Academic Macroeconomics and Policymaking
My BloombergView colleague Noah Smith recently wrote a column in which he asserts that academic macroeconomics is of limited use to economic policymakers such as the Federal Open Market Committee (FOMC). Based on my experience as a policymaker and an academic, I would (somewhat grudgingly!) admit that there is more than a little truth to Noah’s claim. But I don’t think that it is for the reasons that he cites in his piece (false assumptions, lack of empirical fit, etc.). These are, after all, features of all modeling approaches, including ones that policymakers do find of use.
In this post, I argue a different perspective: Academic macroeconomics is specifically designed to be of limited use to policymakers like the FOMC. Academic macroeconomics conceives of policy as a regime (or program or rule) that describes how policymakers will respond to all possible economic contingencies. The science of macroeconomic policy is thought to be about understanding the impact of different regimes on the economy. Such a science should be of potential interest to, say, legislators interested in making long-term (semi-constitutional) reforms in an economy. But it is (again, by design) of little use to policymakers like the FOMC who are asking questions like: Should we raise interest rates at this meeting? I suggest a path forward for academic macroeconomics to make it of more use to such decision-makers.
Rules as the Only Legitimate Objects of Scientific Inquiry
The current thinking about policy in academic macroeconomics can be traced in large part back to the writings of Professor Robert E. Lucas, Jr., of The University of Chicago in the 1970s. Lucas pointed out that, if private sector agents are forward-looking (not necessarily rational!), then their reaction to the specification of a policy choice in a particular economic contingency necessarily depends on their beliefs about how policy choices will change in response to future economic contingencies. Any policy analysis has to fill in those beliefs in some fashion.
Lucas argues that if we think of policy as a series of isolated decisions, there is no systematic way of assessing (especially quantitatively) how one of those decisions will affect agents’ beliefs about future policy choices. That in turn means that we have no way of understanding or modeling the impact of the decisions themselves. Instead, the macroeconomic scientist should view such one-time, isolated, decisions as being outside the scope of inquiry. Macroeconomists should focus on assessing the impact of regimes or rules which describe how policy will react to all possible future contingencies.
Lucas summarized his conclusions in at least a couple of places in a characteristically forceful fashion. In his Critique, he writes, “The only scientific quantitative policy evaluations available to us are comparisons of the consequences of alternative policy rules.” (underlining his). In a beautiful (but less-known) paper called, “Rules, Discretion, and the Role of the Economic Adviser,” he writes, “… recent work on expectations …. emphasizes the fact that analysis of policy which utilizes economics in a scientific way necessarily involves choice among stable, predictable policy rules, infrequently changed and then after extensive professional and general discussion.” This is not an approach to policy that is intended to be of use to the FOMC when it meets later this week.
Over forty years later, we can still see the impact of Lucas’ powerful conclusions in most (all?) academic analyses of monetary policy. The central bank is modeled as a rule that maps economic conditions into interest rates (or asset purchases), perhaps with some random noise. The analysis then centers on how different rules affect economic outcomes.
A Different Way Forward?
When I go back and read Lucas, his thinking seems to me to be very classical (in an econometric sense). When I was a policymaker, I operated in a much more Bayesian way. Taking that kind of approach could allow academic macroeconomists to be of more use to policymakers who have to make some kind of isolated decision. Here’s a sketch of what I mean.
A policymaker has to decide between two options (like raise rates or not). The impact of both options on economic variables of interest is necessarily uncertain. Part of that uncertainty is attributable to the fact that the two options will have different effects on the beliefs of agents about future policy decisions. The policymaker probably starts with prior beliefs of his/her own about the impact of the different options on variables of interest.
In this context, the goal of a macroeconomic adviser is to use all available data (and possibly collect new data) to help the policymaker update his/her prior. The resulting posterior is likely to feature a lot of dispersion. Indeed, in many circumstances, it is possible that the updating might end up being relatively minor (because the available data aren’t that informative).
Buried in the emboldened sentence is a very different vision for macroeconomics than what is typically practiced in the academe. I suspect that it would lead us in a more empirical and less theoretical direction. We would have to learn to live with (and report) a considerable amount of residual uncertainty. Is such an approach “scientific”? I don’t know! But I believe that economists have a lot of tools and data that can help policymakers in updating their priors. And it seems to me like we in the academe could be doing a lot more to systematize that updating process.
Rochester, NY, June 12, 2016
Please address media enquiries and non-academic speaking requests to Monique Patenaude (firstname.lastname@example.org and 585-276-3693).