Believe those who are seeking the truth. Doubt those who find it. Andre Gide

Friday, February 9, 2024

Does high-interest policy constitute fiscal stimulus?

I haven't had much time to blog lately, but I thought I'd weigh in on an interesting discussion I see brewing in Twitterland. The macroeconomic backdrop of the story is how the U.S. economy grew so rapidly in 2023 in the face of a dramatic increase in the Fed's policy rate. Over the period March 19, 2022 - July 27, 2023, the IORB rose from 0.40% to 5.40% and has remained there ever since. In 2023, RGDP grew at 2.5%. PCE inflation in 2023 came in at 3.7% (a decline from the previous year's 6.5%). The unemployment rate remains low (around 3.5%). What's going on here? What happened to the recession so many were predicting?

One idea floating around out there is that high interest rate policy constitutes of a form of fiscal stimulus. Here's Stephanie Kelton expressing the idea:

And here's my friend Sam Levey suggesting the same thing:
For ears attuned to conventional wisdom, this idea sounds bizarre and counterintuitive. But I think there's a way to reconcile these different views. The first step toward reconciliation is to understand the difference between increasing the interest rate and keeping it elevated. That is, we need to make a distinction between change and level.

I think *changes* in the policy rate seem to work the way conventional wisdom dictates (i.e., lowering aggregate demand through a variety of channels). One important channel works through the wealth effect (e.g.):

Once the policy rate remains stable and the transition dynamics work their way through ("long and variable lags"), the higher policy no longer appears contractionary. In fact, high interest rate policy may very well be expansionary, as Stephanie and Sam suggest. How might this work?

In the class of economic models I work with (e.g., see here), monetary (interest rate) policy and fiscal (tax & spend) policy are inextricably linked through a consolidated government budget constraint. A *change* in the policy rate has all the textbook effects of monetary policy--but only in the "short-run." So, for example, while an *increase* in the interest rate puts downward pressure on the price-level, the disinflationary force is transitory (the P-level remains permanently lower if the policy rate remains permanently higher, but the rate of change of the P-level in the long-run remains unchanged).

At least, this is what is predicted to happen if the fiscal policy framework is "Ricardian." A Ricardian fiscal policy is one in which the path of the primary deficit/surplus adjusts over time to anchor a given debt-to-GDP ratio. A Ricardian fiscal regime is often just assumed in economic models. This assumption seems hard to reconcile with the fact that Congress does not appear to implement offsets to Fed policy. At least, it does not appear to do so immediately. It is, however, possible that the offsets (higher taxes, lower spending) are postponed to the future (perhaps after our representatives become alarmed by posts like Marc Goldwein above).

But there is another possibility. It could be that the fiscal regime is "Non-Ricardian." A Non-Ricardian fiscal policy does not anchor fiscal policy in the way a Ricardian regime does--it does not offset higher interest expense (and higher interest income for bond holders) with higher future taxes and/or lower future spending. To finance the added interest expense, it just lets the Treasury issue nominal Treasury securities at a faster pace. If the Fed keeps its policy rate steady (at its higher level), then this additional flow of private sector wealth is likely to manifest itself as stimulus (higher inflation, if the economy is at full employment). Is this where we're at today?

Of course, the Fed is likely to react to the situation described above by *increasing* its policy rate again. But if fiscal policy is Non-Ricardian, the disinflationary pressure induced by the rate change will eventually dissipate. Indeed, it will result in an even higher rate of inflation in the long-run. This is related to the "Unpleasant Monetarist Arithmetic" argument put forth by Sargent and Wallace over 40 years ago; see here: Is it Time for Some Unpleasant Monetarist Arithmetic? And in case you believe this scenario is only hypothetical, consider this paper on the Brazilian hyperinflation: Tight Money Paradox on the Loose: A Fiscalist Hyperinflation. (Note: this is not to suggest that the U.S. is Brazil.)

As always, please feel free to share your thoughts below.



  1. Brilliant as always. And you didnt say "Fisher" once!

  2. Counter-intuitive but thought provoking. Currently, it is interesting to observe yield curves of income instruments remain inverted as they have been for quite some time now.

    Could markets be anticipating yet another policy-driven negative global supply shock and some impact on measured inflation? Moreover, most macro models by necessity blithely assume stable property rights regimes and ignore long-term trends such as hegemonic decline which is generally excruciatingly slow with only the odd agonizing moment.

  3. Hey David, nice post, and thanks for the shout-out :)

    Is Ricardian vs. Non-Ricardian really the issue though? My position is that, given the way that fiscal policy works in reality, it is basically impossible for fiscal policy to be Non-Ricardian. What I mean by that is, to the extent that the government issues liabilities only (or at least primarily) by buying goods and services, it is simply unable to issue liabilities that the public does not expect to be redeemed, because private agents would just choose not to sell goods to the government. As a result, a) the intertemporal government budget constraint cannot *not* hold, b) therefore people today always *expect* it to hold, and c) therefore if the debt increases, that is always met exactly 1-for-1 with expected future surpluses, and actual future surpluses (where at worst, private actors choose to sell less to the government in the future, driving the government's budget into surplus).

    In my model of MMT's price level determination story, I explicitly model the private sector's decision to sell goods to the government, and so this comes up there: (Though that model has a fixed exogenous price level, so not maybe so relevant for your question here.)

    Anyway, all that said, I take the point that it may come down to differing assumptions about how the fiscal authorities react.

    Finally, can you say a little more about the mechanism involved here? The way I see it, if a one-time permanent increase in the interest rate reduces Investment spending (and ignoring the fiscal issue), it would do so on a permanent basis. Then if inflation was driven by the gap between capacity and demand, that would be a permanent reduction in the rate of inflation. Why, in your model, does the inflation rate recover?

  4. What role does QE play? Should the Fed build its balance sheet?

  5. What if we add distribution to the model? Suppose there are two kinds of households. Let’s call them spenders and savers, in honor of Greg Mankiw. Suppose “savers” roll any marginal income they receive into bank deposits (earning FFR) or Treasuries. Spenders but newly produced goods and services.

    Then the degree to which higher rate levels will constitute a kind of baseline stimulus will depend upon the degree to which Treasuries and bank deposits are already held by savers. In extremis, if all Treasuries and bank deposits are held by savers, there will only be the conventional effect, as spenders are deterred by high borrowing costs.

    In general, monetary policy is more predictable and straightforward (conventional) when wealth inequality is high (as the distinction between spenders and savers is likely savers are rich). High interest rates then aggravate wealth inequality, and interest compounds inertly within the bank accounts of the rich (and converts marginal spenders to savers).

    So interest rate policy might simultaneously “work” in the conventionally expected way, and aggravate wealth and more broadly social stratification.

    Maybe it’d be better if it didn’t work.