Page 207 - ePaper
P. 207
Chapter 4: Restoring Convergence between Member States in the EU and EMU
Box 1: Economic convergence, growth models and measures of convergence
Economic convergence and growth models
Economic growth is conventionally attributed to the accumulation of human and physical capital and increased productivity following
technological innovation. The most basic growth model, the Solow model (also called the neoclassical growth model) considers that
technological innovations are exogenous and assumes that capital and labour have diminishing returns. Notably it implies that, in
general, poor countries with less capital per person grow faster (because of diminishing returns to capital), leading to convergence in
GDP per head over time.
In the Solow model, GDP depends on production factors (capital and labour) augmented by technology. Total factor productivity (TFP) is,
by definition, that part of the increase in output that cannot be explained by changes in the other input factors. This residual is seen as a
(proxy) measure of skills, knowledge and technical progress. In empirical analysis, capital and TFP are not easy to separate. This is due
to the fact that technical progress is often embodied in new capital goods. One would underestimate the effect of TFP by assuming that
growth is the result of capital accumulation. Differences in TFP are seen to be important in explaining differences in income and growth
between countries, particularly in the long run when countries can overcome the steady state and grow by inventing new technology.
Decomposition of growth
Trends in GDPpc and GHDIpc are measured in constant prices, since the focus is on real economic and living conditions convergence ( ).
1
Furthermore, the use of GDP in real euros (deflated by the GDP deflator) is preferred to the PPS which are available in nominal values
and are thus more appropriate for cross-section comparisons (since No specific price deflator of PPS values is available).
GDP and growth can be decomposed into several contributions. This section uses a standard simple decomposition of GDPpc trends in
productivity (apparent employment productivity GDP/L), employment rate of the 15–64 population (share of employment in the active
age population) and active age population rate (share of active age population in total population), as reflected below.
GDPpc = GDP /Population = (GDP / L) * (L / POP active age) * (POP active age / Population)
GDPpc = (Apparent productivity) * (Employment rate) * (Share of active age population)
Measures of convergence
Sigma-convergence refers to a reduction of disparities over time between countries, for instance, measured in terms of the standard
deviation or coefficient of variation (the ratio of the standard deviation to the average). Beta-convergence refers to a situation where
incomes in poorer countries grow faster than those in richer ones, usually measured in terms of change over time. The two concepts
of convergence are closely related with Beta-convergence being necessary but not sufficient to achieve Sigma-convergence (see, for
instance, Monfort, 2008).
Other indices exist (for instance, the Gini coefficient, the Atkinson index, the Theil index and the Mean Logarithmic Deviation). It is recom -
mended that we ‘consider a variety of measures to draw firm conclusions about changes in the extent of disparities’ (see, for instance,
Montfort, 2008), and the analysis in this chapter focuses on the coefficient of variation as a main measure of sigma-convergence,
complemented as regards within zones and between zones dispersion by a standard between-within decomposition of total variance
and a decomposition of the Theil index (see Annex 3). An emphasis in the main text is put on the decomposition of total variance
which is closer to the measure of the coefficient of variation and, more specifically, on the share of total variance corresponding to the
between zones component (as the level of variance per se can be misleading, since it is affected by homothetic changes which do not
affect dispersion, the Annex provides additional elements on the level of the between zones contribution to total variance expressed
as an index, based on the first year when data are available).
( ) Furthermore, while entry into the euro is conditional on fulfilling the Maastricht criteria, the euro is intended to support real convergence, defined in
1
terms of per capita GDP, by fostering economic integration (see European Commission, 2008).
2.1.2. Convergence in Europe, • EU-15 Centre (Belgium, Luxembourg, represented 26 % of EU-28 population
trends between and within zones the Netherlands, Germany, Finland, in 2013.
France, Austria) ( ), which represented
8
In order to provide an overview of 36 % of EU-28 population in 2013. • EU-13 Centre and North (Czech Republic,
employment and social convergence Hungary, Poland, Slovenia and Slovakia),
trends in Europe (EU-28) overall, it is use - • EU-15 North (Denmark, Sweden, which represented 13 % of EU-28 popu -
9
ful to reflect not only on overall develop - United Kingdom) ( ), which represented lation in 2013.
ments, but also on changes in dispersion 17 % of EU-28 population in 2013.
both within and between zones. For this • EU-13 South and periphery ( Bulgaria,
purpose, five groups of countries are • EU-15 South and periphery (Greece, Cyprus, Estonia, Latvia, Lithuania, Malta,
considered, reflecting socioeconomic and Ireland, Portugal, Spain, Italy) ( ) which Croatia, Romania) which represented
10
geographical proximity criteria: 8 % of EU-28 population in 2013.
( ) Or in other terms EA-12 Northern countries,
8
see European Commission (2014a).
( ) Which are actually EU non-EA countries.
9
( ) Which are actually EA-12 South and periphery
10
countries, see European Commission (2014a).
205
Box 1: Economic convergence, growth models and measures of convergence
Economic convergence and growth models
Economic growth is conventionally attributed to the accumulation of human and physical capital and increased productivity following
technological innovation. The most basic growth model, the Solow model (also called the neoclassical growth model) considers that
technological innovations are exogenous and assumes that capital and labour have diminishing returns. Notably it implies that, in
general, poor countries with less capital per person grow faster (because of diminishing returns to capital), leading to convergence in
GDP per head over time.
In the Solow model, GDP depends on production factors (capital and labour) augmented by technology. Total factor productivity (TFP) is,
by definition, that part of the increase in output that cannot be explained by changes in the other input factors. This residual is seen as a
(proxy) measure of skills, knowledge and technical progress. In empirical analysis, capital and TFP are not easy to separate. This is due
to the fact that technical progress is often embodied in new capital goods. One would underestimate the effect of TFP by assuming that
growth is the result of capital accumulation. Differences in TFP are seen to be important in explaining differences in income and growth
between countries, particularly in the long run when countries can overcome the steady state and grow by inventing new technology.
Decomposition of growth
Trends in GDPpc and GHDIpc are measured in constant prices, since the focus is on real economic and living conditions convergence ( ).
1
Furthermore, the use of GDP in real euros (deflated by the GDP deflator) is preferred to the PPS which are available in nominal values
and are thus more appropriate for cross-section comparisons (since No specific price deflator of PPS values is available).
GDP and growth can be decomposed into several contributions. This section uses a standard simple decomposition of GDPpc trends in
productivity (apparent employment productivity GDP/L), employment rate of the 15–64 population (share of employment in the active
age population) and active age population rate (share of active age population in total population), as reflected below.
GDPpc = GDP /Population = (GDP / L) * (L / POP active age) * (POP active age / Population)
GDPpc = (Apparent productivity) * (Employment rate) * (Share of active age population)
Measures of convergence
Sigma-convergence refers to a reduction of disparities over time between countries, for instance, measured in terms of the standard
deviation or coefficient of variation (the ratio of the standard deviation to the average). Beta-convergence refers to a situation where
incomes in poorer countries grow faster than those in richer ones, usually measured in terms of change over time. The two concepts
of convergence are closely related with Beta-convergence being necessary but not sufficient to achieve Sigma-convergence (see, for
instance, Monfort, 2008).
Other indices exist (for instance, the Gini coefficient, the Atkinson index, the Theil index and the Mean Logarithmic Deviation). It is recom -
mended that we ‘consider a variety of measures to draw firm conclusions about changes in the extent of disparities’ (see, for instance,
Montfort, 2008), and the analysis in this chapter focuses on the coefficient of variation as a main measure of sigma-convergence,
complemented as regards within zones and between zones dispersion by a standard between-within decomposition of total variance
and a decomposition of the Theil index (see Annex 3). An emphasis in the main text is put on the decomposition of total variance
which is closer to the measure of the coefficient of variation and, more specifically, on the share of total variance corresponding to the
between zones component (as the level of variance per se can be misleading, since it is affected by homothetic changes which do not
affect dispersion, the Annex provides additional elements on the level of the between zones contribution to total variance expressed
as an index, based on the first year when data are available).
( ) Furthermore, while entry into the euro is conditional on fulfilling the Maastricht criteria, the euro is intended to support real convergence, defined in
1
terms of per capita GDP, by fostering economic integration (see European Commission, 2008).
2.1.2. Convergence in Europe, • EU-15 Centre (Belgium, Luxembourg, represented 26 % of EU-28 population
trends between and within zones the Netherlands, Germany, Finland, in 2013.
France, Austria) ( ), which represented
8
In order to provide an overview of 36 % of EU-28 population in 2013. • EU-13 Centre and North (Czech Republic,
employment and social convergence Hungary, Poland, Slovenia and Slovakia),
trends in Europe (EU-28) overall, it is use - • EU-15 North (Denmark, Sweden, which represented 13 % of EU-28 popu -
9
ful to reflect not only on overall develop - United Kingdom) ( ), which represented lation in 2013.
ments, but also on changes in dispersion 17 % of EU-28 population in 2013.
both within and between zones. For this • EU-13 South and periphery ( Bulgaria,
purpose, five groups of countries are • EU-15 South and periphery (Greece, Cyprus, Estonia, Latvia, Lithuania, Malta,
considered, reflecting socioeconomic and Ireland, Portugal, Spain, Italy) ( ) which Croatia, Romania) which represented
10
geographical proximity criteria: 8 % of EU-28 population in 2013.
( ) Or in other terms EA-12 Northern countries,
8
see European Commission (2014a).
( ) Which are actually EU non-EA countries.
9
( ) Which are actually EA-12 South and periphery
10
countries, see European Commission (2014a).
205
