Paul Krugman is one of the few economists at home both in `high theory' and in public economic discourse. He thinks deeply, and he thinks brilliant thoughts. This little book - based on the Olin Lectures he gave in Stockholm - is proof of what his mind can yield, when it sets out to clarify issues.
Development and economic geography, he argues, failed because they did not submit themselves to the discipline of model-building - what might look or even be at first sight downright silly in the end is preferable to the unconscious metaphors of the narrative economic discourse.
For all its clarity, Krugman's argument is deeply flawed. Development and economic geography - together with income distribution - belong to the derelict class of economic problems that addresses the question of historical disparities of wealth in the economic tissue. Why have some countries or regions developed and others have staid behind, why are there poor and rich? Was it done by better use of the available resources, or by impoverishment of other nations or persons? A corollary to this question would be: does our quest for efficiency worsen or reduce disparities? Both Adam Smith and Karl Marx addressed this question, but their observations have been largely forgotten. Pareto and welfare economics picked up the thread, only to conclude platidinuously that the only `good' policies are those that benefit all.
Should the model-building solutions that Krugman suggests be used in development and geography be any good, they might imply that a `big push' applies not just to economic growth, but also to concentration of income - consumer surplus playing the role of `economies of scale'. Interesting. Just as interesting as the metaphor that - as in the `big bang' theory of star formation - the smallest of initial income irregularities (e.g. first predatory capital accumulation) lead to the agglutination of wealth around capitalists. Which, of course, also implies that it is the 90% of dark (workers) matter that keeps the shiny capitalist `stars' in place in a well-ordered and expanding economy.
Toys are useful provided they teach a child the `real thing'. Toy models are not useful when they fail to recognise (let alone address) fundamental issues like that of economic disparity. Models are downright bad when their incautious use leads to blind-sighting in economic policy. Every economist should be made to ponder Kenneth Arrow's Theory of Second Best. Partial optima are bad solutions in the search for an overall optimum.
Can we expect models of income disparity soon? Paul Krugman might devote some of his intellectual powers to construct the simplest of models of income disparity and attempt to integrate it into a growth model - just to disprove (or prove) the widespread intuition that when governments pursue efficiency single-mindedly, the rich get rich and the poor poorer.
Can we further expect a `grand unified theory of everything economic' that would bring together both concerns of efficiency and income distribution into a unified model for development? Don't hold your breath. As Koopmans famously proved, one cannot kill two birds with one stone. Until then, however, efficiency models should either be denied the Warrant Of Fitness for circulation in political circles, and/or carry the label: Efficiency may be harmful to income distribution.


