Growth theory and the wealth of nations (1)

Our friend Ap from cafe salemba, and frequent visitor of "ruang 413" (literally...in the real world) is starting his stint as guess blogger this month. So, without futher ado, here's what Ap has written for our enjoyment...

Growth theory and the wealth of nations

Judging from his posts here, here and here, Yudo seems to be in the 'fascinated' stage with the (neoclassical) growth theory. Indeed, since Robert Solow's 1956 paper, the field of economic growth has become a very interesting subject. The paper provided a scientific framework to analyze what makes a country grow faster while the other don't.

Then, thanks to the availability of both empirical data on cross-country national income, and the advance of statistical tools, starting in the 1980s-1990s economists were able to do empirical analysis on economic growth. Some people have even devoted their life and career on doing such regression, like Harvard's Robert Barro.

The basic, fundamental question for doing cross-country growth analysis is: why some countries are rich and some other poor? Moreover: why some countries have successfully become rich while some others failed?

The irony is, the more we try to explain it, the more likely we fail to do so. The reason became the mechanism of economic growth is very complex to fit into a single model. Consider Solow's famous model. As explained by Yudo, the model predicts that a country's income per worker will grow faster if: i) it can accumulate physical capital ('investment') faster than the rate of its population growth and depreciation ('net depreciation'), ii) its capital-per-worker stock is relatively low.

By way of corollary, Solow model than provides the basic for the famous 'convergence' analysis: at a given rate of investment and net depreciation, poorer countries would grow faster in terms of income per worker, and soon they will catch up the richer ones, who are growing only at the 'steady-state' rate.

In the context of economic development, Solow model suggests that:

1. Level of physical capital should explain cross-country income differential.
2. Investment should affect only the short-run economic growth.
3. In the long-run, economic growth would be determined by the growth rate of something exogenous. Solow labeled it A for technology. What is A, and what affects it, we don't know (he didn't provide any explanations). It is something like the 'manna from heaven.'

When people did the empirical testing for Solow model, (1) and (2) are generally confirmed, but unsatisfactorily. For example, Mankiw (1995) found that variation of physical capital only explains about a third of cross-country income per capita differences in the 1990s. He suggested the term 'capital' should be re-formulated by including 'human' capital. Using secondary school enrollment rate as the proxy for human capital, Mankiw showed that the combination could explain around 80% of income differences.

But still the puzzle still remains. Does secondary enrollment make a good proxy for education in general? Does it mean that poor countries should invest more in schooling? One may tempted to say 'yes.' But other empirical analysis show that Africa remains poor despite heavy investment in education (in terms of school enrollment), as mentioned by William Easterly in his 2001 book. (A chapter in the book discussed how the idea of controlling population growth also did not work well).

There are even bigger puzzles. First, the fact that rich countries are able to sustain, even increase, their growth rates for a very long period. This suggests that the exogenous variable A is more important that just a residual. Perhaps it is not even exogenous; it may be endogenous .

Second, the model has little to say about historical path of economic development. Some today's poor countries were not poor some hundreds and thousands years ago, while the rich group only recently – in the human civilization history scale – became rich. What explains this reversal of fortune? It is very likely that the current per-capita income differences that we observe now have a long and deep historical determinants.

Therefore, the basic, fundamental questions I raised earlier still unanswered by the existing growth models.