Finance: Research, Policy and Anecdotes

Why Paul Romer is wrong (again)!

Let me first say that I have enormous respect for Paul Romer’s contribution to modern growth theory. I first encountered his work as undergraduate student in Germany when I had to write a term paper about one of his papers. Many theoretical models that underpin the empirical finance-growth literature build on endogenous growth theory, so his contribution cannot be underestimated and he is well deserving of the Nobel Prize in Economics.  I have somewhat less respect for him when it comes to his excursion into the policy world, most recently as Chief Economist of the World Bank, a position which he had to leave early due to his rigidity and due to wrong claims he publicly made about the Doing Business methodology and which did cause a minor diplomatic incident between Chile and the World Bank.

 

In a recent op-ed for the Financial Times, Paul Romer gives the still-to-be-elected new president of the World Bank two recommendations, one of which drew my attention. “First, outsource the bank’s research upon which it depends for identifying problems and proposing solutions.”  His main argument: “When complex political sensitivities are allowed to influence research by stifling open disagreement, it ceases to be scientific.” He is certainly not the first to make such a case; others have argued for leaving academic research to universities rather than international financial institutions (IFIs) such as the World Bank or the IMF. I think such a recommendation is wrong, for several reasons (this might be a good point to mention that I am still working as consultant for both World Bank and IMF, so, yes, my views are biased by personal experience and – possibly – personal interest).

 

First, combining policy work with research can be fruitful, both in terms of policy questions/challenges giving rise to research questions and in terms of research feeding back into policy debates. A lot of my early work on financial inclusion (together with Asli Demirguc-Kunt and Sole Martinez) was motivated and informed by policy discussions in the developing world (I still remember the Kenyan Finance Minister asking us in 2003 why bank account fees were so high in his country, a question we somewhat answered in a subsequent paper).  Could such work been undertaken in academia?  Yes, but collecting information from banks and regulators is hard to undertake by academics without the backing of a large IFI such as the World Bank. Further, being challenged on an almost daily basis by operational colleagues can be very helpful in taking a sufficiently granular view in such research (e.g., when designing questionnaires, exploring alternative explanations or understanding the political economy of financial system structures).

 

Second, research done within multinational institutions can result in a dialogue within the institution, much more than research done outside.  Best example is the research undertaken by my former colleague and boss Asli Demirguc-Kunt on deposit insurance in developing countries. While the standard recommendation by IMF and World Bank previously had been to introduce deposit insurance wherever it had not existed before, Asli with several colleagues (some in academia) showed the moral hazard risk, with too generous a deposit insurance raising bail-out expectations and ultimately resulting in the systemic banking crisis it was designed to prevent in the first place.     Yes, such research can be undertaken and had been undertaken in academia before, but mostly if not exclusively focused on US and other advanced countries.  With a publication bias towards US data still very prominent, researchers in multinational institutions have a critical role to play in research on developing countries.

 

But why can such cross-country data not be collected by multinational institutions and then primarily used by academic researchers? Data to be used by researchers is best collected by researchers who can link specific research questions to data and information to be collected. One good example is the Banking Environment and Performance Survey undertaken by the EBRD across Central and Eastern Europe and Central Asia (with data collection facilitated by the relationship the EBRD has with many of these banks), a first round in the early 2000s but with the questionnaire primarily designed by operational staff, while a second round was undertaken in 2011, with the questionnaire designed jointly by EBRD research staff and academics (like yours truly) with very specific research questions in mind, subsequently resulting in several top publications.

 

But what about political sensitivities? Yes, they exist, though in both IFIs and academia. Several years ago I was involved in a research project sponsored by the European Fund for South Eastern Europe and while ultimately everything turned out just fine, there were quite some (sometimes tense) negotiations with the funder and with the banks providing us with data during the process. So, it is not only research undertaken in IFIs where political sensitivities can play a possible role. Where (empirical) research depends on micro-data provided by financial institutions or government agencies, one can never completely ignore political or institutional sensitivities.

 

Where does this leave us? I would argue it leaves us exactly where we are right now – close cooperation between researchers in IFIs and academia, where either side can take the initiative and where cooperation will ultimately result in policy-relevant research with important policy implications. So to come back to Paul’s recommendation: what we need is not outsourcing of research, but increased cooperation and strengthening the independence cum accountability of the research department within the World Bank.