Thursday, June 10, 2010

The transformation of development thinking

David Lindauer and Lant Pritchett discuss the transformation of development thinking, development policy making, and the next big idea for growth. Though this is an old paper, it is quite interesting. May be they should update it considering the ideological shift economics fraternity has seen after the global economic crisis starting 2008.

How has the thinking on economic development changed after 2002 (and before 2008)? It might take different forms depending on several factors that affect the global and local markets. But one thing is certain: “The road to development is extremely complex, and the ultimate guide to that path must therefore be more complex than an arrow pointing confidently in one direction.”

Growth experience between 1980 and 2002:

  • The enormous slowdown in growth that has occurred throughout the developing world: the so-called lost decade(s) in Latin America has its counterpart on other continents. The median growth rate in low- and middle-income economies fell from 2.5 percent in 1960–79 to 0.0 percent (zero!) in 1980–98.
  • The long-awaited transition from stagnating Marxist central planning to a capitalist economy has gone horrifically worse than anyone would have dared predict. While the consensus was that the economies would experience a dip in income as they restructured and as resources were reallocated from old to new activities, no one predicted in 1992 that income in many newly capitalist countries in 2002 would be less than half the level under the Communists.
  • The financial crises—or perhaps the single rolling financial crisis—of the 1990s: Mexico in 1994; Thailand, Korea, and Indonesia in 1997; Russia and Brazil in 1998; Ecuador in 1999; Turkey in 2000; and Argentina today. In each case, something caused a near or actual debt default or a large depreciation (or both).
  • The collapse of sub-Saharan Africa, which has by now become so complete as to force itself into world consciousness. Nearly all of sub-Saharan Africa has been transformed from the heady optimism and enormous promise of early independence to almost unspeakable suffering. The only relatively consistent success cases—Botswana and Mauritius—do not give much hope for an entire continent. The AIDS crisis is so severe that the only historical analogue is the Black Death.
  • The world’s two most populous countries, India and China, have grown rapidly. This is somewhat puzzling because in many ways these countries are slow, cautious reformers that remain among the more closed and restricted economies in the world. Both India and China accelerated by more than 2 percentage points, while the rest of the world decelerated by more than 2 percentage points.

This gives a mixed signal to what and what kind of political and economic system propel growth. So, what should be the agenda for post-Washington Consensus economic growth?

Lindauer and Pritchett argue that a new agenda should “acknowledge the complex, state-dependent, contingent relations between policies and growth.” While doing that there is also a need to recognized that growth regressions have failed to give a clear picture of the different growth experiences of similar and dissimilar countries.

The basic flaw in growth regressions is that they confuse partial correlations with (stable) parameters and confuse empirical variables(that might be associated with policies) with feasible actions to promote growth.

By now, there are thousands of papers that put economic growth on the left-hand side and other stuff on the right-hand side. This research produces empirical findings that are translated, more or less crudely, into policy recommendations: a partial correlation of lagged enrollment rates and subsequent growth is interpreted as proof that education is good for growth, which is then used as the basis for recommending more public spending on education; a partial correlation of trade and growth outcomes indicates that openness is good, which becomes a recommendation to reduce tariffs; a partial correlation of inflation and growth shows that low inflation is good, which leads to the adage that fiscal austerity will promote growth. Sadly, many of these recommendations have not worked in practice, in part because nearly all of the growth regression research is essentially irrelevant to policymaking and policy implementation. The findings do not constitute a credible basis for meaningful development advice, since they are not empirically stable and they are not about policy.

It is not surprising that growth regressions are unstable across countries and over time, because any model in which growth is linearly (or log-linearly) related to any given variable across countries, time periods, levels of development, and circumstances is almost certainly misspecified.

Even growth regressions with right handside variables under direct policy control—say, the budget deficit or tariffs— are still describing a relationship between policy actions and outcomes, not policies and outcomes. A policy is a mapping from states of the world to policy actions.

What is next after the terrible performance and use of growth regressions? They list some better ideas to study growth.

  • A number of intriguing leads suggest future directions for research. One branch of such literature is episodic analysis, in which researchers examine episodes of more or less discrete changes in policy or intermediate outcome variables.
  • A second approach is to examine economic growth directly, either to
    attempt a close reading of the factors that initiated a growth boom or bust or to undertake cross-sectional work explaining changes in growth rates.
  • Although growth regressions are inherently hopeless, a few authors work from the growth partial correlations to plausible microeconomic mechanisms that explain the growth correlation, and then to the actual policy variables that could alter outcomes.

(They argue that economics profession does not need any new great idea in economic growth. It just needs to, as Rodrik says, “figure out how to turn sound economic thinking into useful, context-contingent policy recommendations.”)

For policy advice they propose a diagnostic tree of policy options. They branches they propose are (i) current level of income, (ii) current status of growth, (iii) linkages with the world economy, (iv) government strength, and (v) government capacity.