Comments and Discussion

John Williamson:

I found this a highly stimulating paper. It works from the fact that the aggregate production function is, as we all know,

The traditional model that we all know and love, which Easterly calls "Factor World," assumes that A is common across countries and that all the action comes from variations in (K/L). But way back half a century ago, in one of the very first serious empirical papers in international economics,Wassily Leontief set out to verify the theory and discovered to his discomfort that the United States exported labor-intensive goods and imported capital-intensive goods.¹ He asserted that the problem must lie in bad measurement rather than in the theory; since U.S. workers were really three times more productive than others, one should multiply the U.S. labor force by three, which would resolve the paradox by turning the United States into a labor-abundant country. What Easterly does is provide us with a more congenial way of resolving the paradox: postulate that we live in "Productivity World," where it is A rather than K/L that varies across countries. The Leontief Paradox then falls into place as one more piece of evidence suggesting that this, and not Factor World, is a better representation of the world we live in.

Productivity World is not as endearing as Factor World. Beyond the issue of our familiarity with the latter, there is the problem of understanding. There is a body of theory that tells us how K/L changes, since ΔK = I and there is a literature about what determines I (even if a lot of it is confined to postulating that I = S). In contrast, very little is understood about what lies behind changes in A. Nowadays the tendency is to believe that it stems from "things that people do."² We believe that their willingness to do these things depends upon a society's institutions (including trust). We take it for granted that technology is involved. And we think that maybe human capital should be given a role, though we wonder whether human capital should be modeled by A = f(H), or as causing a change in A, or as a third factor of production. Nevertheless, our understanding of the determinants of A is rudimentary.

I am sympathetic to the author's bottom line, as I will emphasize shortly, but I nonetheless found myself unconvinced by several of his arguments. First, he claims that "the very different performance of developing country regions does not have any obvious Factor World explanation." And yet, it was not that long ago that Alwyn Young and Paul Krugman upset the East Asians by telling them that their rapid growth, like that of the erstwhile communist world,was all driven by factor accumulation rather than productivity growth.³ It is an undisputed fact that investment is far higher in East Asia than in Africa or Latin America. (It is intermediate in South Asia, like that region's growth performance.) Easterly's main positive argument for disputing the importance of factor accumulation in driving growth is that it implies counterfactual predictions for the returns to physical and human capital (for example, wages of engineers vastly higher in India than in the United States). However, there are other assumptions that go into that calculation, like the production function really being Cobb-Douglas and distribution really being determined by marginal productivity. Is it inconceivable that some of those other assumptions might be in error?

Second, he argues that capital and skilled labor are flowing to the richest countries, phenomena that support the view that we live in Productivity World. In both instances it seems to me that the facts are murkier than he represents them. Yes, there is some capital flight from many poor countries, and from some it is substantial (as shown in his table 2). There is also one large rich country that imports large sums of capital. But he does not present figures to challenge the view that the net capital flow has nonetheless been from the rich countries to the poor. With skilled labor...