Monday’s Macro Memo: Behavioral New-Keynesian Macroeconomics

This is my penultimate post featuring research presented at the conference on Applications of Behavioural Economics and Multiple Equilibrium Models to Macroeconomics Policy Conference held at the Bank of England on July 3rd and 4th 2017.

Post Date
11 September, 2017
Reading Time
7 min read


This is my penultimate post featuring research presented at the conference on Applications of Behavioural Economics and Multiple Equilibrium Models to Macroeconomics Policy Conference held at the Bank of England on July 3rd and 4th 2017.

                              Today’s post features two single authored papers: one by Xavier Gabaix and one by Michael Woodford.


Xavier Gabaix, the Pershing Square Professor of Economics and Finance at Harvard University, is one of our leading young economists today. Michael Woodford, one of two keynote speakers at our conference, along with Jim Bullard, is the John Bates Clark Professor of economics at Columbia University. Michael has had a distinguished career as one of the founders of New-Keynesian economics and has been the recipient of many awards.


Xavier has an exciting research agenda that combines ideas from psychology and economics. He is a prolific author who has worked on topics in finance, macroeconomics and behavioural economics.


In Xavier’s own words, “economists usually assume that people know how the economy works. This is a bit strange since economists don’t even know how the economy works”.

These are wise words that you should bear in mind when you assess the credibility of many so-called experts!  You can hear Xavier explain his work in the linked video clip.

Xavier has written a series of papers that have this same theme. In his 2016 NBER working paper, “Behavioral Macroeconomics Via Sparse Dynamic Programming” Xavier asks, rhetorically, “How will a boundedly rational agent behave?” He goes on to answer his own question:


I assume that the agent starts with a much simpler model, [than the complete infinite-horizon stochastic dynamic programme] … where the interest rate and income are constant — this is the agent’s “default” model. …. He knows what to do [in the default model] … (consume a certain fraction of his wealth, and permanent income), but what will he do in a more complex environment, with [a] stochastic interest rate and stochastic income?

Xavier’s answer is that the agent considers approximations to the complete decision rule that are sums of simpler components

and …, for each component, he asks whether it will matter enough for his decision. If a given feature (say, the interest rate) is small enough compared to some threshold … [the agent] drops the feature or partially attenuates it. The result is a consumption policy that pays partial attention to income and possibly no attention at all to the interest rate.


At the conference Xavier presented a paper related to this research agenda, “A Behavioral New Keynesian Model”,  in which he applies the idea of sparse dynamic programming to monetary policy. Xavier builds a model that adds a

… a new parameter, [to the New-Keynesian model] which quantifies how poorly agents understand future policy and its impact … That myopia parameter, in turn, affects the power of monetary and fiscal policy in a microfounded general equilibrium.

Xavier asks and answers a number of questions using this behavioural New-Keynesian framework. I am going to focus here on his proposed solution to the “forward guidance puzzle” since it is most closely related to Michael Woodford’s presentation that I will discuss below.

The forward guidance puzzle is the fact, that in the rational forward-looking New-Keynesian model, believable threats to take policy actions in the future have a bigger impact on present GDP and inflation the further ahead they occur. As Xavier explains in his paper, this puzzle is driven by the unrealistic assumption that people are infinitely powerful computing machines who have perfect knowledge of the infinite future.

The fact that forward guidance is predicted to be so powerful is a puzzle because it is implausible to think that policy actions in the infinite future are more powerful than policy actions a few years ahead. My own view, as I have argued here, is that it is the New-Keynesian framework itself that is broken. The right way to move forward is to ditch the entire New-Keynesian enterprise.

Xavier takes a less radical approach and he chooses instead to try to fix New-Keynesian economics rather than overturn it. Xavier achieves this fix by introducing limits to rationality. In Xavier’s world, where people are boundedly rational in a very particular way, the effects of long-horizon forward guidance are dramatically attenuated. Xavier’s explanation for the forward guidance puzzle is closely related to the research agenda of Michael Woodford. I turn to Michael’s work next.


Michael Woodford was one of our two keynote speakers, along with Jim Bullard, and his keynote address closed the conference.  Michael is one of the founders, and a long-time proponent, of New-Keynesian economics. If you are interested in understanding the New-Keynesian paradigm and its foundations, his path-breaking work Interest and Prices is the book to read.


Michael is a fascinating thinker who has worked on a cornucopia of ideas. We first met when I was an Assistant Professor at the University of Pennsylvania in the 1980s when Michael and I wrote a paper on self-fulfilling prophecies. This was the first paper to introduce indeterminacy as a positive explanation of business cycles and it has the distinction of being the only paper I have ever collaborated on that was published in Spanish before it appeared in English.

Michael went on to become one of the most important early figures in the development of New-Keynesian economics.  He was, and still is, is a key figure. His current research takes on the question of how we should model behaviour when people may not be fully rational. This was the topic of Michael’s talk at the conference.

Michael addresses the question of forward guidance and specifically how central bank announcements will affect the economy when people are forward-looking but not infinitely forward looking. His goal, like Xavier’s, is to fix New Keynesian economics. In his own words, Michael is interested in situations where “people can look some way into the future but not all the way to the end.”  In the attached video clip Michael puts it this way:


economic agents are like expert chess players or go players. They do [engage in] forward planning, …. but they only analyse the consequence of possible moves a finite number of steps ahead.


Michael is concerned that the New-Keynesian model makes a counter-intuitive prediction. If people can plan an infinite number of steps ahead, the model implies that a commitment to very low interest rates by the central bank will be deflationary. Many policy makers would agree with Michael that this is an implausible prediction. They believe, instead, that a low interest rate policy will be inflationary. By restricting the planning horizon of economic agents, Michael shows that forward guidance in the New-Keynesian model implies that low interest rates will be inflationary, a result that accords with central bankers’ intuition.

The second main finding of Michael’s analysis is consistent with Xavier’s research that I discussed above. The NK model, with infinitely forward-looking people, predicts that a commitment to keep interest rates low for a long time will have very big effects on current GDP and inflation and these effects are larger, the further in the future the plan extends. When Michael accounts for an economy populated by clones of Garry Kasparov and Ke Jie instead of omniscient, clairvoyant clones of homo-economicus, he finds that monetary policy works, in economic models, much as central bankers think that it works in the real world.

I am not a fan of the New-Keynesian paradigm for the reasons I discuss here and I have long been an advocate for an alternative approach to macroeconomics that stems, in part,  from the paper that Michael and I wrote together in the 1980s. I have immense respect for Michael and I was personally delighted that he accepted my invitation to present his work at our conference. The work that he and Xavier presented on bounded rationality is important and interesting, but in my view, has more promise if applied to multiple equilibrium models. Rather than try to fix the New-Keynesian paradigm, perhaps it would be more productive to ask how people come to form their beliefs in a world where there are many possible ways to behave rationally. If you are a young researcher looking for a thesis topic, consider applying the insights of Xavier and Michael on bounded rationality to the multiple equilibrium agenda that I discuss in my book Prosperity for All.