ESTUDIOS SOBRE LA ECONOMIA ESPAÑOLA

Leone Leonida Daniel Montolio

EEE 109

July 2001

http://www.fedea.es/hojas/publicado.html

Leone Leonida

Universitá Degli Studi Della Calabria. Dipartimento di Economia Politica. University of York. Department of Economics and Related Studies.

Daniel Montolio

Universitat de Barcelona. Dept. Hacienda Pública and Institut d’Economia de Barcelona. University of York. Department of Economics and Related Studies.

Abstract. The purpose of this paper is to investigate the dynamics of the per capita GDP for the Spanish provinces over the period 1961-1997. A nonparametric density estimation approach is used to show that Spanish provinces had convergent dynamics during years (1961-1981), but divergent dynamics in the last decades (1981-1997). A period of convergence among clusters during the sixties and seventies is found, followed by evidence of intra-clusters convergence (polarization of income) during the eighties. A starting process of divergence among clusters and increasing polarization is found for the last period analysed (1991-1997). Moreover, the analysis allows us to determine the position in the per capita income distribution of each province over time, such that it can be discovered “who is moving where” in the income distribution itself, highlighting the territorial distribution of growth among the Spanish provinces (and regions) over the whole period.

JEL: C14, O18, O40, R11

Key words: Convergence, Polarization, Clusters Dynamics, Non-parametric Density Estimation.

1. Introduction

The convergence issue focuses on whether a set of economies (countries, regions, provinces), starting from different income (or product) levels, will tend to converge to the same per capita income level.

The debate on convergence among macroeconomists and econometricians is still open. From an econometric point of view, if the classical approach to convergence (consisting in regressing the measured growth rate against the initial per capita income level) is used, many questions arise. One of the main drawbacks can be summarised as follows: since only the first two moments of the distribution are involved in the classical analysis, the approach itself is uninformative about the dynamics of the entire income distribution (Quah (1997)).

Various studies assessed, most of them using the “convergence equation”, whether there has been convergence for Spain. However, the analysis did not take complete advantage of the tools recently developed in the convergence literature¹.

This paper purports to assess the convergence issue for the Spanish case by studying the evolution of the per capita income distribution across time. The methodology presented utilises a number of tools recently developed in the convergence debate in order to investigate the dynamics of the per capita Gross Domestic Product (GDP) among the Spanish provinces over the period 1961-1997. To avoid issues linked to the cross-sectional regression and the time series approaches, a non-parametric density estimation approach is used. This overcomes important shortcomings of standard tests commonly used in the convergence debate by making use of the whole distribution and, consequently, results in more information. Furthermore, much of the work dealing with convergence in Spain has been focused at the regional level; by applying this methodology, we can focus on provincial dynamics.

¹ Pioneering works were Dolado et. al. (1994) where the convergence issue was mainly analysed for the Spanish provinces, and García-Greciano and Raymond (1994) who studied regional convergence in Spain. These works were followed by other regional studies such as De la Fuente (1994, 1996), García-Greciano et. al. (1995), Mas et. al. (1994, 1995, 1998), Cuadrado et. al. (1999), Gorostiaga (1999), Salas (1999), García-Greciano and Raymond (1999), among others. Using different specifications and econometric tools, convergence among Spanish regions has been a common result in these works. Recently, Lamo (2000) assessed the issue by studying the conditional distribution, finding no empirical evidence in favour of convergence. Nevertheless, by studying the distribution year by year, as in this work, much more can be said.

We raise, and answer, the following questions: are the Spanish regions and provinces converging? That is, is the poorer part of Spain catching up the richest one? We will show that Spain had convergent dynamics during the period 1961-1981, but divergent dynamics in recent years (1981-1997). Moreover, evidence in favour of inter-group convergence (or polarization of income) is found during the eighties, and a starting process of divergence and increasing polarization among groups of provinces is found for the last period analysed (1991-1997).

Secondly, are there any inter-distributional movements? In other words, is there evolution inside the distribution itself, and what form does this evolution take? We will show that over the first period analysed, beginning in 1961 and finishing in 1981, there was overall convergence dynamics, while in the period 1981-1997 we find overall divergence. These dynamics were essentially due to an incoming and then vanishing “middle class” cluster of provinces.

Third, which part of Spain, from a geographical point of view, is growing faster, and what is the territorial distribution of growth in Spain? We show see that in the South, Centre and the North West of Spain there were provinces with per capita income below the average that gained positions in the income distribution but not enough to join the more developed provinces. Moreover, there was a process of convergence inside the North East part of Spain and Madrid, creating, in 1997, a separated and diverging cluster.

The rest of the paper is organised as follows. Section 2 revises the theoretical and empirical debate on convergence. Section 3 describes the data set used for the empirical analysis. Section 4 deals with the empirical results on polarisation and convergence (4.1), inter-distributional dynamics (4.2), and intra-clusters dynamics and geographical distribution of growth (4.3). Finally, section 5 outlines some conclusions.

2. Convergence: Theoretical and Empirical Debate

The classical approach to the convergence issue, in large part due to Baumol (1986) and extensively used by Barro and Sala-i-Martín (1991, 1992), consists of estimating crosssectionally the measured growth rates against the initial level of per capita income. A negative sign for such an estimate would imply that the greater the initial level of income, the lower the growth rate. From such evidence, it follows that richer countries grow less than the poorer countries, and this has been used as proof of the existence of convergence dynamics among different economies.

Conversely, time series approaches reach radically different conclusions, by testing whether income inequalities are persistent over time. Both cointegration and Kalman’s filter tests tend to reject the convergence hypothesis.

Some researchers have underlined that these approaches are misleading in some ways. For instance Quah (1993b, 1996c) revises the β- and σ-convergence approaches. Using the concept of Galton’s Fallacy, Quah shows that calculating a cross-sectional regression to explain time-averaged growth rates is inadequate in determining how the distribution of income per capita evolves through time, and across economies . Quah also highlights some drawbacks of the σ-convergence approach: the variance of the per capita income across economies can be unchanged through time, showing no converge or no divergence in this sense, but “the economies underlying the cross-section could still be moving about within the invariant distribution” (Quah (1996a)). Therefore, the σ- convergence approach fails to explain inter and intra-distributional dynamics, which is crucial when studying the persistence of income disparities over time (see Quah (1996a,b)). In support of this, Bianchi (1997) stated: “the σ-convergence analysis alone is not sufficient to study convergence unless more information is gained on how units move within the distribution”.

Quah also criticizes one of the main results of the classical convergence theory: poor and rich economies all appear to be converging toward each other at a stable, uniform rate of 2% per year. “The idea here is that such consistency might only reflect something mechanical and independent of the economic structure of growth” Quah (1996b).

Summarising, “cross-section regressions can represent only average behaviour, not the behaviour of an entire distribution” Quah (1996a). Time series tests themselves suffer because of the same issue. They are useful in comparing a pair of economies, but not a large sample of them. In addition, they fail to describe transitional dynamics.

Therefore, it seems more adequate and interesting to estimate the distribution of the per capita income for different years, to analyse the shape, and to see if there is any tendency to collapse toward a unimodal distribution, provided that the initial one is bimodal, or vice versa.

Figure 1 represents an income distribution at time t and another (possible) distribution at time t+s. If the distribution collapses from a unimodal to a bimodal distribution (emerging twin peaks), intra-convergence in groups of per capita GDP but divergence with other group(s) can be found. Alternatively, if the distribution collapses from a bimodal to a unimodal distribution, economies are said to converge over time.

This approach allows us to analyse inter-distributional (inside the per capita income distribution) and intra-cluster (within the different groups) dynamics in which the macroeconomic theories of growth and microeconomic models of cross-sectional interaction can be combined. Therefore, analysing the distribution of the income per capita across countries, regions or provinces, both shape and mobility dynamics can be studied at the same time³.

Figure 1. Emerging Twin-Peaks.

Using a non-parametric approach, and the same data set investigated by other researchers (Summer and Heston (1991)), Quah (1996a, 1997) showed that there is no evidence in favour of convergence among countries in the world. Quah estimated the densities for a number of years and he found a fluctuating bimodal distribution. This evidence has been taken as empirical proof against the convergence hypothesis. The nonparametric density estimation approach is the approach we will use to analyse the convergence issue among Spanish provinces.

3. Data Description: Variables and Data Sources

The variable used to investigate the existence of convergence among the Spanish provinces, and its “dynamics” properties, is per capita Gross Domestic Product (GDP) for the period 1961-1997 for the 52 provinces (including the Spanish cities in the north of Africa: Ceuta and Melilla). This variable is one of the more comparable indexes across different economies (regional or provincial) and time. Moreover, per capita GDP is one of the most common measures of wealth of an area, and hence a good instrument to investigate the convergence process across different economies.

³ On one hand, studying the shape of the distribution we analyse if there is convergence or divergence; if more than two peaks emerge in the distribution (stratification) or if countries are catching up with one another bu only within particular subgroups (convergence clubs). On the other hand, studying the mobility dynamics of a distribution allow us to examine if rich countries at time t are still rich at time t+s (persistence); if some poor countries at time t+s had begun rich (churning or mobility) or/and if some groups of these economies, who were originally close together in the middle class, have separated because of a process of divergence (separa-

The main data source is the recently published database by the Fundación Banco Bilbao Vizcaya (Fundación BBV (1999)), which allows use of a homogeneous series from 1955 to 1997⁴.

In order to avoid the bias that inflation could cause in our analysis, we use the series at 1990 constant prices⁵. Following Quah (1997), we have calculated the ratio between per capita GDP in each year and for each province and the average per capita GDP for Spain. This normalization is very useful in two aspects. Firstly, it is a simple way to abstract from Spain’s total growth and fluctuations. Secondly, since the average for Spain is equal to one, the data assumes the form of dispersion around the mean, and as a result, it is possible to make direct comparisons across years.

4. Empirical Results

4.1. Assessing Convergence across Spanish Provinces

The underlying idea of the non-parametric approach is to “rid oneself of the need to specify in advance a particular functional form (of the underling density)” Johnston and Di

bility).

⁴ We have elected to use the first year of each decade and the last available year in our database (1961, 1971, 1981, 1991 and 1997). In this way we can analyse how the overall distribution evolves during each decade, and more important, in recent years.

⁵ The possibility of using this new database allows us to extend the period of investigation to recent years (up to 1997), while other works on convergence for the Spanish case (mainly using a parametric approach) used data up until the beginning of the 1990’s. This point is crucial when interpreting the results of this study and comparing them with the results of previous studies.

Nardo (1997). Many methods can be used to estimate the density of a given data this study makes use of the Kernel estimator. A Kernel function is defined as:

\[\int_ {x = - \infty} ^ {x = \infty} K (u) d u = 1\tag{[1]}\]

This function is smooth, has derivatives and is used as a weighting function. A broad class of density estimators (the Ronsenblatt-Parzen Kernel density estimators) can be defined as:

\[\hat {f} _ {N, h} (x) = \frac {1}{N} \sum_ {i = 1} ^ {N} \frac {1}{h} K \left(\frac {x - X _ {i}}{h}\right)\tag{[2]}\]

in which K refers to the Kernel function [1], N is the number of observations in the sample, and h is the bandwidth (or window width). In our estimates, we will always use a Gaussian Kernel and the Jones’ rule for the (averaged) optimal bandwidth⁷.

The density estimates for the five years chosen are showed in Figure 2, and the results can be summarized as follows. In 1961, the distribution clearly shows three modes. Provinces were grouped around 0.7, 1 and 1.5 of the income distribution (Spanish average=1). Therefore, in the initial period of our analysis there were three groups (clusters) of provinces: one group that can be considered as “poor provinces” (below the average), another grouped around the Spanish average, that we can call “middle class provinces”, and finally a third group of “rich provinces” (above the average).

From 1961 to 1971, there was a period of convergence across provinces. The poor cluster “moved to the right”: the average value of the per capita GDP of this group of provinces increased (from around 0.7 to 0.8) and approached the middle class provinces (set at about 1.0). Secondly, rich provinces lost positions within the distribution, and were grouped around 1.4 of the income distribution.

⁶ For a good survey of different estimation methods, see Silverman (1986).

⁷ We use the Gaussian Kernel and the Jones’ bandwidth because they are among the most common and less controversial. The bandwidth has to be averaged in order to make comparisons across years by applying the same amount of smoothing to the estimates. A more complete version of this paper, with different Kernels and bandwidths is available from the authors on request.

Figure 2. Kernel Density Estimates (Gaussian) of the Relative per capita GDP with the average of the Jones Bandwidth for the Spanish Provinces (h=0.07920228). 1961–1997.

Estimates Information i. Kernel: Gaussian i Computation Method: FFT i. Automatic Bandwidth Selector: Average of Jones' Bandwidth iv. Bandwidth = 0.07920228

Therefore, during this initial period there was a path of overall convergence⁸: a process of clustering of poor and middle class provinces, and the richest part of Spain losing positions characterizes this convergence pattern.

⁸This result was also found, among others, by Dolado et al. (1994) in their study of the convergence issue

Between 1971 and 1981, there was a process of polarization of income among Spanish provinces. During this period, the convergence process, observed during 1961- 1971, slowed down: in 1981, there were two clear and distant modes, one below the average (0.8) and the other slightly above the average (1.09). Furthermore, rich provinces were still losing positions in the income distribution and approaching the middle class provinces (stratification of income).

For the 1961-1981 period, the rich provinces lost positions in the income distribution, but still created a separate mode (showing persistence of income disparities). Poor provinces increased from 0.7 to 0.8 in the income distribution, but not all of them caught up the middle class provinces. The new and important cluster in 1981 (around 1) was created by rich provinces losing positions in the income distribution, and some of the poor provinces catching up, forming all together a prominent middle class mode (mobility of some provinces). We will analyse in more detail these inter-distributional and intracluster dynamics in the next sub-sections.

The estimated distribution in 1991 is unimodal, and it is mostly concentrated around 0.8. However, there is a little ridge around 1.1 (although this is not important enough to be considered a significant mode). Therefore, the 1981-1991 evolution of the income distribution can be understood as a process of some middle class provinces losing and being made part of the poor mode, and some growing and approaching the rich mode, and thus resulting in the middle class mode vanishing (see figure 3a).

The provinces that remained at 1.1, and started to catch the rich provinces (that were losing at the same time), did not do so enough to create a mode in the estimated distribution.

for the Spanish provinces.

Finally, the 1991-1997 period showed a starting process of divergence. The 1997 distribution was not unimodal anymore; the second mode was clearly more prominent and it increased from 1.1 to about 1.2 of the income distribution (see figure 3b). Provinces were clustered in two levels of income: 0.8 (below average) and 1.2 (above average) implying polarization of income.

Figure 3. a) Comparison between 1981 and 1997 Density Estimates for the Spanish Provinces. Polarization of Income. b) Comparison between 1991 and 1997. Divergence.

We can summarize the evolution of the per capita income distribution across Spanish regions in two main periods. The first period (1961-1981) was characterized by an initial period of convergence and the creation of what we have termed the “middle class” mode. During the second period (1981-1997), there was a process of polarization (vanishing middle class) and an incipient process of divergence (separation of modes).

4.2. Inter-Distributional Dynamics

The previous analysis was concentrated on the evolution of the shape of the entire distribution over time. With this analysis, it is not possible to study the dynamics of a single province. In other words, it is not possible to see “who is moving where” (Jones 1997b) in the distribution over time. To be able to individuate which provinces (and also regions) are gaining and which are losing in the distribution, a bivariate analysis is required.

The definition of the Kernel estimator can easily be generalised for the multivariate case. The multivariate Kernel density estimator with kernel K and bandwidth h is defined as:

\[\hat {f} (\mathbf {x}) = \frac {1}{n h ^ {d}} \sum_ {i = 1} ^ {n} K \left\{\frac {1}{h} \left(\mathbf {x} - \mathbf {X} _ {i}\right) \right\}\tag{[3]}\]

where the Kernel function is now a function defined for d-dimensional x, satisfying

\[\int_ {R ^ {d}} K (\mathbf {x}) d \mathbf {x} = 1\tag{[4]}\]

The main point is that, applying this methodology to two years (d = 2) we have an estimate of the bivariate Kernel (the joint distribution in this case can represent the continuous transition matrix (Quah (1997)), which provides us with information on the dynamics of the clusters over the two years⁹.

Plotting in the bivariate Kernel density estimate the valley¹⁰ positions of the distributions, the 45-degree line and superimposing observations will inform us about the dynamics inside the per capita income distribution which occurred during the two chosen years: if a province crossed the valley over the period analysed it means that it changed from one cluster to another (inter-distributional dynamics). Furthermore, other information can be extracted from the way in which the mass of the estimate is distributed on the space: if the two clusters communicated at some fixed probability level then overtaking occurred. Vice versa, if the two clusters were very distant from each other and there was no communication between them at any probability level then no overtaking occurred.

⁹ Depending on the position in the space of the estimate, we will have information on the divergence (or convergence) process: “if most of the graph were concentrated along the 45-degree diagonal, then elements in the distribution remain where they began. If, by contrast, most of the mass in the graph were rotated 90 degrees counter-clockwise from that 45-degree diagonal, then substantial overtaking occurs” (Quah 1997).

¹⁰ The valleys of the per capita income distribution are the separating points between two modes, or two clusters of provinces found in the univariate case (sub-section 4.1).

The bivariate Kernel estimation is carried out for the overall period and for the two most interesting periods (1961-81 and 1981-97) found in the previous sub-section. Figure 4 presents four graphs corresponding to the bivariate density estimation: a 3D graph showing the bivariate distribution between 1961 and 1997 and its contour plot, and the two contour plots corresponding to the bivariate estimation for the two sub-periods analysed.

For the period 1961-1997, there was a convergent pattern because poor provinces gained positions in the income distribution while rich provinces lost positions in the distribution (the two extreme clusters in 1997 are closer than in 1961). Nevertheless, clearly more can be said analysing the dynamics of two sub-periods defined in the previous subsection.

During the first sub-period (1961-1981), the main pattern observed among provinces below the average per capita income was growth (since this cluster is above the 45-degree line). However, some of the poor provinces changed cluster¹¹ (crossing the valley) and joined what we have termed the middle class cluster observed in 1981. Furthermore, the position of rich provinces (mainly below the 45-degree line) confirms that they lost positions in the income distribution during this period. Two of them joined the middle class cluster, and only one province from the middle class in 1961 caught up rich rich cluster, while others lost positions and joined the poor cluster in 1997. Therefore, the polarization pattern found before is re-confirmed. The rich cluster in 1997 not only received growing provinces from the vanished middle class, but also rich provinces that lost positions in the income distribution during the eighties.

¹¹ In sub-section 4.3, we well individuate which are these provinces.

The analysis of the bivariate density estimation for the period 1981-1997 confirms the results found in the previous section. The poor cluster is nearly on the 45-degree line (i.e. the cluster remained in the same position in 1997 than in 1981) but the middle class cluster found in 1981 clearly vanished. Provinces, which in 1981 were around the Spanish average, followed two different patterns of growth. Some of them grew and moved to the

Summarising, the dynamics across Spanish provinces for the overall period were convergent. However, the sub-periods analysed provide much more information: the period 1961-1981 showed convergence and polarization of income. Instead, the period 1981-1997 was a period of polarization of income and an emerging divergence pattern across Spanish provinces.

4.3. Intra-Cluster Dynamics and Territorial Distribution of Growth in Spain

It is interesting to analyse which are the provinces that have characterised the patterns observed. Do these provinces have something in common? For instance, do they come from the same geographical area? What relation is there among provinces coming from the same region? In studying this issue, we will clarify the territorial distribution of growth for Spanish provinces and regions.

Figure 5 (in the Appendix) shows the bivariate plot of the distributions of the relative per capita income across provinces between two years and for three periods: the overall period (1961-1997), and the two sub-periods previously analysed (1961-1981 and 1981-1997). In Figure 5, provinces have been numbered from 1 to 52 (see Table 1 in the appendix for details), and the modes (full line), the valleys positions (dotted line) of the estimated distributions and the 45-degree line have been superimposed on the graph. Adding the mode position of the income distribution will inform us about the dynamics inside the clusters themselves (intra-cluster dynamics). Therefore, we will be able to analyse how the provinces belonging to each cluster have evolved with relation to the provinces with a similar level of relative per capita income.

During the period 1961-1981, as analysed previously, poor provinces grew, with several catching up provinces with a relative per capita income around the Spanish average. These provinces were Teruel, Sta. Cruz de Tenerife, Las Palmas de Gran Canaria, Burgos, Palencia, Valladolid and Guadalajara. These provinces, belonging to four different regions (Aragón, Canarias, Castilla-León and Castilla la Mancha), grew enough to change from the poor cluster to the middle class cluster. The rest of poor provinces remained in the poor cluster; however, the mode position indicates which regions grew and which lost positions inside that cluster. Provinces mainly stayed in the same relative positions except for Huelva, Soria and Toledo that moved from the lower level of the poor cluster in 1961 to the higher level in 1981.

Middle class regions in 1961 remained mainly in the same position in 1981, growing all with the exception of Zaragoza, Cantabria, Valencia and Navarra, which lost positions during that period. However, Tarragona grew enough to change cluster (it crossed the valley that separated middle and rich clusters in 1981). Among rich provinces, only Baleares and Álava gained positions in the income distribution, while Barcelona, Girona, Madrid, Guipúzcoa, and Vizcaya lost positions. It should be noted that the decrease of positions that the two provinces from País Vasco experienced during the sixties and seventies is quite pronounced.

During the period 1981-1997, poor provinces grew but they remained in the same cluster, but none of them caught up middle class provinces. Only Cádiz, Málaga, Sevilla, León, A Coruña and Ceuta lost positions inside the cluster below the Spanish average in 1997 from 1981.

Now, we can analyse which middle class provinces in 1981 created the polarization pattern explained before (i.e. the vanishing middle class). Asturias, Las Palmas de Gran Canaria, Santa Cruz de Tenerife, Cantabria and Alicante lost positions during the eighties and nineties and at the end of the period formed part of the poor cluster. Moreover, Zaragoza, Burgos, Guadalajara, Lleida, Castellón, Navarra, Vizcaya and La Rioja grew enough to catch the rich provinces and in 1997 created the second and distant mode around 1.2 of the relative per capita income distribution. Furthermore, rich regions such as Baleares, Girona, Madrid and specially Tarragona kept loosing positions. Only Álava and Barcelona gained positions during this period.

Finally, we have divided Spain in three main geographical areas (see Figure 6). South provinces are coloured in black and include provinces belonging to Andalucía, Canarias, Castilla La Mancha, Extremadura, Murcia, Ceuta and Melilla. North West provinces (coloured in dark grey) include provinces of Galicia, Castilla-León, Asturias and Cantabria. North East provinces from Aragón, Balears, Catalunya, Comunitat Valenciana, Navarra, La Rioja and País Vasco are coloured in white. Finally Madrid, due to its special features, has been coloured in light grey.

The first conclusion that we can draw from Figure 6 is that the cluster below the Spanish average was mainly formed by provinces from the South, Center (both Castillas) and North West of Spain. They grew during the overall period (1961-1997) but they did not catch the rich provinces. In the North East of Spain, there were middle class and rich provinces. The main pattern observed was middle class provinces catching rich provinces, while the rich provinces were losing positions.

Thus, on the one hand, the South, Centre and North West of Spain increased their positions in the income distribution, but not enough to join the more developed provinces. On the other hand, there has been a process of convergence inside the North East part of

Figure 6. Regional Distribution of Growth

5. Conclusions

Some conclusions can be drawn from the above analysis despite the complexity of the convergence issue among the Spanish provinces. However, the non-parametric approach used in this paper has shed some light on this issue.

First, during the sixties there was a period of convergence between provinces below the Spanish average and the middle class provinces (around the average). At the same time, rich provinces lost positions but created a distant and very significant mode. This result completes previous studies that found convergence during this period; we can conclude that the convergence process observed during the sixties was mainly caused by middle class and poor provinces converging. In others words, we can speak about both convergence and clustering dynamics during this period.

The seventies was a period where the convergence slowed down. First, some middle class provinces grew and separated from poor provinces that did not grow enough to join them. Second, rich regions kept losing positions in the income distribution, and some of them approached the middle class mode.

During the eighties, the middle class mode estimated at the beginning of the period (1981) vanished. On the one hand, some middle class provinces lost positions and joined poor provinces that grew and grouped in a higher mode. On the other hand, some middle class provinces grew enough to approach rich provinces (grouped in a much lower mode than in 1961).

The final and most interesting period, the 1990’s, shows how the main mode, below the average, remained unchanged, and how some provinces from the vanished middle class caught up rich regions, creating a new mode at around 1.2 of the income distribution. This implied not only a process of polarization of income: not only could we find two separated groups of provinces, one below the average (around 0.8) and another above the Spanish average (around 1.2), but also a process of starting divergence: the two modes started to separate. In the new mode, we could find provinces located mostly in the North East of Spain (and Madrid), implying that the pattern of polarization of income observed during the nineties separated Spain into two diverging entities.

The specific characteristics of the evolution through time of the Spanish provinces makes the topic very interesting for further research to determine the main causes of the evolution pattern observed among Spanish provinces.

6. Appendix

Figure 5. Bivariate plots of the distributions of the relative per capita income.

Figure 5 (cont). Bivariate plots of the distributions of the relative per capita income

7. Bibliography

References

1. Barro, R. (1991), “Economic Growth in a Cross Section of Countries”, The Quarterly Journal of Economics, May.

References

1. Barro, R. and Sala-i-Martín, X. (1991), “Convergence Across States And Regions”, Brookings Paper on Economics Activity, 1.

References

1. Barro, R. and Sala-i-Martín, X. (1992), “Convergence”, Journal of Political Economy, 100,2.

References

1. Barro, R. and Sala-i-Martín, X. (1995), “Technological Diffusion, Convergence, and Growth”, National Bureau of Economic Research, Working Paper: 5151, June 1995.

References

1. Barro, R. and Sala-i-Martín, X. (1996), Economic Growth, New York, McGraw Hill

References

1. Baumol, W. (1986), “Productivity Growth, Convergence and Welfare. What Long-Run Data Show”, American Economic Review, 76, 5, December.

References

1. Cuadrado, J.R., García-Greciano, B. and Raymond, J.L. (1999), “Regional Convergence in productivity and productive structure: the Spanish case”, International Regional Science Review, vol.22.

References

1. De la Fuente, A. (1994), “Inversión pública y desigualdad regional en España 1891-1990”, in J.Mª Esteban and J.Vives (dirs.), Crecimiento y convergencia en España y en Europa. Volumen II, Instituto de Análisis Económica, Barcelona.

References

1. De la Fuente, A. (1996), “On the Sources of Convergence: A Close Look at the Spanish Regions”. CEPR Discussion Paper:1543. December 1996.

References

1. De la Fuente, A.(1998), “Convergencia real? España en la OCDE”, Moneda y Crédito, vol 207, 1998.

References

1. Dolado, J. J.; González-Páramo, J. M. and Roldan, J. M. (1994), “Convergencia económica entre las provincias españolas: Evidencia empírica (1955-1989)”. Moneda y Crédito; 0(198), 1994.

References

1. Durlauf, S. N. (1996), “A Theory of Persistent Income Inequality”, Journal of Economic Growth, 1.

References

1. Fundación BBV (1999), Renta nacional de España y su distribución provincial. Serie homogénea 1955-1993. Avance 1994-1997. Bilbao.

References

1. García-Greciano, B. and Raymond, J.L. (1994), “Las disparidades en el PIB per cápita entre Comunidades Autónomas y la hipótesis de convergencia”, Papeles de Economía Española, nº 59.

References

1. García-Greciano, B. and Raymond, J.L. (1999), “Las disparidades regionales y la hipótesis de convergencia: Una revisión”, Papeles de Economía Española, vol 80, 1999.

References

1. García-Greciano, B., Raymond, J.L. and Villaverde, J. (1995), “Convergencia de las provincias españolas”, Papeles de Economía Española, nº 64.

References

1. Gorostiaga, A. (1999), “Cómo afectan el capital público y el capital humano al crecimiento?: Un análisis para las regiones españolas en el marco neoclásico”, Investigaciones Económicas; 23(1), January 1999.

References

1. Jones, C. (1997a), “On the Evolution of the World Income Distribution”, Journal of Economic Perspective, v. 11, n. 3.

References

1. Jones, C. (1997b), “Convergence Revisited”, Journal of economic Growth, 2, June.

References

1. Jonhston, R. and Di Nardo, R. (1997), Econometric Methods, McGraw Hill, IV Edition.

References

1. Lamo, A (2000), “On Convergence Empirics: Some Evidence for Spanish Regions”, Investigaciones Económicas, Vol. XXIV (3).

References

1. Mas, M., Maudos, J., Perez, F. and Uriel, E. (1994), “Disparidades regionales y convergencia en las CC.AA.”, Revista de Economía Aplicada, Vol. (4).

References

1. Mas, M., Maudos, J., Perez, F. and Uriel, E. (1995), “Public capital and Convergence in the Spanish Regions”, Entrepreneurship and regional Development, (7) 4.

References

1. Mas, M., Maudos, J., Perez, F. and Uriel, E. (1998), “Public Capital, Productive Efficiency and Convergence in the Spanish Regions (1964-93)”. Review of Income and Wealth; 44(3),September 1998.

References

1. Maudos, J., Pastor, J. and Serrano, L. (1998), “Convergencia en las regiones españolas: Cambio técnico, eficiencia y productividad”, Revista Española de Economía, 15(2), 1998.

References

1. Marron, J. S. and Schmitz, H. P. (1992), “Simultaneous Density Estimation of Several Income Distributions”. Econometric Theory, 8.

References

1. Pérez, P. (2000), “Dinámica de las regiones en España (1955-1995)”. Revista de Economía Aplicada. Número 22 (Vol. VIII).

References

1. Quah, D. T. (1993a), “Empirical Cross-section Dynamics in economic Growth”, European Economic Review 40.

References

1. Quah, D. T. (1993b), “Galton’s Fallacy and Test of the Convergence Hypothesis^#, The Scandinavian Journal of Economics, 95

References

1. Quah, D. T. (1996a), “Empirics for Growth and Convergence”, European Economic Review, 40.

References

1. Quah, D. T. (1996b), “Regional Convergence Clusters across Europe”, European Economic Review, 40.

References

1. Quah, D. T. (1996c), “Twin Peaks: Growth and Convergence in Models of Distribution Dynamics”, Economic Journal, 106.

References

1. Quah, D. T. (1997), “Empirics for Growth and Distribution: Stratification, Polarization, and Convergence Clubs”, Journal of Economic Growth, 2.

References

1. Sala-i-Martín, X. (1996), “Regional Cohesion: Evidence and Theories of Regional Growth and Convergence”, European Economic Journal, 40

References

1. Salas, R. (1999), “Convergencia, movilidad y redistribución interregional en España: 1981-1996”, Papeles de Economía Española, Vol. 80, 1999.

References

1. Silverman, B. W. (1981), “Using Kernel Density to Investigate Multimodality”, Journal of the Royal Statistical Society, Series B 43.

References

1. Silverman, B. W. (1983), “Some Properties of a Test for Multimodality Based on Kernel Density Estimates”, in J. F. C. Kingman and G. E. H. Reuter, Probability, Statistic, and Analysis, vol. 79, London Mathematical Society Lecture Note Series, Cambridge University Press.

References

1. Silverman, B. W. (1986), Density Estimation for Statistic and Data Analysis, New York, Chapman and Hall.