Musings on the Fiscal Theory of the Price Level

During my summer hiatus in late August, I managed to dig a bit deeper into the Fiscal Theory of the Price Level (FTPL) and read the manuscript, John Cochrane has kindly made available on his webpage. Cochrane’s book will be published in early 2023 so there is the possibility that there will be changes to the version I have read, but I still wanted to point out some major problems I have with FTPL.

First, as a disclaimer, I am in no position to argue with someone like John Cochrane about the model or the theory in detail. If I would go into an infight with him on the details of FTPL, he would no doubt wipe the floor with me. I am not an academic, I am a practitioner and I judge theories by their ability to explain the world as it is. Thus, many economists – particularly those of a macro inclination – may simply dismiss my musings as coming from someone who doesn’t know what he is talking about. And that is fine.

But when I read the first few chapters of Cochrane’s manuscript where he develops FTPL from a simple one-period model into more and more complex versions of the model, I could not help but wonder if the entire theory is built on sand.

Economic models are famously simplifying the real world just like natural scientists build models of the world that are simplified versions of the real thing. But as a physicist, you are trained not just to simplify the world, but not to oversimplify it. As Einstein has said:

“Everything should be made as simple as possible, but not simpler.”

In my view, FTPL is one of these cases of a theory that has oversimplified the world it tries to describe. For starters, the theory does not include consumption and investments or only to such a rudimentary extent that they have become unrecognisable artifacts. Yes, compared to classic monetarist theories FTPL is more advanced because it creates a link between monetary and fiscal policies. But inflation is not a monetary phenomenon alone. It isn’t even a phenomenon based on the interaction of monetary and fiscal policies. Inflation is very much a real-world phenomenon. Inflation is determined to a large extent by what businesses and households do with the money they earn. In a sense, FTPL seems to largely ignore both consumption and investment activity and focus on paying taxes as the sole purpose of money.

But to give you a pathological example. If everybody would stop consuming and investing, there would be no inflation. No matter the monetary and fiscal policy actions undertaken by the government, if there is zero economic activity, there is no inflation. Of course, in the real world there always is some economic activity, but this example shows that any theory of inflation must include a realistic description of consumption and investment and FTPL simply doesn’t do that, in my view. And because it doesn’t do that it is unlikely to come up with correct policy prescriptions and inflation forecasts. It might do for some episodes (particularly in hindsight) but if FTPL forecasts are correct, they are correct by accident, not by design, I fear.

Second, in chapter 22 of his manuscript, Cochrane tries to dismiss some commonly raised objections to FTPL. For one he zooms in on the use of the net present value of future deficits in FTPL, something that I have struggled with in my initial critique of FTPL as well. Essentially, FTPL assumes that people discount future government (primary) deficits and surpluses to the present day and react to changes in the net present value of these deficits and surpluses. Well, I simply cannot believe that real people act like that. If I would ask my mother what she thinks about inflation and government deficits she would struggle to come up with a coherent answer. But here is what Cochrane has to say about this argument (p. 511):

“The present value equation is part of new-Keynesian and monetarist theories. If somehow the present value equation fails, it rejects those theories equally”.

Exactly. Monetarist theories have failed and so did new-Keynesian theories of inflation. Why does anyone think that using the same fundamental approach to modelling inflation is going to create a theory that suddenly works? Looking at it this way, FTPL looks like yet another attempt to rearrange the deckchairs on the Titanic.

Finally, Cochrane addresses the obvious elephant in the room, which is the case of Japan and I think, he inadvertently points to not only the fundamental problem of FTPL but also a better way forward (p. 511, edits mine to increase readability):

“Discount rates vary. Discount rates are lower in recessions and higher in booms, driving a time-series correlation of inflation with business cycles. The steady downward trend of real interest rates from 1990 to 2020 suggestively correlates with high and rising values of debt in advanced economies, together with low and declining inflation. Just why real interest rates are so low is a good economic question. But it is an economic question for all theories […]. Japan has low real rates. Simplistic r < g calculations say that its present value puzzle is the absence of much greater deflation! […] The Japanese government has a lot of assets. Japan accumulated foreign assets during a long period of trade surpluses. Japan's debt is largely long-term, held by Japanese people and domestic financial institutions. Japan has an inheritance tax. And, perhaps, just wait.”

FTPL cannot explain the case of Japan, and neither can monetarist and new-Keynesian theories of inflation. That is like developing a theory of gravity that works on Earth but not in distant galaxies. You wouldn’t put much faith into such a theory.

Bracket remark: Coincidentally, physicists discovered a long time ago that Einstein’s General Theory of Relativity does explain gravity everywhere in the universe but that distant galaxies do not have enough visible matter to be held together by gravity alone. That is what the entire fuzz about dark matter and dark energy are about. But the point is that physicists are developing entirely new theories to explain these deviations. They don’t just say our theory of gravity is good enough and the fact that other galaxies out there exist that don’t fly apart is not our problem.

In any case, Cochrane in the above quote points to two things that somehow are constantly ignored by macroeconomists: That we have seen a centuries-long decline in real rates (see here for my discussion of this phenomenon) and that discount rates vary over time.

Saying that discount rates vary is like saying that humans exist. A theory of inflation must include the human factor because it is the key driver of inflation both in the short run and in the long run. That FTPL just like new-Keynesian and monetarist theories does not provide any explanation of why and how discount rates vary does not make FTPL a valid alternative to the other theories. It makes all three approaches equally useless.