by Francesco Caselli and Silvana Tenreyro
No. 2, September 2004 - December 2004

Motivation for the Research
Between 1950 and 2000, real labor productivity in some of Western Europe’s richest countries was more than three times that of some of its poorest. By the end of the century, this ratio was well below two. A notable aspect of the decline in cross-country European inequality is the catch-up by the Southerners: Italy first, then Spain, Greece, Portugal, and eventually Ireland (a Southerner in spirit) all had spurts of above-average productivity growth.

Spain’s experience is emblematic: In less than 15 years between the late 1950s and the early 1970s, its labor productivity relative to France’s (the authors’ benchmark for the “average” European experience) went from roughly 65 percent to over 90 percent.

On May 1, 2004, the European Union (EU) admitted 10 new members, primarily from Eastern Europe. To varying degrees, the Easterners’ current relative labor productivities are similar to those of the Southerners before their convergence spurts. This widely noted analogy has given rise to hope that the Easterners will be the new Southerners, and Poland, the new Spain. Indeed, this hope is one of the reasons these countries have wanted to join (and several others hope to join) the EU.

With so many people pinning such great hopes on the continued ability of the European club to generate convergence among its members, this paper revisits the data on the relative growth performance of European countries in the second half of the twentieth century to assess the prospects of the Easterners’ repeating the Southerners’ success.

Research Approach
The overall approach is to look behind the aggregate labor productivity numbers and present a number of different approaches to decompose the overall convergence experience into more disaggregated processes. The discussion is organized around four hypotheses potentially explaining the convergence process:

1. Grounded in the Solovian-neoclassical hypothesis, the first hypothesis holds that initially capital-poor countries have higher marginal productivity of capital, and hence faster growth.

2. The second hypothesis explains the convergence process as the result of technological catch-up. Backward countries converge to the technological leaders mainly through a process of imitation.

3. The third hypothesis interprets the convergence process as driven mainly by gains from trade from European integration, which may have been disproportionately larger for the poorer economies (as a proportion of GDP), both because of their initially more autarchic status and because of their relatively smaller size.

4. The fourth hypothesis holds that convergence is a by-product of structural transformation, which is partly a process of reallocation of resources from low-productivity to high-productivity sectors. If initially poorer countries had a longer way to go in this transformation, this process may itself have been a source of convergence.

Key Findings

• The data suggest that a critical mechanism for Spain’s explosive catch-up was a vast redeployment of labor out of agriculture towards higher value-added sectors.With a much larger agricultural sector than France, Spain benefited disproportionately.
  
  
• In general, Southerners converged to the rest mainly through a faster rate of reallocation of the labor force from low-productivity agriculture to high-productivity manufacturing and services, although in some cases, including that of Spain, within-industry catch-up was also quite important.
  
  
• With respect to the relative contributions of capital deepening and technological change to the reduction of European inequality between 1960 and 2000, physical capital accumulation and total factor productivity (TFP) growth were roughly equally important, although TFP was not always initially lower in poor countries. However, the contribution of human capital accumulation— at least as measured by years of schooling—was negligible.
  
  
• Despite substantial convergence, Spanish average labor productivity has hovered at around 90 percent of French average labor productivity since the mid 1970s. The data indicate that this persistent remaining gap is due mostly to an equally persistent gap in human capital per worker.
  
  
• The new and forthcoming EU members—including Poland, a focus of the authors—exhibit very large labor productivity gaps vis-à-vis Western Europe, and the gaps are accounted for by disparity in physical capital and TFP, but not by disparity in human capital. These countries have substantially larger shares employed in agriculture, which tends to be the least productive sector, than do the Westerners, and their manufacturing and services sectors are also less productive than the corresponding sectors in the West—although the gaps are not as large as in agriculture. There is, therefore, scope for large productivity gains both via labor reallocation out of agriculture and through within-industry catch-up.
  
  
• However, in Eastern Europe the distribution of employment among sectors is much less important as a source of productivity gaps vis-à-vis the rest of Europe than it was in Southern Europe in 1960.
  
  
• For Poland, and for the Easterners generally, the road to convergence passes through physical- capital deepening and through TFP gains at the industry level. This means that convergence may take quite a bit longer than it did for the Southerners. On the other hand, unlike Spain and the Southerners generally, Poland and the Easterners can anticipate a complete catch-up, as they are not hobbled by a human-capital handicap.

Implications
Implications for the first two hypotheses are mixed. Poorer countries experienced faster physical capital deepening, and this explains about 50 percent of their relative gains. Poorer countries also experienced faster TFP growth, accounting for the remaining 50 percent. However, neoclassical growth theorists may be disoriented by the lack of convergence in human capital. And endogenous growth theorists may be disoriented by the fact that not all initially poorer countries lagged the rest technologically, so that their continued faster TFP growth does not square well with the technology catch-up story that these theorists would probably favor.

As an explanation for regional convergence, the trade view (the third hypothesis) runs into some problems: For example, countries with a comparative advantage in agriculture tend to show systematically lower shares of agriculture.

The structural transformation approach fares better as an explanation of convergence by the Southerners in the latter half of the twentieth century. As mentioned above, the Southerners converged to the rest mainly through a faster rate of reallocation of the labor force from low-productivity agriculture to higher-productivity sectors.

Full text of Public Policy Discussion Paper 04-8