document de travail working paper

DOCUMENT DE TRAVAIL WORKING PAPER N°09-03.RS RESEARCH SERIES

EXPLORATORY STUDY ON THE PRESENCE OF CULTURAL AND INSTITUTIONAL GROWTH SPILLOVERS Charles PLAIGIN

DULBEA l’Université Libre de Bruxelles Avenue F.D. Roosevelt, 50 - CP-140 l B-1050 Brussels l Belgium

Exploratory study on the presence of cultural and institutional growth spillovers C. Plaigin Université Libre de Bruxelles - ULB D e p a rtm e nt O f A p p lie d E c o n o m ic s - D U L B E A L a b o ra to ire d e M é th o d o lo g ie d u Tra ite m e nt d e s D o n n é e s - L M T D

Abstract This paper shows an exploratory study on the use of institutional or cultural proximities on growth model. The justi…cation dealing with spillovers problems in growth models with such connection matrices is detailed. The empirical study answers two questions. First, do institutional and cultural weights are really di¤erent to geographical ones? Second, do these proximities have an impact on growth estimation? Analyze concludes that institutional and cultural proximities do di¤ers from geographical proximities. Furthermore, both show autocorrelation with growth. Finally, the spatial estimates point out that institutional weights have a similar impact to geographical weights in growth model. Cultural ponderation seems to be insigni…cant for growth spillovers problems.

1

Introduction

Empirical studies about growth and convergence processes have been the topic of many great authors’ studies. Barro and Sala-i-Martin (1992), Mankiw, Romer and Weil (1992) or Nazrul (1995), among others, have made some of the greatest contributions in this literature. All of these studies have pointed out many di¤erences in the production and the growth values amoung countries. The models are explaining those variations by two sources: physical and human capital. Nevertheless the variation of the residuals may suggest omitted variables in the speci…cation. Indeed, estimations made in the mentioned studies have all 0

Corresponding to [email protected] Author pays special thanks to Professor Enrique López-Bazo for his help and to Professor Stéphane Vigeant for his usefull comments.

1

considered countries as being isolated the one from the others. However, the reality is that interactions exist between observations made in di¤erent places. In other words, border countries are not impervious to economic performance of one another. The treatment of spillovers is still an unresolved issue in growth econometrics and cross-section correlation are important to determine the estimates. Since the works of MRW and Nazrul, spatial econometrics methods appeared in economic studies. The idea that growth observations on di¤erent localizations are mutually dependant has been modelized in estimations. Ertur and Thiaw (2000) showed that Mankiw, Romer and Weil cross sections data’s were biased by a spatial autocorrelation problem. Other studies on growth or productivity have used spatial econometrics methods (Baumont, Ertur and Le Gallo (2000), Beck and Gleditsh (2004), Arbia, de Dominicis and Piras (2005)). All of them have de…ned spatial dependence with respect to geographic distances or contiguity. Durlauf, Johnson and Temple (2005) argued that one can imagine many reasons for presence of spatial correlation in the error terms. Akerlof (1997) said that countries are perhaps best thought of as occupying some general socio-economic-political space de…ned by a range of factors. Durlauf and Al. (2005) pointed out that nothing in the empirical growth literature suggests that issue of long term development can be dissociated from the historical and cultural factors. So far, no consensus has emerged from the literature about the fact that geographic distance or contiguity constitutes the optimal way to weight observations. No testing methods exist yet to determine which spatial weights methods are optimal. A valuable generalization of the spillovers treatment is still under investigation. Interesting works from Hall and Jones (1999) are taking into account the social infrastructure in order to explain di¤erences among countries’ productivity per capita. In their work they used the institutions and government policies that provide the incentives for individuals and …rms in an economy. In this paper, the social infrastructure will take a formal de…nition as institutions or informal as culture. This paper will integrate formal and standard aspects of the social infrastructure de…ned by institutions as well as a more informal approach by integrating the concept of "culture" in the analysis. Indeed the proximity or the distance for countries pair with respect to these concepts can have an impact to the transmission of productivity and thus, growth. Some empirical studies have already tried to determine the impacts of cultural behavior on the economic performance. Duane Swank (1996) 2

starts from the idea that communitarian policies countries should have a better economic growth than non communitarian policies countries. The statistical analyze suggests that human capital investment and communitarian policies is empirically more powerful than the alternative. Marcus Noland (2003) has studied the relationship between culture, religion and economic performance. He …nds correlations between religious af…liations, the intensity of religious belief and cultural tendencies. Unfortunately, in his study there were no correlations between cultural measures and economic performance but religion is always correlated with economic performance in panel and cross-section estimates. Easterly and Levine (1997), Alesina and al. (2003), Masters and McMillan (2001) have showed that ethno-linguistic fractionalization has a negative impact on economic growth. Barro and McCleary (2004) have shown a strong relationship between economic growth and religiosity. They found out that religious beliefs stimulate growth but that church attendance tends to reduce it. Their interpretation is that religious beliefs may help to sustain aspect of individual behavior. The theoretical literature agrees with the hypothesis of cultural behaviors in‡uencing a country’s economic growth. Indeed cultural contents of a country are strongly related to the social organization, the networks and the trust of a population. Languages, religions, origin of law and G. Hofstede indexes (individualism, uncertainty avoidance, distance to power and masculinity) are determining the culture in this work. They are vectors of social rules and trust that can increase the organization of common action. Acemoglu, Johnson and Robinson (2004) as well as Hall and Jones (1999) though that cross countries di¤erences in output can be explained by the presence of di¤erent institutional system. They argue that the way the government a¤ects physical and human capital accumulation by taxes and rules has an indirect impact on economic productivity per capita. The way that redistributions are made a¤ects as well the economic performance. Another channel for the State to insure productivity of individuals and …rm is the way it guarantees the security inside the country. Indeed, crime level a¤ects human and physical capital accumulation. Another stylized fact over institutions is that democracy appears as a precondition for prosperity. Indeed, all the developed countries are running from democracy system. A few poor countries, as India does, sustain a democracy in a long run period. All emerging countries have chosen a democratic system as Chili, South Africa and South Korea. Indeed, living in an autocracy cannot insure individuals and …rms the private property and represents a capital accumulation-proof fence. Acemoglu, Johnson and Robinson (2004) founding suggests that low growth countries have chosen poor institutions as well, considering institutions 3

as endogenous. Giving for example east and West Germany, North and South Korea, China, Taiwan and Hong-Kong, all of these have started with same health level and have grown di¤erently by choosing di¤erent institutions. Anyway, this question is still under investigation, and this work will consider institutions as exogenous factor that could help to provide explanation on the variability of error terms in growth estimations. Another justi…cation of using cultural and institutional matrices of weights is the importance of such distances on trade and foreign direct investments (FDI). Indeed, these features can be seen as factors of capital accumulation. Trade e¤ects on growth level have been pointed out by several authors as Dollar and Kraay (2002) that are showing cross-country estimations that show out a great impact of trade level on growth. Romer and Frankel (1999) had large and robust e¤ects of trade on growth but slightly signi…cant. Other authors as Beugelsdijk, de Groot, Linders and Slangen (2007) or Linders, Burger and Van Oort (2008) have shown the importance of institutional and cultural distances in trade using gravity models. The social infrastructure distance are considering as a cost of trade and invest. Culture may have a positive role on trade and FDI as it stays at low level of di¤erences. But up to a certain level it increases the cost of trade and FDI by complicating interactions, make them prone to fail and hinder the development of trust. They point out that the cost of employees’ managing increases as well because of the interactions with local stakeholders as employees, unions, suppliers, government agencies. Institutions by legal, political and administrative systems determine the international attractiveness of a location. The aim of this work is to determine a cultural and a institutional weights matrices and to answer two questions : Does cultural and institutional proximities really di¤er from geographical ones? Does those type of proximities have an impact on growth estimations ? The …rst chapter presents the empirical growth model from Mankiw, Romer and Weil (1992). Then, an overview of spatial econometrics is showed. Second chapter present how cultural and institional matrices are build. Finaly, third chapter will show a exploratory study which point out that di¤erences between all notions of proximities really di¤er. Then, regressive spatial results evaluates the impacts of the di¤erent weights matrices.

4

2

The empirical model

2.1

Economic model

This paper is based on the neo-classical model of growth, called the Solow-Cass model. In this section, the development of the augmented model made by Mankiw, Romer and Weil (1992) by including human capital will be explained in detail. The aims of this study are to determine if a cultural proximity application of spatial distance could improve the MRW estimates and if there is a spillover of productivity per capita or growth due to a cultural proximity. The production function of a Solow-Cass model including human capital is Y (t) = K(t) H(t) (A(t)L(t))1

;

(1)

where Y is output, K is capital, L is labor and A the level of technology. H is the stock of human capital. MRW assumes that a constant part sk of capital and a constant fraction sh of human capital are invested. Following MRW’s paper (1992, pp 416-418), the evolution of the economy is determined by _ = sk y(t) k(t) _ = sh y(t) and h(t)

(n + g + )k(t); (n + g + )h(t);

(2) (3)

where y = Y =AL, k = K=AL, and h = H=AL. And ; n and g are respectively the capital depreciation rate, the population growth rate and the rate of technology. The economy converges to a steady state de…ned by

k =

s1k sh n+g+

and h =

sk s1h n+g+

!1=(1 1=(1

)

(4) )

:

(5)

Substituting those results in the production function and taking the log provides the equation below :

5

+

ln [y(t)] = ln A(0) + gt +

ln (sk ) +

1

ln (n + g + )

1

ln (sh ) :

1

(6)

This equation developped by MRW shows how income per capita depends on population growth and accumulation of physical and human capital. In addition to the Solow-Cass model, MRW (1992, pp 422-423) developped an endogeneous growth equation. y is the steady state level of income per e¤ective worker and its level is given by equation (6). Approximating around the steady state, the speed of convergence is given by d ln (y(t)) = [ln(y ) ln(y(t))] ; (7) dt where = (n + g + )(1 ) is the convergence rate. Equation (7) suggests a natural regression to study the convergence rate and implies that ln(y(t)) = (1 e t ) ln(y ) + e t ln(y(0)); (8) where y(0) is the income per e¤ective worker at an initial date. Substracting ln(y(0)) from both sides of equation (8) gives ln(y(t))

ln(y(0)) = (1

e

t

) ln(y )

(1

t

e

) ln(y(0)):

(9)

By equation (6), it is possible to replace ln(y ) and to obtain ln(y(t))

ln(y(0)) = (1

t

e

)

+(1

e

t

(1

e

t

(1

) )

(10)

ln (sk )

1

ln (sh )

1 +

1 t e ) ln(y(0)):

ln (n + g + )

(11) (12)

Equation (11) lead us to a classical estimation equation : Y = X + "; That is yi = 1 xi1 + 2 xi2 + ::: +

k xik

where i = 1; ::; N is the number of spatial units,

(13) (14)

+ "i =

1; 2 ; :::;

k

0

is the vector of parameters, and the "i ’s are the iid error terms with expectation equal to zero and variance equal to 2 : 6

2.2

Estimation methods

Observations over economic growth has oftenly been considered as independant from each others. In fact, those observations have ’spillovers’ problems : a problem of propagation. Indeed, good economic results in a country could have repercussions on its neighbours. Those spillovers are results of spatial autocorrelation between data taken in di¤erent places. In order to resolve those problems, we use statistical technics called spatial econometric. The aim of this method is to weight, by a function of proximity, observations in order to take into account those spillovers. In this section, I will present the two most used models and the way to interpret them. Those model are taken from Anselin (2003) and from Ertur and Thiaw (2005). 2.2.1

Spatial lag

The principle of the spatial lag model is to introduce a spatial lag on the dependant variable : Y = W Y + X + "; (15) where W is a NxN matrix representing a spatial relation between each entities, with the diagonal elements equal to zero and lines normalized to 1. " is the the iid error terms with expectation equals to zero and variance equal to 2 . By reducing we obtain Y = (I

W)

1

X + (I

W)

1

":

(16)

The presence of (I W ) 1 with X and the error term in equation (14) implies two spatial e¤ects. Firstly, there is a spatial multiplyer e¤ect which means that for every localisation, Y depends on the explaining variables from its own localities but also from observations in the other localities. Secondly, there is a spatial di¤usion e¤ect such that every exogen shock from a localisation a¤ects its own dependent observation (yi ) but also all dependent observations in the other localisations. 2.2.2

Spatial error

In this model, spatial autocorrelation is de…ned by the correlation between the interest variable’s value and the spatial units proximity where such variable is observed. The spatial model is based on the adoption of a spatial process in the error term ": Y = X + " whith " = W " + ;

(17)

where is the autoregressive coe¢ cient of the error term " and is the iid error term with expectation equal to zero and variance equal to 2 I: 7

By reducing this speci…cation : Y = X + (I

W)

1

:

(18)

The presence of (I W ) 1 is related only to the error term. That means that only an exogeneous shock in a spatial unit of the model has some repercussion on Y in all spatial units. Those repercussions are bigger when spatial units are close and smaller if spatial units are far from each other. 2.2.3

Weight matrices

The most important step in a spatial econometric study is to de…ne the dependence between localities. The …rst proposition have been made by Moran (1948) and Geary (1954). The methods were based on a contiguity relation between all the geographic entities present in the study. The notion of neighborhood was de…ned by a binary relation. Since that, other notions of spatial dependence have been used. Some are still based on a contiguity relation, other are taking into account the geographic distances between entities or exportation ‡ux between countries. Spatial contiguity is a binary relation between countries i and j. A relation between two localisations can be de…ned as cij =

1 0

if i and j are contigue otherwise,

(19)

and cij = 0 if i = j: This speci…cation implies that only neighborhood is important in the relation between two geographical entities. The weight matrix W is then de…ned as the matrix of contiguity where lines are normalized: cij wij = PN

j=1 cij

:

(20)

Another way to de…ne interaction between entities i and j is to take the real distances between theirs centers. De…ne dij as the distance between i and j: Then, the weight matrix W can be written with wij =

1 dij 1

PN

j=1

with wij = 0 for i = j:

8

, dij

(21)

The inverse distance is taken in order to de…ne a bigger relation for entities near to each other. In general, every decreasing function of distance which gives a bigger relation to close entities can be used : f (dij ) wij = PN , f (d ) ij j=1

with wij = 0 for i = j and f 0 (dij ) < 0.

9

(22)

3

Weights matrices

Most of the studies are based on geographical spatial dependence. The consideration of this type of dependence has provided good results in many of them but none of them has brought the proof that geographical distances are optimal. Furthermore, it could be interesting to test the impact of other type of interdependence on growth estimation. It is indeed possible that spillovers of growth or productivity can be channeled by social, cultural or political proximities. The problem is how to modelize such qualitative distances by the use of a weight matrix W in order to introduce it in spatial dependence econometrics estimations.

3.1

Geographical weights

The geographical matrix computed with the inverse of distances. Those distances are computed following the great circle formula1 which uses latitudes and longitudes of the most important cities or agglomerations (in terms of population).

3.2

Cultural Weights

The modelization of cultural weights is build with a tool called proximities between objects2 . Let us compare two vectors of observations relative to two entities i and j : xi ; xj where xi = (xi1 ; xi2 ; ... ,xik )0 and xj = (xj1 ; xj2 ; ..., xjk )0 with xik and xjk 2 f0; 1g for all k: Then four cases can appear : xik = xjk = 1; xik = xjk = 0; xik = 0 ; xjk = 1; xik = 1 , xjk = 0: Then, we can de…ne :

1

P a1 = Pk=1 I(xik = xjk = 1); P a2 = Pk=1 I(xik = 0; xjk = 1); P a3 = Pk=1 I(xik = 1; xjk = 0); P a4 = Pk=1 I(xik = xjk = 0):

The great-circle distance is the shortest distance between any two points on the surface of a sphere measured along a path on the surface of the sphere . 2 W.Härdle, L.Simar, "Applied Multivariate Statistical Analysis", Lehrstuhl für statistik wirtschaftswissenscha‡iche fakultät Humboldt, Universität zu Berlin, 2004.

10

Finally, proximity between objects can be computed as dij =

a1 + a4 ; a1 + a4 + (a2 + a3 )

(23)

where and are weight factors which can take di¤erent values. gives more weight to zero similiraties and gives more weight to disimilarities. In this work, Jaccard speci…cation with = 0 and = 1 and Tanimoto’s with = 1 and = 2 will be used3 . Finally, the elements of the weight matrix W are the proximities with lines normalized to 1 : dij wij = PN

j=1

with wii = 0

dij

(24)

The last point is how to build binary vectors of comparison between countries with qualitative or quantitative variables ? Let zik be a quantitative variable of comparison. An element of tik can be computed by tik =

1 0

_

if zik > z :k . Otherwise

(25)

Let zik be a qualitative variable with n modalities. Then, n binary variables can be created indicating the presence or absence of each modality. The vectors of comparison have been built with variables of religions, languages, legal origin and the Hofstede’s four dimensions of culturality for each country of the study. The religion variable had four modalities: percentage of Catholics, Protestant, Muslim or percentage of other religions. Each one of these modalities had been transformed in a binary variable. The legal system origin has been divided in …ve binary variables: British, French, socialist, German or origin. Languages were divided in seven categories: the six languages o¢ cially spoken at the ONU (English, French, Spanish, Arabic, Russian, Chinese) and a category of other languages. Each country could only have one legal system origin but it is possible to have several variables of o¢ cial languages equal to one. Finally, the four dimension of Hofstede estimations on cultural indexes were added4 : 3

W.Härdle, L.Simar, "Applied Multivariate Statistical Analysis", Lehrstuhl für statistik wirtschaftswissenscha‡iche fakultät Humboldt, Universität zu Berlin, 2004. 4 Hofstede G., (2001), "Culture’s consequences: Comparing values, Behaviors, Institutions and Organizations across Nations", Second edition, Thousand Oaks: Sage publications

11

Power Distance Index (PDI) that is the extent to which the less powerful members of organizations and institutions (like the family) accept and expect that power is distributed unequally. Individualism (IDV) on the one side versus its opposite, collectivism, that is the degree to which individuals are integrated into groups. Masculinity (MAS) versus its opposite, femininity, refers to the distribution of roles between the genders which is another fundamental issue for any society to which a range of solutions are found. Uncertainty Avoidance Index (UAI) deals with a society’s tolerance for uncertainty and ambiguity; it ultimately refers to man’s search for Truth. It indicates to what extent a culture programs its members to feel either uncomfortable or comfortable in unstructured situations. Comparing each country score to the average, these ones were changed to binary data. After doing this work, 90 countries were still used in the estimations for cultural weight. The cultural weights gives an equal ponderation to similar 1 in the vector of comparison and to disimilarities. No weight are given to similar 0 in the vectors ( = 0 and = 1).

3.3

Institutional weights

Institutional weights are build with the observations of Kaufmann and al. (2002) of governance indexes. The worldwide governance indicators are provided as a World Bank database for the years 1996-2007. For this study, only the observations of this index for the year 1996 were used. Kaufmann and al. identi…ed six dimensions of governance and de…ned them as5 : Voice and accountability : measuring perceptions of the extent to which a country’s citizens are able to participate in selecting their government, as well as freedom of expression, freedom of association, and a free media. 5

Kaufmann, Daniel, Kraay, Aart and Mastruzzi, Massimo,Governance Matters VII: Aggregate and Individual Governance Indicators, 1996-2007(June 24, 2008). World Bank Policy Research Working Paper No. 4654

12

Political stability and the absence of violence : measuring perceptions of the likelihood that the government will be destabilized or overthrown by unconstitutional or violent means, including politically-motivated violence and terrorism. Government e¤ectiveness : measuring perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government’s commitment to such policies. Regulatory quality : measuring perceptions of the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development. Rule of law : measuring perceptions of the extent to which agents have con…dence in and abide by the rules of society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence. Control of corruption : measuring perceptions of the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as "capture" of the state by elites and private interests. These indicators were constructs on the basis of factor analyses and give a pretty good description of all the matters related to institutionalism. In this paper, our interest is to build some institutional proximity in order to have a new weight matrix in spatial analyses. Firstly, it is necessary to compute an institutional distance for each country pair i and j: The Kogut and Singh index (1988) gives a tool to compute distances between the two vectors of indexes: " 6 # 1 X (Iki Ikj )2 IDij = 6 k=1 Vk Where Iki is the k th institutional index for the country i and Vk is te variance of the k th institutional index.

As these formula gives us distances, we need to inverse the result to reach a value of proximities beetwen two entities i and j. Then : Inst wij =

13

1 IDij

Finally, this weight matrix has to be row normalized to run into the spatial econometrics methods.

14

4

Results

4.1

Exploratory analysis

The aim of the exploratory study is to answer to several questions and justify the estimations using spatial methods. As the use of geographical weights has already been documented in literature, this paper will focus on the utilization of institutional and cultural weights. Thus, two question have to …nd an answer in this exploratory analyze, …rstly, does institutional and cultural weights really di¤er from geographical weights? Secondly, does the use of such matrices is justi…ed in growth estimations? The next section will show descriptive analyze of di¤erent weight type in order to be able to di¤erentiate geographical proximities and institutional or cultural ones. To answer the second question, an analyze of global and local autocorrelation on growth will be detailed before going to spatial estimates. 4.1.1

Matrices analyses

In order to see if institutional or cultural proximities are di¤erent from geographical ones, we have to compare which countries are institutionally or culturally near to a precise location. Figure 1 shows in green the 44 nearest countries to Spain and the red colored ones are the 45 countries under the median.

Fig. 1 Institutional proximities compare to Spain median The map clearly shows that nearest countries from Spain are Europeans ones but also a great part of North, Central and South America. Some countries of Africa are considering as near from Spain and Asian countries as South Korea, Japan, Malaysia or Philippines. Australia and 15

New Zealand complete the list of countries above the median of proximities from Spain. Do the nearest ones geographically are the nearest ones institutionally? Figure 2 shows the Spain quartiles distribution and Figure 3. These standing compare to Spanish’s institutional proximity average.

Fig. 2 Quartiles of Spain’s institutional proximities

Fig. 3 Instutional proximities compare to Spain Average The fourth quartile represent Canada, United States, Chili and Costa Rica for America, All European countries except Norway and Sweden, then Korea, Japan and Australia are composing the nearest institutional neighbors from Spain. Figure 3 shows that fourth quartile countries adding to New Zealand, Norway, Sweden, Costa Rica and Paraguay are above the average of institutional proximities. Figure 4, 5 and 6 are showing the same maps for institutional proximities from India.

16

Fig. 4 Institutional proximities compare to Indian median

Fig. 5 Institutional proximities - Indian quartiles

Fig. 6 Institutional proximities compare to Indian average The same results are obtained with India: the countries above the average are those of the fourth quartile of the institutional proximities with India. Looking further to explain this fact, Figure 7 shows the box plots for several countries in di¤erent continent in order to see the institutional proximity distribution shape. 17

Fig. 7 Box plot - Institutional proximities All countries’box plots have a very asymmetric shape. Furthermore, they do have outboxes high values. These values have a great impact on average that becomes really high compare to the median. It points out that one country have strong institutional relations with few countries. For example Spain main relations are Barbados, Japan and France. Peru has a really strong link with Turkey, Columbia, Mexico and India. The most in‡uent neighbors of India are Ecuador, Nicaragua, Brazil and Bangladesh. Do these conclusion are the same for cultural proximities? Figure 8 and 9 shows cultural proximities compare to the average for Spain and France respectively. The connections from cultural point of view are di¤erent from institutional ones. Indeed, France and Spain are well connected to South and Central America but North America stands under the average for both. France has good cultural connection with Africa (Fig. 9) and Spain.

18

Fig. 8 Cultural proximities compare to Spain average

Fig. 9 Cultural proximities compare to French average For cultural connections, median’s separation is quite similar to average one due to a more symmetric shape of cultural proximities distribution as shown in …gure 10. These facts have two meanings, …rst connections are above the average for half of the countries present in the study and second they do not have really strong connection with some countries.

19

Fig. 10 Cultural proximities box plots All the maps are showing connections quite di¤erent from geographical ones from institutional and cultural point of view. Indeed, those weights are connecting countries that would not be looking only from geographical distances. Furthermore, institutional and cultural weights do not reject connections for countries near from each other but does not connect systematically countries from the same regions. A last step to ensure that the three matrices are really di¤erent with each other’s is to compute correlation between average vectors for each type of distances. Taking weights matrices in absolute value and doing a average of weights for each line gives a vector with one average weight value for all the countries in the study. AverageG;I;C i

=

1 N

1

N X

G;I;C wik

k=1

Table 1 gives the correlation matrix between the three average vectors. Correlation between vectors is really low. Furthermore, correlation between institutional and cultural vectors is low and negative. Even if these correlations are made with countries’average values for each type of proximities, it is possible to conclude that the di¤erent weights do not have similar variation on their value. It means that when a country has a great amount of connection from one point of view, he could have very low level of connections from another one. This ensures even more that 20

the three ways to ponderate the spatial estimations are di¤erent to each other and will in‡uence results on their own way. Table 1 : Correlation between average weights vectors Correlations Geography Institution Culture Geography 1 Institutional 0,267 1 Cultural 0,232 -0,223 1 4.1.2

Impact on growth

The last step of the explorative study is to determine if the weights really have an impact on growth. These part show a global and local autocorrelation analyse. In order to tell if the is a presence of global autocorrelation over growth datas, two tests will be made : Moran’s I test of global autocorrelation and Geary’s C test of global autocorrelation. Moran’s I is testing the nul hypothesis of no global autocorrelation over a variable.

I=

N X

_

wij xi

N i;j=1 N S0 X

x

xi

xj _ 2

_

x With i 6= j

x

i=1

P P Where S0 = i j wij : The Moran’s I value has to be compare to its expected value E (I) = 1= (n 1) : Above these value, x has a positive autocorrelation due to spatial weight that means that neighbours have a similar value on x. Under, x has a negative global autocorrelation and neighbouring countries have unsimilar value. Moran’s I statistic follow a N (0; 1) and it is then possible to test I compare to its expected value. If it is signi…cant, it means that the nul hypothesis of no global autocorrelation is rejected. Table 2 shows the result of global autocorrelation Moran’s I tests for the Geographic, Institutional and cultural weights matrices. Table 2 : Moran’s I global autocorrelation tests Geography Institution Culture Moran’s I 0,215 0,173 0,087 E (I) -0,011 -0,011 -0,011 p-value 0,000 0,000 0,000 21

The results show that the three tests have a highly signi…cant positive global autocorrelation. That means that one country’s geographical, institutional or cultural neighbors have similar growth value. Comparing the Moran’s I value, Geographical en institutional weights have a best positive autocorrelation on growth values. Geary’s C is the second global autocorrelation test on x and has the following expression: N X

wij (xi xj ) N 1 i;j=1 C= N 2S0 X _ 2 xi x

With i 6= j

i=1

where the de…nition of the elements are the same as the Moran’s I expression. Geary’s C has to be compared to its expected value E (C) = 1. A signi…cant test on Geary’s C value means a positive global autocorrelation if C < 1 and negative global autocorrelation if C > 1: Table 3 shows the results for Geary’s C global autocorrelation test with geographical, institutional and cultural weights. Table 3 : Geary’s C global autocorrelation tests Geography Institution Culture Geary’s C 0,767 0,789 0,878 E (C) 1,000 1,000 1,000 p-value 0,000 0,000 0,000 Geary’s C result con…rm exactly those obtained with Moran’s I global autocorrelation tests. The three weights matrices have a positive autocorrelation impact on growth. Geographical and institutional connections seem to have the best impact on growth. A way to visualize the relation between one country growth and neighboring growth is to draw the Moran’s scatterplot for each of the connection matrix made for this study. Moran’s scatterplot (Anselin, 1996) plots the spatial lag of centered dependant observations W z with the centered values z in order to detect stability. The dependent values are centered. The plot is divided in four quadrants: the upside right called HH, the downside left LL, the upside left HL and …nally the downside right LH. The quadrant HH is the one in which the both values of z and W z are positives and refers to the cluster of countries with positive 22

spatial interaction. The quadrant LL is the one in which the both values are negative and refers to the cluster of countries with negative spatial interaction. The others two quadrants refer to atypical localization: a high value surrounded by low values or the inverse. Furthermore the slope of the regression line has the value of the Moran’s I statistic. Figure 11 shows the geographical weights Moran’s scatterplot. As showed, the scatterplot has a good positive shape and regression slope points out a real positive impact of geographical proximities on growth. Indeed, a majority of the scatters is in the HH and the LL quadrant and means that high (low) growth countries are surrounded by high (low) growth countries. Moran scatterplot (Moran's I = 0.215) g 1

5 48 38 31 716432 8 66 60 68 782324 65 40 41 3 30 79 3644 46 75 51 842943 45 73 54 2 83 28 67 13 86

14

Wz

72 26

0

87

63

80 81 62 89

21 17 61

55 74 34 27 10

39 90

4

47 1 56 11 70 85 69 57 3577 20 5219 7 22

37 185888 49

76

25 12 15

50 16 82 5342

59

9

33 6

-1 -3

-2

-1

0 z

1

2

3

Fig. 11 Geographical weights Moran’s scatterplot Figure 12 and 13 show the Moran’s scatterplot for the institutional and cultural weights, respectively. Both have a good positive relation and scatter positioned for the most of them in the HH and LL quadrants. It means that the institutional and cultural neighbors of high (low) growth countries are high (low) growth countries. Nevertheless, regression’s slope shows a di¤erent impact of cultural weights on growth. As institutional weights have the same impact level on growth, cultural one seems to be slight. These observations con…rm the results obtained with global autocorrelation tests.

23

Moran scatterplot (Moran's I = 0.173) G

Wz

1

0

-1 -3

-2

-1

0 z

1

2

3

Fig 12 Instutional weights Moran’s scatterplot Moran scatterplot (Moran's I = 0.087) G

Wz

1

0

-1 -3

-2

-1

0 z

1

2

3

Fig. 13 Cultural weights Moran’s scatterplot Finally, a last way to analyse the clusters of countries with growth is to make a Moran’s I local analyse. The expression of a local Moran’s Ii is : X zi Ii = N wij zj X j2J i zi2 =N i=1

24

where zi is normalized value of the interest variable (growth) for country i and Ji the connections with i’s neighbors. A positive (negative) value of Ii means that country i is in a clusters of similar (dissimilar) countries. A local Moran’s I is calculated for each countries and then tested at 10% signi…cance level. In our case, a map would not help to visualize the clusters as neighbors are not only geographical but also institutional and cultural. Table 4 and 5 show a list of countries by continents that have a signi…cant value of local Moran’s I. Normal font countries are signi…cant for institutional (cultural) proximities only, Italic font are only for geographical weights and bold font countries are signi…cant for both. Table 4 : Countries with signi…cant Moran’s I test for institutional weights compare to geographical weights Europe America Africa Asia Romania Dominican republic Madagascar China Belgium Guatemala Kenya Iran United kingdom Paraguay Algeria India Luxembourg United states Zimbabwe Pakistan Spain Ecuador Mauritius Bangladesh Portugal Venezuela Gambia Nepal Ireland Barbados Zambia Malaysia Norway Honduras Comoros Israel Poland Peru Togo Japan Bolivia Cameroon Korea Nicaragua Rwanda Hong-Kong Chile Niger Philippines Chad Sri Lanka Burundi Indonesia Nigeria Thailand South Africa Australia Mozambique Côte d’ivoire Most of the countries that are signi…cant with institutional weights are also with geographical ones. Very few countries have a signi…cant negative local test but those who have are di¤erent from geographical or institutional point of view. Countries that are in a cluster of dissimilar geographical neighbors are Chile, Barbados, Dominican Republic, Paraguay and Mauritius. Some are high growth countries surrounded by low growth neighbors as Chile. Others are low growth countries 25

that are near enough from United States to say that neighbors are high growth countries as Dominican Republic. Countries that have a negative local test from institutional weights are China, Iran, India, Dominican Republic, Pakistan, Bangladesh, Romania, Nepal and Malaysia. All of those countries are countries with growth level above the average with institutional connections with lower average growth countries. Table 5 : Countries with signi…cant Moran’s I test for cultural weights compare to geographical weights Europe America Africa Asia Poland Mauritius South Africa Indonesia Norway Dominican Republic Gambia China Germany Chile Mali Nepal Austria Paraguay Chad Hong-Kong Romania Panama Burundi Israel Ireland Peru Nigeria Malaysia Luxembourg Honduras Rwanda Pakistan United Kingdom Venezuela Ivory Coast Sri Lanka Guatemala Togo Korea Ecuador Niger Japan Bolivia Cameroon Thailand Nicaragua Comoros India Barbados Kenya Philippines Madagascar Iran Mozambique Bangladesh Zimbabwe Zambia For cultural proximities, very few countries have a negative local Moran’s I value. Mauritius, Dominican Republic, South Africa, Chile and Indonesia are countries that are clustered with culturally dissimilar countries. All these results con…rm that geographical, institutional and cultural weights are di¤erent from each other. Even if 36 countries have a locally signi…cant value in both, institutional and geographical or cultural and geographical, they do not have the same clusters. As proximities di¤er from the point view and as growth values show an autocorrelation either with geographical than institutional or cultural weights, it is interesting to go through a spatial analyze of growth models with those three type of ponderation.

26

4.2

Regressive results

The estimation is made on classical MRW model with human capital for growth and productivity per capita. All variables needed for the MRW have been found in the Penn World Table Version 6.1 initially build by Summers and Heston in 1988. Other data come from the World bank Database : GDF & WDI database. Investment share has been taken in average on the period 1981-2000. The human capital variable is the school enrollment percentage in secondary school taken in average between 1981 and 2000. Population growth rate has been taken in average on the same period augmented with rate of technology and of capital depreciation (g + ) of 0.05 as in the MRW model. The logarithm of each variable is used. After cleaning the data, 92 countries were remaining. Spatial results using cultural or institutional links may lead to endogeneity problems. Indeed, the spatial structure used in spatial econometrics methods has to be constant with respect to time period. The volatility of institutional or cultural data is, for sure, higher than geographic values. In this work, I assume that, for the set of countries I used, the volatility of institutional and cultural data is null. Endogeneity problems using such qualitative connection matrices are a big issue in spatial econometrics and have to be the subject of a speci…c study, which is not the aim of this paper. 4.2.1

Classical results

Table 1 summarizes the results obtained in the estimation of growth rate equation. Those results are quite similar of those obtained by other authors in the literature. As expected we obtained a positive impact of schooling and of investment share on the growth rate and a negative impact of population growth rate and also of initial level production. The value of the estimates is all signi…cant at the 1% level except for the constant which is not signi…cant. Those estimates are also a little bit smaller than the literature’s one.

27

Table 6 : Classical results on growth OLS Constant 0.13 ln(School) 0.19 (***) ln(n+g+ ) -0.11 (***) . ln(I/Y) 0.33 (***) ln(Yinit.) -0.18 (***) R2 0.44 Countries 92 (* * * ) : sig n i…c a nt a t 1 % , (* * ) : sig n i…c a nt a t 5 % , (* ) sig n i…c a nt a t 1 0 % .

4.2.2

Spatial results

All the results have been computed with Stata spatial econometrics toolboxes. Firstly, the results obtained with growth model are showed. Two types of estimations are presented: the spatial error model estimates and the spatial autoregressive model estimates. Those results are compared to classical OLS estimates. The three tests presented in the table are testing the null hypothesis =0. A signi…cant test means a signi…cant weights’impact on the estimates with the spatial error model. Table 7 : Growth model - Spatial OLS ML Geo Constant 0.13 0.07 ln(School) 0.19 (***) 0.19 (***) ln(n+g+ ) -0.11 (***) -0.08 (*) ln(I/Y) 0.33 (***) 0.26 (***) ln(Yinit.) -0.18 (***) -0.15 (***) . 0.66 (**) 2 R 0.44 0.47 Countries 92 90 Wald test 5.73 (**) Likelihood test 3.12 (*) Lagrange test 2.36

error estimates ML Inst ML Cult 0.21 0.08 0.20 (***) 0.21 (***) -0.10 (**) -0.10 (**) 0.34 (***) 0.33 (***) -0.19 (***) -0.18 (***) 0.31 0.10 0.47 0.47 90 90 0.38 0.02 0.37 0.02 0.16 0.01

(* * * ) : sig n i…c a nt a t 1 % , (* * ) : sig n i…c a nt a t 5 % , (* ) sig n i…c a nt a t 1 0 % .

Results obtained (Table 7) with spatial error model are similar to OLS estimation with institutional or cultural weights. Only geographical weights change the results and have a signi…cant estimate for . Anyway, robust test of Lagrange is insigni…cant for the three connection matrices thus spatial error model is not the best model for spatial study on growth. 28

Indeed Table 8 shows that robust test of Lagrange are signi…icative for geographical and institutional proximities with spatial lag model. The three tests presented in the table are testing the null hypothesis =0. A signi…cant test means a signi…cative weights’ impact on the estimates with the spatial autoregressive model. The intuition of the exploratory study stays true in the estimates: geographical and institutional weights have a similar impact on growth model with estimates around 0,5 and cultural is …nally insigni…cant in both, spatial error and spatial lag study. Table 8 : Growth model OLS Constant 0.13 ln(School) 0.19 (***) ln(n+g+ ) -0.11 (***) ln(I/Y) 0.33 (***) ln(Yinit.) -0.18 (***) . 2 R 0.44 Countries 92 Wald test Likelihood test Lagrange test -

Spatial autoregressive estimates ML Geo ML Inst ML Cult 2 0.08 0.35 0.07 0.18 (***) 0.21 (***) 0.20 (***) -0.07 (*) -0.08 (*) -0.10 (**) 0.26 (***) 0.33 (***) 0.32 (***) -0.17 (***) -0.23 (***) -0.18 (***) 0.76 (***) 0.54 (**) 0.21 0.54 0.51 0.47 90 90 90 17.58 (***) 5.38 (**) 0.27 10.01 (***) 4.17 (**) 0.25 13.40 (***) 3.84 (**) 0.24

(* * * ) : sig n i…c a nt a t 1 % , (* * ) : sig n i…c a nt a t 5 % , (* ) sig n i…c a nt a t 1 0 % .

29

5

Conclusion

Countries are connected to each others in many ways. The most obvious one is naturally geographical situation that is clearly a channel for growth spillovers. This study has showed other connections between countries, taking into account cultural and institutional similarities. Those type of proximities could indeed have an important role to play for growth propagation. Two countries that have similar institution or cultural behavior reduce the cost and the barriers to trade, FDI and economic collaboration. This paper has showed how to build institutional and cultural proximities. The obtained weight matrices are really di¤erent to each other’s and mainly from the geographical one. The spatial exploratory analyze have showed that growth is su¤ering global spatial autocorrelation problems with respect to geographical, institutional and cultural spatiality. Growth observation of one country depends on growth observation of its neighbors, institutional neighbors and cultural neighbors. The way that proximity a¤ects growth is positive for all weights matrices. Even in a local study, a very few countries are clustered with unsimilar countries. The spatial exploratory study shows that geographical and institutional weights have a great positive impact on growth when cultural similarities show slight but signi…cative e¤ects. This intuition is con…rmed with spatial estimations on growth MRW model. The spatial error model does not have a signi…cative robust test for its use and thus the spatial lag model is the best one for growth estimation. Assuming institutional and cultural proximities as constant over the time period, only geographical and institutional weights have signi…cative impact on growth estimations. Once again cultural proximities do not show a real impact on growth spillovers. To conclude, this paper shows that geographical and institutional proximities are di¤erent and both are channel of growth spillovers. Cultural weights do not seem to play an important role in growth propagation or maybe it may contains ambiguity in the way to in‡uence propagation. Institutional di¤erences seem to be a higher constraint for growth propagation. These results are opening new research perspectives: choosing the best spatial models for each spatial weights, geographical analyses of growth di¤usion, estimating relative importance of di¤erent weight matrices, going further by including geographical and institutional externalities on growth. 30

6

Bibliography

References [1] Acemoglu D., Johnson S. and Robinson J., 2004, Institutions as the fundamental cause of long-run growth, Handbook of economic growth. [2] Akerlof G., 1997, Social distance and economic decisions, Econometrica 65 (5), 1005-1027. [3] Alesina A., Devleeschauwer A., Easterly W., Kurlat S. and Wacziarg R., 2003, Fractionalization, Journal of economic growth 8 (2), 189-211. [4] Anselin L., 1988, Spatial econometrics : methods and models, Studies in Operational Regional Science, Kluwer Academic Publishers. [5] Anselin L., Bera A.K., Florax R. et Yoon M.J., 1996, Simple diagnostic tests for spatial dependence, Regional Science & Urban Economics, 26, 77-104. [6] Anselin L., 2003, Spatial externalties, spatial multipliers and spatial econometrics, International regional science review, 26, 153. [7] Arbia G., de Dominicis L. et Piras G., 2005, The relationship between regional growth and regional inequality in EU and transition countries : a spatial econometric approach, Department of Spatial Economics, Free University of Amsterdam. [8] Barro R.J. and Sala-i-Martin, 1992, Convergence, The Journal of Political Economy, Vol.100, N 2, 223-251. [9] Barro R.J. and McCleary R., 2004, Religion and economic growth, The Milken Institue review, second quarter 2004, 36-45. [10] Baumont C., Ertur C. et Le Gallo J., 2000, Convergence des régions européennes, une approche par l’économétrie spatiale, Université de bourgogne, Pôle d’Economie et de Gestion. [11] Beck N., Gleditsh K.S. et Beardsley K., 2004, Space is more than geography : using spatial econometrics in the study of political economy, New York university, department of political science university of California, San Diego, Paper of September 29. [12] Bénassy-Quéré A., Coupet M. and Mayer Thierry, 2007, Institutional determinants of foreign direct investment, The World Economy, pp 764-782. [13] Beugelsdijk S. et van Schaik T., 2003, Social Capital and regional economic growth, Faculty of economics, Tiburg University. [14] Beugelsgijk S., de Groot H., Linders G-J and Slangen A., 2007, Curtural distance, Institutional distance and international trade, Faculty of economics, Tiburg University. 31

[15] Clemente J., Marcuello C., Montañés A; et Pueyo F., 2004, Social spending, public generation of social capital and economic growth, Faculty of economics, University of Zaragoza. [16] Dollar D. and Kraay A., 2002, Intitutions, trade and growth, World Bank, April 2002. [17] Durlauf S.N., Johnson P.A. and Temple J.R.W., 2005, Growth econometrics, Handbook of economic growth, Volume 1A, pp 555663. [18] Durlauf S.N. and Quah D.T., 1998, The new empirics of economic growth, Centre for Economic Performance, discussion paper N 384. [19] Easterly W. and Levine R., Nov. 1997, Africa’s growth tragedy: policies and ethnic divisions, The Quarterly Journal of Economics. [20] Elhorst J.P., 2001. Panel data models extended to spatial error autocorrelation or a spatially lagged dependent variable, Research Report 01C05, University of Groningen, Research Institute SOM (Systems, Organisations and Management). [21] Ertur C. et Thiaw K., 2005, Growth and spatial dependence, the Mankiw, Rome rand Weil model revisited, Laboratoire d’économie et de gestion , Université de Bourgogne [22] Fingleton B. and López-Bazo E., 2006, Empirical growth model with spatial e¤ects, Papers in regional science, Vol. 85, N 2: [23] Frankel J. A. and Romer D., 1999, Does Trade Cause Growth?, American Economic Review, American Economic Association, vol. 89(3), pages 379-399, June. [24] Gapany H. et Friboulet J.-J., 1998, La dynamique culturelle du développement : concepts et mesures, Working paper N 307, Faculté des SES, Université de Fribourg. [25] Hall R.E., Jones C.I., 1999, Why do some countries produce so much more output per worker than others?, The Quarterly Journal of Economics, Vol. 114, N 1, 83-116. [26] Härdle W. et Simar L., 2003, Applied multivariate statistical analysis, MD tech –Methods and Data technologies. [27] Heston A., Summers R. and Aten B., Penn World Table Version 6.1, Center for International Comparisons at the University of Pennsylvania (CICUP), October 2002. [28] Hofstede G., 2001, Culture’s consequences: Comparing values, Behaviors, Institutions and Organizations across Nations, Second edition, Thousand Oaks: Sage publications [29] Holtz-Eakin D. et Schwartz A.E., 1995, Spatial productivity spillovers from public infrastructure : Evidence from State Highways, International Tax and Public Finance, 2:459-468. [30] Hsiao C., 2003, Analysis of panel data –second edition, Econometric 32

[31] [32]

[33]

[34]

[35]

[36]

[37] [38] [39]

[40] [41] [42]

[43] [44] [45] [46]

[47]

society monographs, Cambridge University Press. Islam N., 1995, Growth empirics : a panel data approach, The Quarterly Journal of Economics, Vol.110, N 4, 1127-1170. Kaufmann D, Kraay A. and Mastruzzi M., Governance matters VII: Aggregate and individual governance indicators 1996-2007, World Bank Institute, Global Gorvernance Program 4654, June 2008, Le Gallo J. and Ertur C., 2003, Exploratory spatial data analysis of the distribution of regional per capita GDP in europe 1980-1995, Paper in regional science, Vol 82, 175-201. Linders G-J, Burger M. and van Oort F. G., 2008, A rather empty world: the many faces of distance and the persistent resistance to international trade, Cambridge Journal of Regions, August 2008. López-Bazo E., Vayá E. and Artís M., 2004, Regional externalities and growth : evidence from european regions, Journal of regional science, Vol. 44, N 1; 43-73 Mankiw N.G., Romer D. et Weil D.N., 1992, A contribution to the empirics of economic growth, The Quarterly Journal of Economics, Vol.107, N 2, 407-437. Moreno R. et Trehan B., Dec 1997, Location and the growth of nations, Journal of Economic Growth, 2:399-418. Noland M., 2003, Religion, culture and economic performance, School of public policy and management, working paper 03-13. Ord K., 1975, estimation methods for models of spatial interaction, Journal of the American Statistical Association, Vol 70, N 349, 120126. Perret B., 2002, Indicateurs sociaux, Etat des lieux et perspectives, Les papiers du CERC, N 2002-01. Putnam, R.; Leonardi, R.; and Nanetti, R.Y. 1993. Making democracy work. Princeton, New Jersey: Princeton University Press. Plaigin C., 2006, Econometrie spatiale et distances non géographiques dans un modèle de croissance, PhD attendance thesis, Université Libre de Bruxelles. Quah D.T., 1996, Empirics for economic growth and convergence, European Economic Review, 40, 1353-1375. Rothstein, B., 2001, Social Capital in the Social Democratic Welfare State, Politics & Society, 29(2), pp. 207-241 Rappaport J., 1999, Local Growth Empirics, CID working paper N 23. Swank D., 1996, Culture, institutions and economic growth : theory, recent evidence and the role of communitarian polities, American journal of political science, Vol. 40, N 3, Aug. 1996, 660-679. Temple J. et Johnson P.A., 1998, Social capability and economic 33

growth, The Quarterly Journal of Economics, Vol. 113, N 3, 965990. [48] Verardi V., 2006, Applied microeconometrics with applications to political economics, Cours doctoral, Université Libre de Bruxelles.

34

DULBEA Working Paper Series 2009 N°.09-03.RS Charles Plaigin « Exploratory study on the presence of cultural and institutional growth spillovers » January 2009. N°.09-02.RS Thierry Lallemand and François Rycx « Are Young and old workers harmful for firm productivity » January 2009. N°.09-01.RS Oscar Bernal, Kim Oostelinck and Ariane Szafarz « Observing bailout expectations during a total eclipse of the sun » January 2009.

2008 N°.08-24.RS Leila Maron, Danièle Meulders and Sîle O’Dorchai « Parental leave in Belgium » November 2008. N°.08-23.RS Philip De Caju, François Rycx and Ilan Tojerow « Rent-Sharing and the Cyclicality of Wage Differentials » November 2008. N°.08-22.RS Marie Brière, Ariane Chapelle and Ariane Szafarz « No contagion, only globalization and flight to quality», November 2008. N°.08-21.RS Leila Maron and Danièle Meulders « Les effets de la parenté sur l’emploi », November 2008. N°.08-20.RS Ilan Tojerow « Industry Wage Differential, Rent Sharing and Gender in Belgium », October 2008. N°.08-19.RS Pierre-Guillaume Méon and Ariane Szafarz « Labor market discrimination as an agency cost », October 2008. N°.08-18.RS Luigi Aldieri « Technological and geographical proximity effects on knowledge spillovers: evidence from us patent citations », September 2008. N°.08-17.RS François Rycx, Ilan Tojerow and Daphné Valsamis « Wage differentials across sectors in Europe: an east-west comparison », August 2008. N°.08-16.RS Michael Rusinek and François Rycx « Quelle est l’influence des négociations d’entreprise sur la structure des salaires ? », July 2008. N°.08-15.RS Jean-Luc De Meulemeester « Vers une convergence des modèles ?

Une

réflexion à la lumière des expériences européennes de réforme des systèmes d’enseignement supérieur », July 2008. N°.08-14.RS Etienne Farvaque and Gaël Lagadec « Les promesses sont-elles des dettes ? Economie Politique des promesses électorales », June 2008.

N°.08-13.RS Benoît Mahy, François Rycx and Mélanie Volral « L’influence de la dispersion salariale sur la performance des grandes entreprises belges », May 2008. N°.08-12.RS Olivier Debande and Jean-Luc Demeulemeester « Quality and variety competition in higher education », May 2008. N°.08-11.RS Robert Plasman, Michael Rusinek and Ilan Tojerow « Les différences régionales de productivité se reflètent-elles dans la formation des salaires ? » April 2008. N°.08-10.RS Hassan Ayoub, Jérôme Creel and Etienne Farvaque « Détermination du niveau des prix et finances publiques : le cas du Liban 1965-2005 », March 2008. N°.08-09.RS Michael Rusinek and François Rycx « Rent-sharing under Different Bargaining Regimes: Evidence from Linked Employer-Employee Data », March 2008. N°.08-08.RR Danièle Meulders and Sîle O’Dorchai « Childcare in Belgium », March 2008. N°.08-07.RS Abdeslam Marfouk « The African Brain Drain: Scope and Determinants », March 2008. N°.08-06.RS Sîle O’Dorchai « Pay inequality in 25 European countries », March 2008. N°.08-05.RS Leila Maron and Danièle Meulders « Having a child: A penalty or bonus for mother’s and father’s employment in Europe? », February 2008. N° 08-04.RR Robert

Plasman,

Michael

Rusinek,

François

Rycx,

Ilan

Tojerow

« Loonstructuur in België », January 2008. N° 08-03.RS Caroline Gerschlager « Foolishness and Identity: Amartya Sen and Adam Smith », January 2008. N° 08-02.RS Michele Cincera « Déterminants des oppositions de brevets: une analyse micro-économique au niveau belge », January 2008. N° 08-01.RR Robert Plasman, Michael Rusinek, François Rycx, Ilan Tojerow « La structure des salaires en Belgique », January 2008.

2007 N° 07-22.RS Axel Dreher, Pierre-Guillaume Méon and Friedrich Schneider « The devil is in the shadow Do institutions affect income and productivity or only official income and official productivity », November 2007. N° 07-21.RS Ariane Szafarz « Hiring People-like-Yourself: A Representation of Discrimination on the Job Market », November 2007. N° 07-20.RS Amynah Gangji and Robert Plasman « Microeconomic analysis of unemployment in Belgium », October 2007.

N° 07-19.RS Amynah Gangji and Robert Plasman « The Matthew effect of unemployment: how does it affect wages in Belgium », October 2007. N° 07-18.RS Pierre-Guillaume Méon, Friedrich Schneider and Laurent Weill « Does taking the shadow economy into account matter to measure aggregate efficiency», October 2007. N° 07-17.RS Henri Capron and Michele Cincera « EU Pre-competitive and Near-the-market S&T Collaborations », October 2007. N° 07-16.RS Henri Capron « Politique de cohésion et développement régional », October 2007. N° 07-15.RS Jean-Luc De Meulemeester « L’Economie de l’Education fait-elle des Progrès ? Une Perspective d’Histoire de la Pensée Economique », October 2007. N° 07-14.RS Jérôme de Henau, Leila Maron, Danièle Meulders and Sîle O’Dorchai « Travail et Maternité en Europe, Conditions de Travail et Politiques Publiques », October 2007. N° 07-13.RS Pierre-Guillaume Méon and Khalid Sekkat «Revisiting the Relationship between Governance and Foreign Direct Investment», October 2007. N° 07-12.RS Robert Plamsan, François Rycx and Ilan Tojerow « Wage Differentials in Belgium : The Role of Worker and Employer Characteristics », October 2007. N° 07-11.RS Etienne Farvaque, Norimichi Matsueda and Pierre-Guillaume Méon« How committees reduce the volatility of policy rates », July 2007. N° 07-10.RS Caroline Gerschlager «Adam Smith’s Account of Self-Deceit and Informal Institutions », May 2007. N° 07-09.RS Marie Pfiffelmann « Which optimal design for lottery linked deposit », May 2007. N° 07-08.RS Marc Lévy « Control in Pyramidal Structures », May 2007. N° 07-07.RS Olga Bourachnikova «Weighting Function in the Behavioral Portfolio Theory», May 2007.  N° 07-06.RS Régis Blazy and Laurent Weill « The Impact of Legal Sanctions on Moral Hazard when Debt Contracts are Renegotiable », May 2007. N° 07-05.RS Janine Leschke «Are unemployment insurance systems in Europe adapting to new risks arising from non-standard employment? », March 2007.

N° 07-04.RS Robert Plasman, Michael Rusinek, Ilan Tojerow « La régionalisation de la négociation salariale en Belgique : vraie nécessité ou faux débat ? », March 2007. N° 07-03.RS Oscar Bernal and Jean-Yves Gnabo « Talks, financial operations or both? Generalizing central banks’ FX reaction functions », February 2007. N° 07-02.RS Sîle O’Dorchai, Robert Plasman and François Rycx « The part-time wage penalty in European countries: How large is it for men? », January 2007. N° 07-01.RS Guido Citoni « Are Bruxellois and Walloons more optimistic about their health? », January 2007. 2006 N° 06-15.RS Michel

Beine,

Oscar

Bernal,

Jean-Yves

Gnabo,

Christelle

Lecourt

« Intervention policy of the BoJ: a unified approach » November 2006. N° 06-14.RS Robert Plasman, François Rycx, Ilan Tojerow « Industry wage differentials, unobserved ability, and rent-sharing: Evidence from matched worker-firm data, 1995-2002» N° 06-13.RS Laurent Weill, Pierre-Guillaume Méon « Does financial intermediation matter for macroeconomic efficiency? », October 2006. N° 06-12.RS Anne-France Delannay, Pierre-Guillaume Méon « The impact of European integration on the nineties’ wave of mergers and acquisitions », July 2006. N° 06-11.RS Michele Cincera, Lydia Greunz, Jean-Luc Guyot, Olivier Lohest « Capital humain et processus de création d’entreprise : le cas des primo-créateurs wallons », June 2006. N° 06-10.RS Luigi Aldieri and Michele Cincera « Geographic and technological R&D spillovers within the triad: micro evidence from us patents », May 2006. N° 06-09.RS Verena Bikar, Henri Capron, Michele Cincera « An integrated evaluation scheme of innovation systems from an institutional perspective », May 2006. N° 06-08.RR Didier Baudewyns, Benoît Bayenet, Robert Plasman, Catherine Van Den Steen, « Impact de la fiscalité et des dépenses communales sur la localisation intramétropolitaine des entreprises et des ménages: Bruxelles et sa périphérie», May 2006. N° 06-07.RS Michel Beine, Pierre-Yves Preumont, Ariane Szafarz « Sector diversification during crises: A European perspective », May 2006. N° 06-06.RS Pierre-Guillaume Méon, Khalid Sekkat « Institutional quality and trade: which institutions? which trade? », April 2006.

N° 06-05.RS Pierre-Guillaume Méon « Majority voting with stochastic preferences: The whims of a committee are smaller than the whims of its members », April 2006. N° 06-04.RR Didier Baudewyns, Amynah Gangji, Robert Plasman « Analyse exploratoire d’un programme d’allocations-loyers en Région de Bruxelles-Capitale: comparaison nternationale et évaluation budgétaire et économique selon trois scénarios », April 2006. N° 06-03.RS Oscar Bernal « Do interactions between political authorities and central banks influence FX interventions? Evidence from Japan », April 2006. N° 06-02.RS Jerôme De Henau, Danièle Meulders, and Sile O’Dorchai « The comparative effectiveness of public policies to fight motherhood-induced employment penalties and decreasing fertility in the former EU-15 », March 2006. N° 06-01.RS Robert Plasman, Michael Rusinek, and François Rycx « Wages and the Bargaining Regime under Multi-level Bargaining : Belgium, Denmark and Spain », January 2006. 2005 N° 05-20.RS Emanuele Ciriolo « Inequity aversion and trustees’ reciprocity in the trust game », May 2006. N° 05-19.RS Thierry Lallemand, Robert Plasman, and François Rycx « Women and Competition in Elimination Tournaments: Evidence from Professional Tennis Data », November 2005. N° 05-18.RS Thierry Lallemand and François Rycx « Establishment size and the dispersion of wages: evidence from European countries », September 2005. N° 05-17.RS Maria Jepsen, Sile O’Dorchai, Robert Plasman, and François Rycx « The wage penalty induced by part-time work: the case of Belgium », September 2005. N° 05-16.RS Giuseppe Diana and Pierre-Guillaume Méon « Monetary policy in the presence of asymmetric wage indexation », September 2005. N° 05-15.RS Didier Baudewyns « Structure économique et croissance locale : étude économétrique des arrondissements belges, 1991-1997 », July 2005. N° 05-14.RS Thierry Lallemand, Robert Plasman, and François Rycx « Wage structure and firm productivity in Belgium », May 2005. N° 05-12.RS Robert Plasman and Salimata Sissoko « Comparing apples with oranges: revisiting the gender wage gap in an international perspective », April 2005.

N° 05-11.RR Michele Cincera « L’importance et l’étendue des barrières légales et administratives dans le cadre de la directive ‘Bolkestein’ : Une étude comparative entre la Belgique et ses principaux partenaires commerciaux », April 2005. N° 05-10.RS Michele Cincera « The link between firms’ R&D by type of activity and source of funding and the decision to patent », April 2005. N° 05-09.RS Michel Beine and Oscar Bernal « Why do central banks intervene secretly? Preliminary evidence from the Bank of Japan », April 2005. N° 05-08.RS Pierre-Guillaume Méon and Laurent Weill « Can Mergers in Europe Help Banks Hedge Against Macroeconomic Risk ? », February 2005. N° 05-07.RS Thierry Lallemand, Robert Plasman, and François Rycx « The EstablishmentSize Wage Premium: Evidence from European Countries », February 2005. N° 05-06.RS Khalid Sekkat and Marie-Ange Veganzones-Varoudakis « Trade and Foreign Exchange Liberalization, Investment Climate and FDI in the MENA », February 2005. N° 05-05.RS Ariane Chapelle and Ariane Szafarz « Controlling Firms Through the Majority Voting Rule », February 2005. N° 05-04.RS Carlos Martinez-Mongay and Khalid Sekkat « The Tradeoff Between Efficiency and Macroeconomic Stabilization in Europe », February 2005. N° 05-03.RS Thibault Biebuyck, Ariane Chapelle et Ariane Szafarz « Les leviers de contrôle des actionnaires majoritaires», February 2005. N° 05-02.RS Pierre-Guillaume Méon « Voting and Turning Out for Monetary Integration: the Case of the French Referendum on the Maastricht Treaty », February 2005. N° 05-01.RS Brenda Gannon, Robert Plasman, Ilan Tojerow, and François Rycx « Interindustry Wage Differentials and the Gender Wage Gap : Evidence from European Countries », February 2005.

Brussels Economic Review University of Brussels DULBEA, CP140 Avenue F.D. Roosevelt, 50 B-1050 Brussels Belgium ISSN 0008-0195

Apart from its working papers series, DULBEA also publishes the Brussels Economic Review-Cahiers Economiques de Bruxelles.

Aims and scope First published in 1958, Brussels Economic Review-Cahiers Economiques de Bruxelles is one of the oldest economic reviews in Belgium. Since the beginning, it publishes quarterly the Brussels statistical series. The aim of the Brussels Economic Review is to publish unsolicited manuscripts in all areas of applied economics. Contributions that place emphasis on the policy relevance of their substantive results, propose new data sources and research methods, or evaluate existing economic theory are particularly encouraged. Theoretical contributions are also welcomed but attention should be drawn on their implications for policy recommendations and/or empirical investigation. Regularly the review publishes special issues edited by guest editors.

Authors wishing to submit a paper to be considered for publication in the Brussels Economic Review should send an e-mail to Michele Cincera: [email protected], with their manuscript as an attachment. An anonymous refereeing process is guaranteed. Additional instructions for authors and subscription information may be found on the Brussels Economic Review’s website at the following address: http://homepages.vub.ac.be/~mcincera/BER/BER.html