I am not surprised by the reactions of Mennes, Gunning and Elbers to my column, ‘Weak links’ (The Broker 10). In a nutshell, they argue that cross-country regressions may be imperfect but can be useful nonetheless, and that properly conducting such regressions is not a trivial exercise. I do not disagree. Indeed, as observed by the critics, I sometimes use this methodology myself. Cross-country regressions make it possible to explore issues that are hard to analyze otherwise. While most individual studies suffer from shortcomings, together they comprise an evolving ‘body of knowledge’ that could inform policy makers and academic.

So what is the issue I tried to raise in my column? The point was obviously not too warn economics professors about the dangers of cross-country regressions. In fact, as I wrote earlier, ‘Most economists are well aware of the potential pitfalls of such work, and take the outcomes with more than a single grain of salt’. However, the majority of The Broker readership are not economics professors. And to my experience such ‘non-insiders are easily overwhelmed by impressive tables suggesting ‘hard’ and precise results – conveying the illusion of deep truths’. The point was to convey to a broad audience that cross-country results – and especially individual studies – are often far from robust truths. It is instructive to quote two undisputed experts, Brock and Durlauf (WBER 2001:232)*, on this:

This literature does rely on assumptions that may be argued to be […] dubious and whose implausibility renders the inferences typically claimed by empirical workers to be […] suspect”

As a result, there are studies to support almost any opinion or policy position – trade can be ‘good’ or ‘bad’ for economic growth, and institutions may be ‘key drivers’ of development, or not. Evidence for all of this, and much more, is ‘out there’. It takes some effort and skill to see the forest for the trees.

Upon additional reflection, there are two points I would like to make about cross country regressions. First, there is too much econometrics that is not guided by theory or informed ideas. Deaton & Miller (1995, p.11) argue that econometric models should complement case studies. Case studies and theory generate in depth insights and hypotheses, and econometrics can be used to explore whether these insights hold up to scrutiny and may be generalized, or not. Too often, however, ad hoc econometrics is agenda setting.

Second, it would perhaps be best if everybody would acquaint himself with the details and pitfalls of cross-country studies so as to be able to “weigh” the evidence that is presented. However, in the absence of such a Nirvana, it would simply be good to treat the various models and pieces of evidence with caution. This was the only message I could squeeze into a 600 words column.

*Reference

Deaton, A. & Ron Miller (1995) International Commodity prices, macro-economic performance and politics in sub-Saharan Africa. Princeton Studies in International Finance, 79, Oct 1995.