999) Employment and profit margins in UK water companies: implications for price regulation proposals, report written by D.Hall and E. Lobina, London: Public Services International Research Unit. Sappington, D.E. and Stiglitz, J.E. (1987) ‘Privatization, Information and Incentives’, Journal of Policy Analysis and Management, vol.6, no.4, pp.567-82. Shaoul, J. (1997) ‘A Critical Financial Analysis of the Performance of Privatised Industries: The Case of the Water Industry in England and Wales’, Critical Perspectives on Accounting, vol.8, pp.479-505. Smith, A.D., Hitchens, D.M. and Davies, S.W. (1982) International Industrial Productivity, Cambridge: Cambridge University Press. Tullock, G. (1976) The Vote Motive, London: Institute of Economic Affairs. Vass, P. (1997) ‘The Methodology for Resetting X’, in (ed.) P. Vass, Regulatory Review 1997, London: Centre for the study of Regulated Industries. Vickers, J and Yarrow, G. (1988) Privatization; An Economic Analysis, Cambridge, Mass.: MIT Press. Vickers, J. and Yarrow, G. (1991) ‘Economic Perspectives on Privatization’, Journal of Economic Perspectives, vol.5, no.2, pp.111-132. Water UK (2000) Waterfacts 2000, London: Water UK. Waters II, W.G. and Tretheway, M.W. (1999) ‘Comparing Total Factor Productivity and Price Performance: Concepts and Applications to the Canadian Railways’, Journal of Transport Economics and Policy, vol.33, part 2, pp.209220. Willner, J. (1996) ‘A Comment on Bradburd’, “Privatisation of Natural Monopolies”’, Review of Industrial Organization, vol.11, pp.869-882.

37


WSA (1997, and earlier) Waterfacts, London: Water Services Association of England and Wales.

38


39