Neoclassic growth theory suggests that different capital and labor productivities across countries lead to differential productivity growth rates and hence to conditional convergence across different economies. Rostow (1990) coined the term that the "poor get richer and the rich slow down." The convergence theorem of neoclassic theory arises from diminishing returns on capital. Economies that tend to have less capital per worker tend to have higher rates of return and hence higher growth rates (see Abramovitz, 1986). Conversely, economies with high capital intensity (which, because of the relationship between capital intensity and productivity, are closer to or at the productivity frontier) tend to have lower growth rates. Evidently, economies differ in more respects than their capital intensities, and hence even neoclassic theory only postulates conditional convergence (after accounting for all other factors).
The neoclassic concept of capital is usefully extended to include also human capital in the form of education, experience, and health (see, e.g., Lucas, 1988; Barro and Sala-I- Martin, 1995). Thus, additional convergence potentials accrue for economies with a proportionally higher ratio of human-to-physical capital. Equally, the generation and adoption of new technologies is facilitated by high human capital. Yet, even with the inclusion of human capital, long-term per capita growth must eventually cease in the absence of continuous improvements in technology. This cessation, however, mostly affects economies at the productivity frontier and not those that lag behind. For the latter, both theory and empirical evidence seem, all else being equal, to indicate conditional convergence (i.e., "the poor can get richer."). The ceteris paribus condition is an important qualifier for the convergence theorem. Evidently, the potential for conditional convergence and economic catch-up cannot be realized in an economy struck by civil war, poor institutions, or even low savings rates (related to the demographic transition discussed in Section 3.2). The recent Asian financial crisis also demonstrated that differential capital productivity indeed can lead to vast influxes of capital into developing economies, and appropriate assimilative capacities (banking systems, functioning legal system, institutions, etc.) need to be in place to use such capital flows productively (World Bank, 1998b).
|
Figure 3-11: Residual GDP per capita growth rates as a function of GDP per capita (log scale). The residual growth rate is that per capita GDP growth not explained by other factors such as education, terms-of-trade, institutional factors, etc., in Barro's multi-factor analysis of per capita GDP growth. Data source: Barro, 1997. |
In terms of a functional relationship, therefore, per capita GDP growth rates are expected to be higher for economies with low per capita GDP levels. Notwithstanding many frustrating setbacks, such as the recent "lost decade" for economic catch-up in Africa and Latin America, empirical data indicate that the convergence theorem holds. Figure 3-11 illustrates some empirical evidence put forward by Barro (1997) based on the experiences of some 100 countries in the period 1960 to 1985.
Similar convergence trends have also been identified within economies. For instance, Barro and Sala-I-Martin (1995) find significant convergence trends across individual states in the USA, between prefectures in Japan, and between different regions in Europe. Barro (1997) concludes in his analysis that the conditional convergence rates across these countries is statistically highly significant, and proceeds rather slowly at 2-3% per year. It may take an economy 27 years to reach 50% of steady-state levels (the productivity frontier) and some 90 years to achieve 90% of that level. Based on this convergence criterion alone, it may well take a century (given all other factors set favorably) for a poor economy to catch-up to levels that prevail in the industrial countries today, never mind the levels that might prevail in affluent countries 100 years in the future. Barro's analysis indicates a threshold GDP per capita level at approximately US$3000 per year. Below that level, additional productivity growth potentials result from catch-up; beyond that level, higher per capita GDP levels make further productivity growth ever more difficult to achieve (as indicated by the negative values of the residual GDP per capita growth rates in Figure 3-11).
Given the wide range in historical experiences and the slow rates of convergence suggested by neoclassic growth theory, it is not surprising that the available scenario literature takes a cautious view on economic catch-up. Whereas convergence tendencies are generally evident in scenario assumptions (see the significantly higher GDP per capita growth rates for currently developing countries compared to industrial countries in Figure 3-10), long-term convergence rates are low. For instance, from all six IS92 scenarios only one (IS92e) assumes that developing countries outside China may eventually reach present OECD income levels, and even in this most optimistic scenario it is assumed to occur only after 2080 (Pepper et al., 1992). Even in this convergence scenario per capita income differences remain large - a factor of five by the end of the simulation horizon (US$31,000 per capita GDP per year in developing countries outside China versus US$150,000 OECD average). In an influential critique Parikh (1992) referred to the IS92 scenario series as being "unfair to the South," a point also taken up in the evaluation of the IS92 scenarios. Alcamo et al. (1995) concluded that new IPCC scenarios "will be needed for exploring a wide variety of economic development pathways, for example, a closing of the income gap between industrial and developing countries." With a few notable exceptions (e.g., the scenario developed by Lazarus et al. (1993) and the Case C scenarios presented in IIASA-World Energy Conference (WEC) (IIASA-WEC, 1995) and Nakic�enovic� et al. (1998a)), the challenge to explore conditions and pathways that close the income gap between developing and industrial regions appears to have been insufficiently taken up in the scenario literature, a gap this report aims to begin to fill. Chapter 4 describes two scenarios in which the ratios between regions of GDP/capita decline and the absolute differences increase.
|
Other reports in this collection |
|
