The BIS Quarterly Review recently published an article on r* (the preferred term for the “natural rate of interest”) by Benigno, Gianluca, Boris Hofmann, Galo Nuño Barrau and Damiano Sandri. “What vadis, r*?” The natural interest rate after the pandemic.
This article is an example of why I have largely abandoned the DSGE literature. From my point of view, the content can be summarized as follows:

The authors describe r*, which constitutes a necessary empirical complement to the dynamic stochastic general equilibrium (DSGE) literature. The DSGE literature assumes that the policy rate (and its expected change) is a key driver of macroeconomic dynamics. The value of r* is necessary both to assess the accuracy of models and to provide policy guidance.

They then examine various estimates of r*and note that they have all evolved dramatically due to the pandemic and the monetary policy response to it.

They discuss various theories as to why r* maybe increased, or maybe decreased.

They then note that the uncertainty regarding r* must be taken into account when reviewing the policy.
The article constitutes a readable summary of the subject as one might hope to find, and contains 40 references (some recent, but many older which deal with r* estimation techniques). It contains nice graphs showing the evolution of r* estimates. For those interested in the subject, this is a good place to start.
My concerns
Beyond noting that the article is a good summary, I won’t dwell on it too much. The reason is simple: despite the extensive study of the articles which I did not bother to read, it did not contain any information that I would consider new to me. Unfortunately for readers familiar with my writing, this means that the rest of this article is also entirely predictable.
The problem is simple: the entire DSGE literature is now based on the assumption that there is no other way to think about macroeconomics. The fundamental assumptions of the DSGE macro must be almost correct, even if it is accepted that existing models need new mathematical bells and whistles to achieve a state of modeling nirvana. That is, it is possible to object to parts of standard DSGE models, but the “correct” model must have similar properties (at least with respect to interest rates). Thus, macroeconomics must react to the real policy rate in such a way that it is possible to estimate r*, r* will evolve in an intelligible manner, and that political decisionmakers with a precise estimate of r* will evolve as they expect.
My counterargument is that the most parsimonious view of macroeconomics is that the economy does not respond to interest rates in the way conventionally assumed. As long as macroeconomic variables do not fluctuate wildly – which happens most of the time outside of crises – any econometric method of estimation r* will converge to something resembling a moving average of the real policy rate. And since the policy rate is set by convention – and not by natural law – any estimated value of r* simply tells us the historical trajectory of policy makers’ conventions. This is useful for those who spend their days guessing where the policy rate might go, but it doesn’t necessarily mean they are able to guess where the economy might go.
You don’t need a PhD in economics to understand why my proposal was not adopted by the many leading luminaries cited in the article, but only by a cynical understanding of institutional psychology. None of them will be willing to commit to a position that the intellectual edifice they have spent almost their entire adult life developing is completely worthless and must be abandoned. And since adopting macro DSGE is a necessary step to advance at elite institutions, don’t hold your breath waiting for a socalled “paradigm shift.”
If I had had time to dissect the document, I would have been happy to do so. I don’t have time right now, so I just wanted to point out the existence of the article, and I made my stock response.
Email Subscription: Go to https://bondactivities.substack.com/ (c) Brian Romanchuk 2024