To make a case for reform, you need to argue that performance has been bad, based on observable metrics. In his first press conference in June, Warsh says:

We recognize that inflation has been running well ahead of the Fed’s long-stated inflation goal of 2 percent that’s been going on for more than five years.

And, in the second press conference in July:

…we talked a lot about the implications of the past five years of high inflation on the current policy conjuncture. To echo an old phrase, has the past really passed?

What’s he talking about? Every January, the FOMC revises its Statement on Longer-Run Goals and Monetary Policy Strategy The Fed’s mandate is defined in the Federal Reserve Act, and in other pieces of legislation as, roughly, maintaining price stability and maximum employment - a dual mandate. That’s rather vague, and the statement of longer-run goals fleshes it out. Part of what it spells out, consistent with most central banks in the world, and consistent with the first part of the mandate, is an inflation target. This is specified as a 2% target for the 12-month increase in the personal consumption expenditures (PCE) deflator. So, in principle that’s easy to understand and to use as a measure of the Fed’s performance, at least for one part of the dual mandate.

Here’s data on the 12-month PCE inflation rate, and the secured overnight financing rate. I’m looking at the latter as I think that’s the best measure of the Fed’s interest rate policy, rather than the fed funds rate (more on that in the future).

Figure 1: PCE Inflation and Overnight Interest Rate

So what Walsh says is true in that, indeed, inflation has been above 2%, consistently, for the last five years. But, from late 2023 to early 2026, inflation was between 2% and 3%, which is actually pretty good as inflation targeting goes - the science is imperfect, and even if it were perfect there are forces driving inflation other than monetary policy. And those non-monetary factors can be hard to predict. You can’t nail inflation at 2% even if you want to and know what you’re doing. For example, the Bank of Canada states their inflation target as a range of 1% to 3%.

Speaking of forces beyond the central bank’s control, what could we say about the period from early 2021 to late 2023, which is part of the 5-year high-inflation period Warsh points to? Leading up to that, you’ll remember the period from 2010 to 2020 as one of persistent undershooting of the Fed’s inflation target. Following that, the COVID period of high inflation took virtually everyone by surprise. As far as I can tell, that period is still not well-understood, but I think it’s uncontested that this was driven by COVID-19. And essentially all countries of the world had roughly the same experience with high inflation over this period.

Also note in Figure 1 that, by late 2024, PCE inflation appeared to be on the way to 2%, though with some bumps in the road. Of course, you know what happened at the beginning of 2025 - we had a new White House occupant. Two hits to the PCE - tariffs and the Iran War - are level effects (they shouldn’t cause sustained increases in PCE inflation) and are perhaps temporary - but in the adjustment period they’ll show up as increases in the inflation rate.

So, I think it’s easy to argue that the inflation-targeting performance of the Fed over the last five years has been good, considering. That is, it’s not hard to blame high inflation on factors outside of the Fed’s control - COVID-19 and DJT. You could quibble with the Fed’s response during 2021-2023, but it’s quite possible that this burst of inflation would have gone away on its own.

But is Warsh blaming high inflation on the Fed, or is he just “stating facts?” I’m suspicious. In the context of other things he’s saying - criticism of how the Fed communicates, the task forces, it seems clear he wants to claim the Fed is not performing well, but he’s smart enough to avoid being caught dissing Jay Powell. Interestingly, Warsh likes to stress the five years of high inflation which, coincidentally, lines up with the beginning of Biden’s term. There are other ways he could say this - ‘inflation is currently above target,’ for example.

Implicit criticism of one’s predecessor and the institution, along with public calls for task forces and reforms are, as far as I know, unprecedented. There have been FOMC chairs who changed Fed policy in radical ways. For example, Paul Volcker implemented a radical approach to inflation control, but did so without task forces or, to my knowledge, criticism of Burns and Miller. Bernanke was a big advocate of inflation targeting, but he didn’t blame Greenspan for not introducing it (or for the financial crisis, for that matter).

But why the task forces? As I noted in my last post, these task forces are supposed to address five areas of concern:

1. Fed communications.

2. Fed balance sheet.

3. Data.

4. Productivity and jobs (AI?)

5. Inflation frameworks.

In principle this is supposed to be about outside experts giving advice to the Fed about what to do. But it’s window dressing in the sense that there are plenty of experts employed as economists in the Fed system, and the interaction between the Fed and people on the outside has been good.

The fact that the Fed system is decentralized is one of its strengths. There are 12 semi-independent regional Feds, each of which has its own research department, and ideas compete across the System. Those research departments have come to have very high standards, with economists assessed in those places in much the same way you would do it in academia. The Board of Governors in D.C. also has a very large group of well-trained economists. Coupled with that, the System takes great pains to encourage interaction with outside people. In my career, I’ve spent a lot of time in Federal Reserve Banks, and at the Board, both as an employee and an academic visitor. Question is, why would the Fed need an explicit outside task force to do something, other than to make it look like something is being done, perhaps as cover for stuff that will make our stomachs turn?

Conclusion: As far as I can tell, the Fed is a healthy central banking institution, and its performance, given the extreme nature of some of the shocks it has faced, has been good. I would think the primary goal for an incoming FOMC chair would be getting up to speed with a group of highly-qualified professionals, not carping about the past. There’s something going on here that smells bad, and I hope things turn out well.

I’ll base most of this piece on Warsh’s first two press conferences, with transcripts (and video, if that’s what you like) posted here and here. I’ll structure this in 3 parts.

Forward Guidance

Basically, Warsh doesn’t like it. But what is forward guidance about, anyway, and why would anyone think this is something a central bank should be engaged in? Economists became more interested in forward guidance post-financial crisis, as another item, along with balance sheet policy and negative nominal interest rate targets, in the unconventional monetary policy toolbox. Of course, all these things are now so commonplace that we shouldn’t be calling them unconventional.

Typically, central banks have implemented two kinds of forward guidance, calendar and state-based. With calendar based guidance, the central bank promises to take an action at a particular date in the future. And, state-based guidance specifies that a particular action will take place under a particular set of contingencies in the future. I’m hard pressed to think of an example of calendar based forward guidance, but the Fed does a lot of state based forward guidance, sometimes vaguely, in the FOMC statement, and in the Summary of Economic Projections (SEP).

Just for fun, I compared the March 2010 FOMC statement, with the last one for July 2026. The first runs to 442 words, while the latter is 166. In March 2010, the FOMC wanted to tell us a lot about current economic conditions, why things were going the way they were, and what to expect in the future. The July statement, as Warsh says, sticks to the facts - sort of. One place where it doesn’t is in this paragraph:

Inflation remains elevated relative to the Committee's 2 percent goal, in part reflecting supply shocks that have driven price increases in certain sectors, including energy. The Committee will deliver price stability.

That last sentence is important. That’s a promise about the future, which is forward guidance of course. But it’s not a promise to deliver a 2% inflation rate, but a promise to deliver something undefined - price stability. That’s what’s in the Fed’s legislative mandate (which is appropriately vague), so why did anyone think it was important to put it in the statement?

There’s plenty to take issue with on the forward guidance front. Sometimes forward guidance is non-specific gibberish. FOMC members can agonize over seemingly trivial word choices, intended perhaps to bind their colleagues rather than to inform the public. On some occasions, FOMC members are over-confident in their ability to predict the future, and find themselves in situations where they have to abandon apparently ironclad commitments. The SEP, which seems to be intended as a forecast (including a forecast of future interest rate policy - the “dot plot”) is often interpreted, incorrectly, as a commitment.

So, there are good reasons why the Fed should avoid making rash promises in the form of forward guidance. But what does Warsh have in mind? Apparently, something I would not have guessed he’d be concerned with. In the July 29 press conference, Warsh takes note of what he thinks is one of “two economic developments worth highlighting.” He claims that, currently, “nominal and real yields are materially higher across the Treasury curve.” So, what’s he talking about? If this is true for the entire yield curve, it must be true for 5-year and 10-year Treasuries, nominal and TIPS.

So, I see some increase in the past weeks, but it’s not notable, in the context of what we’ve seen over the past 20 years. Just for good measure, we could look at 5-year and 10-year breakeven inflation rates (nominal yield minus TIPS yield), which is sometimes used as a measure of average anticipated inflation over the relevant horizon.

Nothing much going on there either, I think.

But what is Warsh seeing in the yield curve data? This is connected to forward guidance. Warsh’s notion is that, since he thinks he’s giving financial markets no information, what we’re seeing in asset prices reflects the fact that “Market participants are learning to play the ball, not the referee…”

This is weird, if you ask me. First, no matter what Warsh says, market participants know that Fed actions can matter in a big way for asset prices, and that there are potentially big gains or big losses at stake, depending on portfolio allocations. Warsh can’t stop these market participants from making inferences about the implications of economic news for Fed behavior, no matter what nonsense he throws at them.

Second, Warsh’s approach appears designed to increase uncertainty about Fed behavior. For a long time (50 years?) it has been widely accepted that there are benefits to predictable policy. Predictability need not (and maybe should not) come from explicit promises about what the Fed will do in the future, but from careful explanation of why an action is being taken. Monetary policy will be predictable, if two conditions are met: (i) Fed officials understand what they are doing; (ii) Fed officials can successfully communicate that understanding to the public. Unfortunately, Warsh seems not to understand what he’s doing, and he ain’t a good communicator.

Pressed - by the press - about what exactly the yield curve is communicating to him, he can’t say. He seems to want to criticize the previous FOMC regime for the high inflation we’ve seen in the recent past, but he can’t tell us how he plans to bring it down. He claims that what he’s seeing in the yield curve implies that the market is somehow tightening financial conditions on its own, which makes no sense to me.

Task Forces

In his June 17 press conference, Warsh announced that he was setting up five task forces:

1. Fed communications.

2. Fed balance sheet.

3. Data.

4. Productivity and jobs (AI?)

5. Inflation frameworks.

These are basically committees (“task force” sounds more active, I guess) that are supposed to examine these separate topics. Basically, the idea is to revisit essentially all of the key elements of policymaking at the Fed. The task forces will be composed of people from outside the Fed. I have seen some names mentioned. Maybe better choices could have been made, but in general these are people who know something about monetary policy and how it should be conducted.

But this setup - the task forces - is something I’ve never seen before. A new leader at the Board of Governors, or at one of the regional Feds, sometimes (though not always) will hire new advisors, or will find people from within the organization who they think should be closer to the policy process, given what they want to do. But there are two problems here: (i) Most of this stuff has been analyzed extensively, and there are no good reasons to change things out of the gate - Warsh would be better off waiting until he figures the job out. (ii) Why can’t the Fed’s own staff work on these things - why do you need Tom Sargent or Mervyn King to tell you what to do?

Maybe the task forces are just about communication? Window dressing, basically. Perhaps Warsh wants to send a message to someone that he’s serious about reforming the institution. That someone certainly can’t be the executive branch of the U.S. government. We know those guys don’t care about expertise. But otherwise, where’s the constituency for Fed reform?

Does This Guy Know Anything Useful?

You’ll have to pardon me now for being elitist - that just the way I am. Can’t help it. Warsh has a bachelors degree in public policy from Stanford and a law degree from Harvard. I’ve heard of those schools, I think, but in my experience central banks run better when someone with a PhD in Economics (an MA sometimes works) runs the place. Sometimes a person with an Economics PhD (in this case a D Phil - same thing) even runs the country, after running two different central banks - I can’t say enough good things about that guy.

A really smart person can, of course, learn on the job. After all, Warsh has had five years, 2006-2011, on the Board of Governors already, in trying times, which should have been educational. But his two press conferences don’t demonstrate that he picked up much. Sometimes he’ll toss off economic jargon to, apparently, demonstrate that he’s fluent, such as:

I think there was a misimpression by some in financial markets, by some households and businesses, that central bankers like me, we set a 2 percent inflation target, but maybe we were more tolerable of a somewhat higher inflation target. In economics we’d call that the “revealed preference”.

First, I think implicit in that last sentence is the claim that he’s an economist. Second, I think he means “tolerant,” not tolerable. Third, that’s not revealed preference. Revealed preference comes up in micro theory. The basic idea is that you cannot observe what someone’s preferences are with respect to goods and services. But we can, in principle, see what they buy. If it was feasible to buy a set of stuff A, but I bought a set of stuff B, we say that, by revealed preference, I prefer B to A. But if actual inflation is 3%, and not at the 2% target, you can’t say that the FOMC has revealed a preference for 3%. The Fed, of course, does not choose the inflation rate. They choose some administered interest rates, a target for an overnight interest rate, and some balance sheet actions. So, inflation could be 3% even though the Fed wants 2%, because the Fed doesn’t know what it’s doing, or it’s not feasible to achieve 2% on that particular date.

Here’s another one. This is in part Warsh’s answer to a reporter’s question about what inflation measure Warsh relies on.

But I’ll say this, some version of the Lucas critique, some version of Goodhart’s Law in economics, should remind us that when we talk about measures of inflation or something else, and we describe those measures as being consistent with our objectives, we might make them such that they’re not very good measures or very good objectives. Broadly, if you said to me standing in front of you, I abide fully by the strategy document, we’re going to deliver 2 percent inflation, and not a whisper more, but to achieve that, I’m looking at a broader set of inflation data than PCE. So, without sort of fully revealing my cards, I’m trying to understand like my colleagues, what’s the underlying generalized change in prices that are happening in the economy. It is not a perfect science. I might have said 42 days ago, I’ve got a task force for that. But we have a data project that’s trying to look and see whether we can’t separate the noise from the signal. And so if you would hear a message from me, yes, I care about what the PCE prints are, I care about what the contributions are from CPI and everything else, but my -- my lens is broader than that. Even though the remit is quite narrow.

So, there are two economic ideas in here, the Lucas Critique, and Goodhart’s Law. The Lucas critique is about how economists use macroeconomic models to draw policy conclusions. The idea is that the policy conclusions can be wrong because a change in economic policy - a change in the policy rule - can cause significant changes in the parameters of the model. Typically macroeconometric models are not “structurally invariant” to changes in policy, calling into question their usefulness in policy evaluation. Goodhart’s Law is related - though it’s more central banking lore than an issue discussed by academics. Charles Goodhart - a one-time Governor of the Bank of England, argued that an attempt by the central bank to target something would make it a bad target. The example he was thinking about was money growth targeting.

So, first, what Warsh is talking about is not the Lucas critique. The question he’s answering simply regards how you measure inflation, not how we decide on a policy rule. Second, Goodhart’s law is goofy in this context, if not goofy in general (warning: if someone is calling an economic idea a “law,” it’s probably wrong). You’re telling me your choice of the PCE deflator to measure inflation means that targeting PCE inflation causes PCE inflation to be a poor measure of inflation. This makes no sense. And most of the quote above is gibberish - I’m sure this helped no one in the room understand anything.

Conclusion

I’m not impressed. Warsh seems to be trying to put something over on us - using words that might work in some boardroom, but aren’t fooling the people who pay attention to monetary policy. In particular, the media people at the last press conference didn’t seem to be buying it.

Is this a disaster? Maybe. Maybe not. The optimistic view would be that there’s enough institutional inertia to protect us from this nonsense.

Let’s start with the state of affairs in the U.S. Not everyone seems to know this, but the FOMC claims that it is targeting the PCE deflator. By that measure, inflation is essentially at target, as these things go:

The last observation in the chart is 2.5%. People looking for evidence to support their personal narrative can of course find other inflation measures to make their case, but typically there’s little in the way of theory or empirical evidence to support looking at anything other than the measure the Fed chose.

On the real side, the labour market is less tight than it was, but the numbers are still showing strength:

And real GDP is growing at a good clip:

As well, to the extent we can measure anticipated inflation, that’s under control, as evidenced by 5 and 10-year breakeven rates, which are about where they were before the financial crisis:

So, the state of the U.S. economy, by these measures, appears to be as close to perfect as the FOMC might want, given its dual mandate, including its 2% PCE inflation target. Then, given that the median FOMC member still thinks the long-run nominal interest rate should be about 2.5%, if I had been asleep for 20 years, just woke up, looked only at these charts, and knew the 2.5% long-run estimate, I would wonder why the overnight rate was not 2.5% instead of 5.3%.

[And note - here’s another beef - that I’ve plotted the secured overnight financing rate, as that is what the Fed is currently targeting. Basically, the Fed’s setting for the interest rate on its reverse repo facility currently determines the overnight repo rate. The Fed claims to be targeting the fed funds rate “in a range,” but that’s bullshit.]

So, what could the FOMC people be thinking? They’re all eager to be Paul Volcker, or what? Maybe they’re worried about CPI inflation, which was up to 3.5% year-over-year in the last report. But, again, CPI inflation is not what the FOMC chose to target. My best guess, though, is that the FOMC isn’t so worried about the actual inflation numbers as the state of the real economy. That is, the idea that the inflation rate is controlled by controlling the unemployment rate persists at the Fed, as it does at most central banks in the world, despite plentiful evidence to the contrary. Fed officials were fond of telling us, pre-Covid, how the Phillips curve was “flat” (perhaps “nonexistent” would have been a better word), but that didn’t stop some people (notably Larry Summers) from claiming that to bring inflation down from where it was in mid-2022, we needed a long period of high unemployment. Well, apparently not. Inflation came down with essentially no upward movement in the unemployment rate. The Phillips curve narrative persists, as it somehow allows monetary policy committees to achieve consensus, and it’s easy to explain to the public - however wrong it may be. Unfortunately, if central bank officials speak Phillips curve language for long enough, they start to believe it, which produces bad monetary policy.

The job of the FOMC going forward is to invent a narrative that allows them to justify bringing the overnight nominal interest rate down to something commensurate with 2% inflation in the long run. That may be an overnight rate of 2.5%, but it could be higher. I haven’t been impressed with existing models of the long run real interest rate, but U.S. TIPS yields may indicate that the long run real rate has risen:

This might tell us that the long-run real rate is closer to 2% than to 0.5%, the FOMC consensus. If so that’s good news for inflation targeting. That is, if the policy rate stayed at 5.3% forever, and the long-run real interest rate is 2%, then we would expect inflation to come in at 3.3%, which is better than 4.8% under the FOMC consensus.

The bad news is that the FOMC, along with other central bankers in the world, may be setting itself up for persistent overshooting, just as central banks tended to undershoot their inflation targets from 2010 to 2020. That’s all part of the same phenomenon, which is not recognizing long-run Fisher effects - and this may set in sooner than people seem to think. Basically, consistent with all mainstream macroeconomic theory, a persistent increase in the central bank’s nominal interest rate target produces a long-run increase in inflation, not a decrease. What allows a central bank to hit its inflation target is a widely-held belief that the central bank will always revert to a nominal interest rate target consistent with its inflation target and Irving Fisher. If that widely-held belief falters, the game’s up.

As is typical, the macroeconomic situation in Canada is not so different from the U.S., though the case here is even stronger for interest rate cuts. Canadian CPI inflation is now within the Bank of Canada’s 1% to 3% target range for 12-month inflation, and 3-month inflation rates have fallen below 2%:

Though employment growth is high, the unemployment rate has increased recently:

Real GDP growth is somewhat anemic, in part reflecting poor productivity growth:

So, inflation performance is good, the real economy is weakening, but the Bank of Canada persists with a policy rate of 5%, in spite of its own estimate of a “neutral” nominal interest rate of 2% to 3%:

Again, central bankers seem to think that they have all the time in the world to contemplate whether or not to reduce their nominal interest rate targets. But, as usual, doing nothing amounts to something, and it’s not good. Stop dragging your feet, folks!

First, let’s review where we are, in terms of inflation. In the US, 12-month and 3-month headline CPI inflation rates look like this:

So, things have certainly improved from mid-2022, but the 12-month CPI inflation rate is still well above 2%, at 3.7% in October. And the 3-month CPI inflation rate is certainly not coming down, which is troubling. I’m focusing here on CPI inflation, rather than PCE inflation, which is what the FOMC claims to be explicitly targeting, so I can draw a better comparison to Canada.

At this point you may wonder why I wouldn’t look at core measures of inflation - measures with volatile prices stripped out in various ways. I think there are good reasons to ignore core measures. One typical justification for looking at core measures is that they are claimed to forecast headline measures of inflation. So, first, I’m not sure that’s correct. Here’s headline CPI inflation CPI inflation ex food and energy for the US:

So, eyeballing that chart, I’d say it looks like headline leading core, not the reverse. Second, if what you are interested in is forecasting future headline inflation, I think you could do a better job by using all the available information, rather than stripping some prices out of the headline index and using that. Finally, there’s no theory that gives us any confidence in a specific core measure. I understand the motivation, which is to try to look past volatility in the headline inflation measures, but I don’t think this works well in practice.

Anyway, on to what’s going on in Canada:

So, that’s essentially the same story as in the U.S. I’ve showed the Bank of Canada’s 1-3% target range for headline CPI inflation. As in the U.S. 12-month inflation has come down substantially, but it’s still above the target range, at 3.7%, and the 3-month inflation rates are not coming down.

So, what have the Fed and the Bank of Canada been doing?

I’ve shown overnight rates in the two countries, which are more or less the same. The Bank of Canada began interest rate increases earlier in 2022 than did the Fed, but did not go as high - the overnight interest rate in Canada is about 30 basis points lower than the effective fed funds rate.

We’ve seen so far that U.S. and Canadian experience during this hig-inflation episode is quite similar, both in terms of the problem, and the monetary policy response to the problem. And, the narrative that BoC officials and Fed officials spin is very similar as well. Basically, according to these people, a disinflation is produced by high interest rates. “Demand” is too high relative to “supply,” an increase in the nominal rate of interest increases the real rate of interest, which reduces demand, which reduces the rate of inflation. It’s basically an IS/LM/Phillips curve story.

So, how does that narrative mesh with what has actually occurred? There’s been a disinflation alright, but here’s what’s been going on in labor markets:

There’s been a slight increase recently in the unemployment rate in Canada, but unemployment rates in both the U.S. and Canada are still in the ballpark of where they were pre-pandemic. So if the central bank narrative is correct, why did inflation come down? In the basic narrative that reduction in demand shows up as a decline in economic activity and an increase in the unemployment rate - it’s not something I can’t see. So, if we believe the narrative, then the inflation rate actually came down on its own. Whatever made it go up - and there’s certainly no consensus about that, other than that it’s pandemic-related - those factors worked in reverse to make inflation come down. So, that’s troubling, as it says that it’s possible that central banks could have done nothing and been in the same place - or better possibly.

But what about the standard narrative? Does it make any sense? Let’s go way back - to the Volcker era in the U.S. When Volcker became FOMC chair, in 1979, inflation was high, and there was a consensus that something needed to be done about it, though there were differing views about how to engineer a disinflation (monetary policy vs. wage and price controls, essentially). Volcker’s FOMC seemed convinced that monetarist principles would work - follow a Milton Friedman prescription of reducing money growth, and inflation would surely come down.

So, if we think of monetary policy as managing money growth, the idea is that we ignore the path of interest rates, reduce money growth, and hope for the best. A worry in the old days was that non-neutralities of money - sticky wages and prices, credit market effects, whatever - would kick in, and we could suffer a long period of high unemployment. But that would be a symptom of disinflation. Ultimately, long-run neutrality tells us that, if we wait long enough, then inflation and nominal interest rates will be low, and the real economy will be humming along.

Fortunately, though the 1981-82 recession was relatively severe, inflation fell, nominal interest rates fell, and by the mid-1980s the U.S. economy was humming along. But central bankers’ interest in monetarism didn’t last long, as money growth targeting proved to be a poor approach to ongoing inflation control. So, what evolved is what we have now, in most countries: a central bank with an inflation target, with the primary means for achieving the target being some implicit policy rule for adjusting a nominal interest rate target, given observables.

But, given this modern framework for monetary policy, central banks have a rather strange view of how inflation control works. For example, central bankers want to engineer a disinflation not through some means where we know where we’re going in the long run (low inflation, low nominal interest rates, economy humming along), and anything bad that happens is due to non-neutralities of money. Instead, they seem to think that disinflation works through the non-neutralities of money, as if Volcker reduced inflation because he induced a recession. Basically, we control inflation by controlling the unemployment rate. And we’ve known for a long time that that’s a messed-up approach - or maybe some people forgot.

The danger here is the following. Long-run neutrality - inherent in all the dynamic models we work with, essentially - says that higher nominal interest rates ultimately engender higher inflation. That’s just Irving Fisher. That’s why the long run world with low inflation has low nominal interest rates. So, in a disinflation, engineered by the modern central banker, that central banker eventually has to find a reason to reduce nominal interest rates. What would make our central banks cut interest rates? More unemployment? Inflation at target? Both?

The risk is that, if high nominal interest rates persist, then so does high inflation. How high? Well, this data is encouraging in that respect:

This is the yield on U.S. 10-year Treasuries, inflation-indexed. This is now above 2%, about where it was just prior to the financial crisis. Canadian and U.S. central bankers, the last I checked, seem to think that the long-run short-term real rate of interest is in the range of 0% to 1%. So, for example, if nominal interest rates persist at about 5%, then we should ultimately see inflation in the range of 4% to 5%. But suppose, in line with the above chart, that the long-run real interest rate is about 2%. That means long run inflation of about 3%, given a 5% nominal interest rate - not so bad, though still not at target.

Serious disinflation is something that has never been done before in the context of modern central banking frameworks (inflation targeting and nominal interest rate rules). I wish I were more confident in BoC and Fed people, but they worry me.

So, how does the Fed want to control inflation? We know what the FOMC has done thus far. In the last three meetings, the Fed has been unusually aggressive, increasing the overnight policy rate target by 150 basis points, with the fed funds rate currently at 1.6%, where it was before the pandemic. In its last projections, FOMC members’ dot plot looks like this:

Chart 1

To the far right in Chart 1, we can take the dots to represent committee members’ estimates of the “neutral nominal interest rate,” or the long-run overnight nominal interest rate consistent with 2% inflation, given the absence of macroeconomic shocks for some time, or some such. The median is 2.5%, which comes from a median estimate by the committee members of a long-run real interest rate (or “natural real interest rate”) of 0.5%. Then, using Fisherian logic, add 2% (the inflation target) to get 2.5%. So, the median view on the committee is that, to sustain 2% inflation in the long run, an overnight nominal rate of 2.5% is appropriate.

But you can see that the committee thinks that increasing the overnight interest rate target by another 90 basis points to 2.5% isn’t sufficient to control inflation. The median view as that there should be an overshoot of the neutral nominal rate. The median FOMC person thinks the policy rate should be about 3.4% at the end of this year, and 3.75% at the end of 2023. If you look at the inflation projections included with the dot plots, the FOMC is expecting that inflation will still be above 2% at the end of 2024, with the policy rate above 2.5%.

There are many people that I’ve talked to, and people whose prognostications I’ve read, who think that the FOMC’s projected policy is not aggressive enough. I’ll take Larry Summers’s piece in the Washington Post as representative of this view. Summers says:

…it [the FOMC] is not raising real rates above the neutral level. On the Fed’s forecasts, real rates — interest rates adjusted for inflation — will be negative and well below their neutral level for this year and next. On market forecasts, real rates will not reach zero, let alone their neutral level, within the next decade.

Basically, Summers thinks the real interest rate needs to be sufficiently high to bring inflation under control.

When Summers says “interest rates adjusted for inflation,” I’m interpreting this as the ex-post real interest rate, that is the current nominal interest rate minus the actual rate of inflation. Here’s what that looks like:

Chart 2

In terms of how Summers thinks about the world, that’s very alarming. He thinks low real interest rates cause high spending, which makes unemployment low, which makes inflation high, via a Phillips curve (PC) effect. So, according to this view, real interest rates cause inflation, which is consistent with a standard narrative about what caused high inflation in the 1970s. That is, according to that view, the low real rates you see following the 1974-75 recession in Chart 2 caused the high inflation that Paul Volcker had to deal with during his two terms at the Fed (1979-1987).

Summers’s logic is basically the same as what’s behind the Taylor rule, specified in its original form as:

(1) R = r* + i* + a(i - i*) + by,

Here, R is the target for the overnight nominal interest rate, r* is the natural real interest rate, i* is the target rate of inflation, i is the actual rate of inflation, and y is the output gap, i.e. some measure of slack in the economy. Finally a and b are parameters with a > 1 (the Taylor principle) and b > 0. For our purposes, set b = 0, supposing we’ve got a policymaker with their mind set on inflation control. Then, if r* = 0.5, as the FOMC seems to think, and i* = 2, then the Taylor rule tells us the current target for the overnight nominal interest rate should be somewhere north of 7.3%. Further, note that the Taylor rule implies that the ex-post real interest rate, when b = 0, from (1), should be

(2) R - i = r* + i* + (a - 1)(i - i*)

So, the Taylor rule dictates a nominal policy rate above the neutral rate, r* + i*, so long as inflation is above target, which I think is what Larry Summers wants the Fed to do.

It’s clear from (1) that when the central bank is achieving its goals, then the overnight nominal rate will be equal to the neutral nominal rate, so if the central bank knows r* (an important issue, but suppose they get it right) then the Taylor rule is at least consistent with a long run in which, if the central bank achieves its goals, its policy rate is equal to the neutral rate. But how do we know that the Taylor rule takes us to Nirvana?

Typically, in standard monetary models we like to play with, including those with sticky prices, the Taylor rule generally doesn’t take us to Nirvana. There are Taylor rule perils, as Benhabib and coauthors originally pointed out. Basically, in models where we can work out the global inflation dynamics (not linearized New Keynesian models, for example) a Taylor rule with the Taylor principle produces multiple equilibria, with many of those converging to a zero-lower-bound steady state with zero nominal interest rates and low inflation. That happens because Fisher effects are more important than typically realized, in standard monetary models. That is, an aggressive central banker, seeing inflation below target, lowers the target nominal interest rate, which further lowers the inflation rate, which further lowers the nominal interest rate target, etc., until the nominal interest rate goes to its lower bound. And there it stays, so long as the central banker adheres to the Taylor rule.

But what happens when inflation is above target? In some models, I think inflation above target can’t be sustained, but in a baseline case with money neutrality (say an endowment economy with cash in advance), the Taylor rule will sustain an equilibrium in which inflation increases forever. Basically, higher nominal interest rates effectively accommodate the higher inflation — supported of course by open market operations. So, Summers, and many others like him, think that low real interest rates cause high inflation, but typical dynamic models say that, if you follow policies based on that idea, high nominal interest rates will act to propel high inflation.

For example, going back to 1970s inflation:

Chart 3

In Chart 3, the blue line is the fed funds rate, red is the PCE inflation rate, and green is the ex post real interest rate. So, in the chart, is the low real interest rate causing the inflation rate to go up, or is the nominal interest rate going up, thus accommodating the high inflation? Note in particular that there are times when the increase in the fed funds rate leads the increase in the inflation rate.

The idea that Fisher effects are important for monetary policy, sometimes called Neo-Fisherism, isn’t just part of mainstream monetary theory, there’s empirical work on this too. In particular, work by Martin Uribe and Azevedo, Ritto, and Teles argues that the data support the notion that temporary increases in the nominal interest rate reduce inflation, but persistent increases result in inflation going up. And the long run where the Fisher effect dominates comes much sooner than you might think.

John Cochrane thinks that FOMC members are effectively Neo-Fisherites (though maybe they don’t know it). They’re not like Larry Summers, in that they don’t want to adhere to an aggressive Taylor rule, thinking that high inflation will pretty much go away on its own. Though you’ll notice that the June FOMC dot plots are more aggressive than the March ones were, so maybe the FOMC is moving in a Summers direction. So, I see at least three possible outcomes here:

1. Inflation starts coming down of its own accord sooner than most people expect, the FOMC backs off some, and there’s a smooth path to R = 2.5% and 2% inflation with no recession (ha ha).

2. Inflation is more persistent than I think it will be, the Fed gets more aggressive, the high nominal interest rates accommodate the high inflation, and then we really have a serious inflation problem, but no recession.

3. Inflation is more persistent than I think it will be, the Fed gets more aggressive, we have a recession, the Fed reduces its nominal interest rate target and, by Fisherian logic, inflation comes down. The Fed pats itself on the back, Jay Powell is a big hero, starts smoking cigars like Paul Volcker, and Larry Summers is surprised that we didn’t need 5 years of 5% unemployment to bring inflation down.

Sometimes people think of an increase in oil prices as a negative supply shock. What about that?

Chart 9

So, in dollar terms, oil prices aren’t that much higher than in the 2010-2014 period (and I checked this today, and this price is down to $105). Further, if you adjust for inflation, you get this:

Chart 10

So, in relative terms (deflated by the CPI), oil prices are higher than before the pandemic, but lower than in the 2010-2014 period. So, relative to April 2020, we’re paying four times as much in goods and services for oil, but in April 2020 oil was really cheap. If you can remember as far back as 2014, oil is relatively cheap now.

It’s Fiscal Policy?

This seems to be a more specific supply/demand story. Larry Summers and Olivier Blanchard are often credited with prescience for warning in advance that overly generous federal government transfers could lead to high inflation. And Summers and Blanchard have a common AD/AS/IS/LM/PC view of the world.

A key feature of the pandemic was of course generous federal income support — transfers financed by debt issue. That shows up as a high deficit, made larger by the falloff in tax revenue due to the decline in economic activity early in the pandemic period.

Chart 11

And, the temporarily high deficit has left the U.S. with a higher level of debt, relative to GDP.

Chart 12

So, first, the deficit, as a fraction of GDP is actually a little smaller than just before the pandemic, and the debt/gdp ratio is somewhat higher, but maybe not so much higher, and it’s no longer growing. Thus, any effect of fiscal policy should be temporary, barring for example the effects on debt service of higher interest rates. Second, there is no clear regularity in the relationship between government debt and inflation. Indeed, Japan famously ran government debt up in excess of 200% of GDP, with no apparent inflationary effects. And the runup in debt from 60% to 100% of GDP in the US, from 2008 to 2015, had no discernible effect on inflation.

Inflation is Always and Everywhere a Monetary Phenomenon?

That’s of course a quote from Milton Friedman who, along with his fellow quantity theorists, thought that the quantity of money — defined as the quantity of assets in existence that are being used for transactions purposes — determines the price level. And growth in the money supply, according to Friedman, is what explains inflation. So, if you want to bring inflation down, you reduce growth in the money supply. A quaint idea now but, since people seem to now be having fond memories of Paul Volcker — and since Jay Powell now seems to want to be the modern-day Paul Volcker — I should point out that Volcker was actually consciously implementing a monetarist deflationary plan. He didn’t actually “jack up interest rates,” as the standard narrative goes. The FOMC at the time decided that they were not interested in targeting an overnight interest rate, but in reducing the rate of money growth and letting market interest rates do whatever. Volcker actually finished his 8 years as FOMC chair with lower inflation and lower nominal interest rates than when he started.

We could argue that monetarism worked in bringing down inflation during the Volcker era, but it proved a failure in ongoing inflation control — that’s why most central banks opted for inflation targeting backed up by day-to-day overnight interest rate targeting. Basically, monetarism works only if there is a systematic and predictable relationship between aggregate economic variables and money — a stable money demand function. But money demand functions are as notoriously unstable as the Phillips curve.

However we should look at what is going on with standard monetary quantities during the pandemic, as that might tell us something useful, and because some people (including John Cochrane) have tied monetary phenomena to fiscal policy in trying to explain what’s going on with inflation. So, first we’ll look at some standard monetary aggregates: currency, the monetary base, M1, and M2 (on a log scale):

Chart 13

What stands out in that figure is the huge increase in M1. That’s an increase of about $11 trillion, and most of it happens from April 27 to May 4, 2020. But, not to worry, that’s just the effect of a regulatory change that effectively eliminated the difference between a savings account and a checking account. So, best to ignore M1. We know more or less what’s driving the monetary base - that’s the Fed’s quantitative easing (QE) program. Focusing on 12-month growth rates in currency and M2:

Chart 14

So growth in currency outstanding and in M2 was unusually high during the pandemic, with growth rates peaking in early 2021. Those growth rates are now close to where they were pre-pandemic, though in level terms consumers are holding more liquid assets. M2 rose from about 70% of GDP before the pandemic to about 90% now. If you’re thinking that’s some money multiplier story associated with QE, it certainly doesn’t look like it. The increase in M2 had to be matched by an increase in assets on banks’ balance sheets, and the assets acquired by the banks were mainly Treasuries, reserves, and privately-issued securities, not loans. That is, the government increased its dissaving, issuing a lot more debt during the pandemic. Ignoring the foreign sector, the private sector had to dissave and acquire the debt. Basically, they did a good portion of that through the banking sector, holding the assets in the form of insured deposits backed by government debt, some of which the Fed had turned into reserves by way of QE.

As a check, we saw an almost identical phenomenon in Canada.

Chart 15

The measures of money are different (it’s a different banking system), but you can see a similar pattern, with growth rates peaking around the same time as in the U.S., and growth rates currently approaching pre-pandemic rates.

So, what’s going on? We’re not in a Milton Friedman world in which the central bank can control the quantity of means of payment by controlling outside money, with a fixed relationship between the two. Though there’s a story out there, which John Cochrane was discussing, that the federal government and the Fed coordinated a “helicopter drop” during the pandemic. That is, according to the story, the government issued debt to finance transfers, and the Fed bought the debt with money — thus, effectively money was issued to the private sector, and it’s obvious this causes inflation (according to the story), just like in Milton Friedman’s parable. To me, this doesn’t make sense. For example, note in Chart 10, as I mentioned earlier, that federal government debt goes from about 60% of GDP before the financial crisis, to about 100% of GDP in 2015. And over that time the Fed converted a load of that into reserves. So where was the inflation then?

What seems to be going on is that transactions accounts at banks are demand determined, in that the assets acquired by banks to back the increase in deposits could as easily be held outside the banking sector, by mutual funds for example. Even reserves can migrate outside the banking sector in the form of reverse repurchase agreements — indeed, with assets of about $9 trillion, the Fed currently has about $2 trillion in reverse repos outstanding.

Of course, people accumulated balances as bank deposits and currency during the pandemic for a reason, or reasons. And we can certainly construct a story that in some ways fits the standard narrative about the current sources of high inflation. That is, at the beginning of the pandemic, people accumulated assets in forms they could readily spend, because they were postponing consumption, and were uncertain about when restrictions would relax so they could resume normal activity. But once things did open up, all that deferred spending was unleashed. As well, people needed transactions balances to buy the houses and consumer durables that were rapidly changing hands during the pandemic (new and used). But if the accumulation of money balances is at worst an underlying symptom of the inflationary phenomenon associated with the pandemic, it looks like the inflation cannot sustain itself. That is, M2 relative to GDP has leveled off, and should start to fall.

Chart 16

But, the high level of bank deposits and currency could, at least in part, be due to a flight to safety. The world is now a more scary place, and insured deposits and US currency may be seen as a good place to park your wealth. Indeed, in Chart 14, you can see that M2/GDP rose during the financial crisis, then stayed high, in fact increasing from about 60% to 70% from 2010 to the beginning of 2020. So, who knows, M2/GDP may not decrease further, because bank deposits and currency are being held as safe assets, not for transactions purposes.

Conclusion

The bottom line is that, if we sift through all the inflation narratives out there, some of them make no sense, and some of them make a little sense. Phillips curve narratives won’t help us understand inflation — these narratives are at best descriptive, and won’t help us formulate good monetary policy. In the supply/demand narratives, the idea that “supply” is constraining activity isn’t any more true than it normally is, in spite of the fact that some goods are hard to find. Though gasoline prices loom large in popular discussion, oil is actually relatively cheap now compared to 2014.

What makes most sense to me here is a monetary story about high inflation, related to the pandemic — and of course it has to be a pandemic story, otherwise we would still be in the low-inflation pre-pandemic world we used to complain about. But the monetary story isn’t a Milton Friedman story, according to which the Fed’s actions caused inflation, through QE for example — it’s a story where the accumulation of liquid assets is a symptom. People deferred consumption, they accumulated transactions balances, and then reduced the growth in transactions balances over time. And the economy is beginning to look more like its pre-pandemic self in many ways. The fed funds rate is even where it was at the beginning of the pandemic, at least for the time being. What still sticks out is of course high inflation, and the high job vacancy rate, which may be due to low immigration. I know it’s become commonplace to panic about high inflation now, but I’d say it looks transitory. Next, on to policy issues.

For the impatient, here’s an executive summary of what I’ll discuss in this two-part post, and the policy post to follow: The consensus view on the Federal Open Market Committee (FOMC) is that, if inflation is above (below) target, it is brought down by increasing (decreasing) the target for the overnight nominal interest rate. If we accept that view, then the FOMC made a serious policy error. Inflation has been above the Fed’s 2% inflation target since March 2021, but the FOMC waited a full year to take action in the form of a target interest rate increase. Why did it wait? Because it was overconfident in its ability to understand how the economy works - particularly during an unprecedented pandemic - and overconfident in its ability to forecast future events. That overconfidence led the Committee to commit in 2020 to a policy framework based on a faulty evaluation of post-financial crisis macro history —serving to illustrate why policy commitment can actually be a bad thing.

What caused the high rate of inflation as we exited the throes of the pandemic? It’s hard to have confidence that any of the typical culprits is to blame, but one could make a case that the Fed is now making another error. The FOMC eventually decided that high inflation is not transitory, and that it needed to be assaulted with aggressive interest rate increases. But maybe the transitory view was not so crazy. Maybe high inflation would indeed go away on its own, but for the fact that the Fed’s actions could act to accommodate it.

The Problem

First, to remind you what’s going on, here’s a time series for the PCE deflator.

Chart 1

Note that I’m using the PCE deflator as PCE inflation is what the FOMC claims to be targeting. So, in eyeballing the time series you might notice a break in the trend in early 2021. Or, if we look at the year-over-year measure of the inflation rate, as is typical:

Chart 2

So, in March 2021, the inflation rate exceeded the 2% inflation target. But in June 2021 (to account for data lags relative to actual decisionmaking), the FOMC says in its statement:

With inflation having run persistently below this longer-run goal, the Committee will aim to achieve inflation moderately above 2 percent for some time so that inflation averages 2 percent over time and longer‑term inflation expectations remain well anchored at 2 percent. The Committee expects to maintain an accommodative stance of monetary policy until these outcomes are achieved.

And in the dot plots published after the June 2021 FOMC meeting, the median committee member was predicting an overnight rate target of 0.125% at the end of 2022, with the overnight rate still below 1% by the end of 2023. By October 2021, in Chart 2 the inflation rate had exceeded 5% and was still increasing. What did the FOMC tell us in December 2021?

With inflation having exceeded 2 percent for some time, the Committee expects it will be appropriate to maintain this target range until labor market conditions have reached levels consistent with the Committee's assessments of maximum employment.

Recall that an issue with the Fed’s new policy framework, adopted in 2020, is that it appears to define “maximum employment” as unattainable - thus shortfalls in employment from the “maximum” could always be used to justify an accommodative policy. But what was actually going on in the labor market at the time?

Chart 3

In October 2021, the job openings rate had reached a historical high of 6.8%, and the unemployment rate in November 2021 was 4.2% — not as low as in February 2020, but given the high vacancies, by some metrics the labor market was tighter at the end of 2021, than before the pandemic. Not “maximum employment,” though, according to the FOMC. So the policy rate didn’t move, in spite of the fact that inflation had been above target since March 2021, and well above target by December 2021. However, committee members were anticipating an earlier tightening than before. For example, in the December 2021 dot plots, the median committee member was thinking that the overnight rate would be at 0.75% by the end of 2022.

By March, however, panic had set in, and the FOMC proceeded to increase the overnight rate target in three moves (March, May, June 2022) by 150 basis points, which is unprecedented in size in the upward direction. As well, Fed asset purchases were stopped as of March 2022, and assets are now being permitted to run off the Fed’s balance sheet.

What Caused the High Inflation?

Did the Phillips Curve Do It?

A standard narrative before the pandemic was that the Phillips curve (PC) had become “flat.” You can find plenty of Fed insiders who talked about this, including Lael Brainard. Hardcore adherents of the PC theory of inflation seem to think that central banks control inflation by controlling the unemployment rate, so long as inflation expectations are “anchored.” So, if you’re one of those people, and you think the Phillips curve is flat, you think that the only avenue for controlling inflation is to influence inflation expectations, thus shifting the “inflation-augmented PC,” as it was called in the old days. Here’s a scatter plot of PCE inflation vs. the unemployment rate, both pre-pandemic (January 2010 to February 2020) and pandemic (post-February 2020).

Chart 4

You can see a “flat” PC (though more like a Phillips curve with the wrong sign) in the pre-pandemic data. Then, after the onset of the pandemic, the PC woke up, and we observe a traditional negatively sloped relationship between the inflation rate and the unemployment rate. I think this makes PC adherents happy.

But, if you’re a PC person, and you take the pandemic observations in Chart 4 as tracing out the now-wide-awake PC, what would you conclude about how we should control inflation? We can make a case that inflation expectations are still “anchored,” in that it’s hard to trust survey measures of inflation as measuring anything that people are actually acting on, and bond market measures of anticipated inflation don’t show anything alarming. For example, the breakeven inflation rate we back out from 10-year nominal Treasury yields and 10-year TIPS is at 2.6%. So, adjusting for the fact that TIPS are indexed to the CPI, not the PCE, and that people may be willing to pay a lot for inflation insurance now, maybe anticipated inflation isn’t far off 2%. So, if you buy into PC theory, it takes a recession to bring inflation down - perhaps an unemployment rate of about 6%, from Chart 4. But before the pandemic, Fed people didn’t seem to think that an unemployment rate as low as 3.5% was consistent with “maximum employment,” so with expectations anchored at 2%, why did the natural rate of unemployment rise to about 6%? A hardcore PC person might argue that we’re currently “overheating.” As the argument goes (and you can find thinking like this in Ben Bernanke’s new book), supply constraints shift the Phillips curve too, so presumably the natural rate is now higher. But the supply constraints are going away, so why won’t inflation just go down on its own?

I don’t find Phillips curve reasoning helpful. First, it relies too much on unobservables — in particular the “natural rate of unemployment” and some measure of “slack,” as well as inflation expectations. That’s especially problematic in a policy setting — people can cook their measures of whatever to manipulate policy decisions for other reasons than pure economics. Second, the PC is notoriously unstable. I don’t know of PC types who, in 2020, were forecasting a Phillips curve awakening. The FOMC, which seems to consist primarily of PC types, predicated its new policy framework on a flat Phillips curve. They were thinking hard about the Phillips curve, but that led them to make a policy error, in terms of how they do inflation control. If you can’t tell me why the PC is flat, or tell me when the PC will wake up, or when it will go to sleep, what good is the PC in formulating policy?

It’s All About Demand and Supply?

Econ 101 demand/supply reasoning sometimes works well — just read the boxes in your average Principles of Micro book. However, in helping us understand and deal with inflation, this type of approach leaves something to be desired.

First, demand/supply reasoning sometimes manifests itself in what I’ve seen called the “micro approach to inflation.” This involves going market by market, and pointing out why prices, in dollar terms, are up in each market either due to a shift in the supply curve, a shift in the demand curve, or both. Supply of cars is down because car manufacturers can’t get chips, demand for cars is up because the government gave everyone a pile of cash, for example.

What’s the problem here? Inflation is a macro phenomenon, of course. In a partial equilibrium demand/supply diagram, supply and demand depend on relative prices, but inflation involves the rate of change over time in the price of goods and services in terms of the unit of account/medium of exchange. To think about that, we need general equilibrium, we need dynamics, we need forward-looking behavior, and we need to be thinking not only about goods and services, but assets, and how payments are made.

The second typical manifestation of demand/supply analysis is what shows up in most principles and intermediate macro textbooks. That’s aggregate demand/aggregate supply (AD/AS). So, that addresses some of the problems with the micro approach to inflation, as there are assets in there (money and bonds, implicitly), and it’s sort of general equilibrium. The idea is to make use of a tool an undergraduate already knows something about - supply and demand - in constructing a macro model which we can reduce to an upward-sloping supply curve and a downward-sloping demand curve, in (y,P) space, where y is a quantity (aggregate output of goods and services), and P is the price of goods and services in units of money.

Unfortunately, AD/AS doesn’t get us to where we want to go either. First, it’s explaining the price level, not its rate of change. Second, in general equilibrium, we can’t separate demand from supply. For example, the households in this world are on the supply side of the labor market (which helps to determine aggregate supply), on the demand side of the market for goods and services, and on the demand side of the market for money. So exogenous factors affecting households will shift both the AD and AS curves. As well, as people explained at least 45 years ago, the forward looking behavior of economic agents, which is critical for thinking about monetary phenomena, implies that everything depends on everything else — you can’t identify aggregate demand and aggregate supply curves through exclusion restrictions. That’s what Chris Sims called “incredible identification.”

Typically, in undergrad macro, people will try to explain inflation through some combination of the demand side of AD/AS (that’s IS/LM) and a PC, with exogenous inflation expectations. That of course misses the supply side factors that people are claiming are important now, and also misses the dynamics - we want to explain where the inflation expectations come from. New Keynesian models involve an attempt to bring AD/AS into the modern age. NK is general equilibrium, with optimizing behavior explicitly modeled, and forward-looking economic agents with rational expectations. And, with the aid of some tricks, you can reduce that to an effectively demand-determined system (in a precise sense) with a PC. Though sometimes sold as some kind of microfoundation for IS/LM/PC, NK actually isn’t. I don’t think Keynes, Hicks — and whoever made up AD/AS — would recognize it.

One might be able to explain a rise in inflation in an NK model due to technological factors which affect the output gap, though I’m not sure. But if you can, I think you end up in the same place as with the raw PC reasoning I went through above. That is, the efficient level of output is lower, implying a lower level of output consistent with the same inflation rate, given inflation expectations. But then the high inflation goes away on its own, as the technological constraints relax.

A standard narrative seems to be that pandemic restrictions and lockdowns have disrupted supply chains, so that output is supply-constrained. And demand has been high. Central bankers (and I’ve heard these words come out of the mouths of both Jay Powell and Tiff Macklem, the Governor of the Bank of Canada) then call this a situation with “excess demand” which purportedly propels inflation.

So what if we forget about models, and look at the outcomes for real variables to see if we can extract anything from that. What I’m looking for is some evidence of constraints on production and, maybe, “excess demand.”

First, go back to Chart 3, which shows a persistently high level of job vacancies and a very low unemployment rate. Jay Powell mentioned this in his last press conference, and proposed the idea that it might be possible, through tight monetary policy, to bring down the vacancy rate with no effect on the unemployment rate (“bringing supply into line with demand,” or something like that). Well, good luck with that. Typically in a recession, vacancies and unemployment move in opposite directions, and we might expect that a monetary-policy-induced recession would produce that sort of movement along the Beveridge curve (a well-established regularity). Even worse, one feature of the financial crisis recession in 2008-09 was that the Beveridge curve shifted — we were observing more unemployment given the vacancy rate. So maybe a recession would produce not so much movement down in vacancies, and lots of movement up in unemployment.

Next, here’s the path of real GDP:

Chart 5

So, except for the downward blip in 2022Q1, real GDP is pretty much back to its pre-pandemic trend. In terms of private domestic expenditures:

Chart 6

Chart 7

And finally, imports and exports:

Chart 8

So, I don’t see any indication in the above data that expenditure is being constrained by technology (which includes the state of supply chains). Domestic private spending is roughly on the pre-pandemic trend, and imports are booming. Exports, however, are well below the pre-pandemic trend. But if international trade is supposed to be so arduous, why are imports so high?

I know what you’re going to say. Supply chains! I can’t buy a car. I can’t buy cat food. But technology is always constraining output, which I hope doesn’t surprise you, and people can substitute, as they’re obviously done. If technological constraints were a big deal at the aggregate level, you would see it in the aggregate quantities.

The excess demand people of course might argue that what we’re seeing is “overheating,” which means output in excess of efficient output. How could we get that? In theory, either someone is getting fooled — workers think their real wage is high when it’s not, firms think their relative price is high when it’s not — or some wage or price is sticky. With sticky wages, firms and workers somehow can’t bear to negotiate a higher wage, but workers for some reason want to keep working even given the lower real wage. With sticky prices, some firms can’t bear to increase their prices, even though there is a load of customers coming in the door, and they proceed to hire the workers to serve that load of customers. The interesting part of this “overheating” phenomenon is that I don’t see anyone checking to see if there is widespread fooling going on, or whether there are firms and workers complaining because the costs of increasing prices and wages are so damn high. The workers may be complaining that the firms aren’t paying enough, but that’s a different story. I certainly haven’t noticed the price of anything being stuck recently.

[Continue on to Part II.]

So, at its core, the agreement says:

• The target will continue to be defined in terms of the 12-month rate of change in the total CPI.

• The inflation target will continue to be the 2 percent mid-point of the 1 to 3 percent inflation-control range.

• The agreement will run for another five-year period, ending December 31, 2026.

That’s been the fundamental part of the agreement since 1991. It’s important to recognize what’s useful about this as a central banking goal. Basically,

1. It’s easy to understand what the Bank’s goal is.

2. It’s simple to evaluate the Bank’s performance relative to the goal.

3. The bank is capable of performing well relative to the goal, based on what we’ve observed for the last 30 years.

4. If the Bank performs well relative to the goal, this arguably is the best the Bank can do, relative to any other goals it might specify.

The trouble is in part with this:

“The Government and the Bank also agree that monetary policy should continue to support maximum sustainable employment, recognizing that maximum sustainable employment is not directly measurable and is determined largely by non-monetary factors that can change through time.”

So what’s “maximum sustainable employment?” The language here does not help me understand what the concept is, nor how to measure it. Indeed, we’re told that it’s “not directly measurable.” So there seems to be a goal here, but it’s ill-defined, and the agreement is up-front that we’re not going to be able to evaluate the Bank’s performance in achieving maximum sustainable employment, whatever that is. So why put that language in there?

And there’s more information here:

“Major forces, including demographics, technological change, globalization, and shifts in the nature of work, are having profound effects on the Canadian labour market. These evolving forces have increased uncertainty about the level of maximum sustainable employment (i.e., the level of employment beyond which inflationary pressures arise).”

So, again, if all these evolving forces have increased uncertainty about maximum sustainable employment, why are we even talking about it? This passage does, however, attempt to define what the concept means which is “the level of employment beyond which inflationary pressures arise.” So, that sounds something like the natural rate of unemployment. But how would this help us think about what we saw in the 1970s, or what we’re seeing now? In the 1970s, seemingly we saw “inflationary pressures” arising when employment was low and unemployment was high. And, currently, the Bank is on the record expressing the view that, in spite of the fact that there seems to be a lot of inflationary pressure, and actual inflation, there’s still slack in the economy, which Bank officials seem incapable of defining or measuring. So, apparently, maximum employment is not necessary for inflationary pressure. If you’re now hopelessly confused, it’s not your fault.

Here’s another issue:

“The Bank will also continue to leverage the flexibility of the 1 to 3 percent range to help address the challenges of structurally low interest rates by using a broad set of tools, including sometimes holding its policy interest rate at a low level for longer than usual.”

A lower-for-longer policy is what the Bank signed on to last year. Here’s the relevant language in the December 8 press release:

“We remain committed to holding the policy interest rate at the effective lower bound until economic slack is absorbed so that the 2 percent inflation target is sustainably achieved.”

So, that commitment hasn’t panned out well. The passage reads as if the Bank is controlling “economic slack” - presumably a negative deviation from maximum sustainable employment - so as to control inflation in the long run. But here’s what the path of the consumer price index looks like, from January 2018, relative to a 2% inflation path (I think this is up to date - Statcan is currently down).

So, the Bank’s notion, given the Phillips curve model that seems deeply embedded in Bank officials’ minds and in this agreement was that, coming out of the pandemic, inflation would return to the 2% target from below, as the economy returned to maximum sustainable employment. But of course, that didn’t happen, leaving the Governing Council confused about what to do, other than to stick to their commitment to do nothing until next year. And note that one could even make a case that we’re currently at maximum sustainable employment, given the November labour force data.

So, we can now see the Bank performing badly, relative to its primary stated goal, but doing nothing about it, because they made an unwise commitment last year of lower-for-longer. But in spite of the fact that lower-for-longer now looks questionable, there it is in the new inflation-targeting agreement.

I won’t say this new IT agreement is surprising, as it reflects a standard inclination of central bankers, which is to invest excessively in words that serve only to confuse people. I think I know why the new language is there - which is to placate one or Governing Council members with strong ideas about dual mandates. But the additions don’t serve any useful purpose, unfortunately.

Let’s look at the Bank’s policy statements first so we can, hopefully, understand what they think they’re doing. Then, we’ll look at the data to see if the policy makes any sense. Since early in the pandemic, the Bank has included forward guidance in its post Governing-Council-meeting policy statements. From the September 8, 2021 meeting, there are two key pieces of information, the first relating to the future path of the policy rate, and the second to the rate of asset purchases. On the policy rate,

“We remain committed to holding the policy interest rate at the effective lower bound until economic slack is absorbed so that the 2 percent inflation target is sustainably achieved.”

Here, by “effective lower bound,” the Governing Council means 0.25%. But what do they think “economic slack” is, and how are they measuring it? There’s more information in a speech given by Deputy Governor Larry Schembri on November 16. Schembri says:

“Overall, the labour market has recovered the pandemic-induced job losses, but considerable excess capacity remains. The rates of unemployment and underemployment remain elevated, especially for certain groups in the labour force.”

So according to him, slack, or “excess capacity,” can in part be measured by the unemployment rate, the “underemployment rate” (presumably some measure of discouraged workers — people not in the labor force who want to work). This statement also indicates that the Bank is interested in disaggregating data to see how different groups are being affected. Schembri’s talk goes on to talk about “maximum sustainable employment” as a concept, the output gap, and the Phillips curve as an organizing principle in inflation control. The general theme of the talk is that there is a lot of uncertainty. Maximum sustainable employment is hard to measure, and the Phillips curve may be broken, apparently. Thus there is a lot of uncertainty about what is going on, but Bank officials seem sure that, whatever slack is, it’s still there, and it’s “considerable.”

An interesting part of the Bank’s forward guidance, quoted above, is the link that’s made between disappearing slack and the Bank’s inflation target. Schembri elaborates on this in his talk:

“With an inflation target, our normal practice is to look ahead and adjust the degree of monetary stimulus to affect the level of aggregate demand and close any gap with potential output or aggregate supply. With inflation expectations well anchored at 2 percent, inflation should sustainably come back to target when economic slack is absorbed and the economy returns to maximum employment.”

So, that’s an odd way to put it. If the Governing Council is guided by macroeconomic theory, best guess is that’s New Keynesian (NK) macro. But in a baseline NK model, the policy rule Schembri specifies would not “anchor” inflation expectations at 2%. For example, consider two policies: (i) low nominal interest rate target forever; (ii) high nominal interest rate target forever. In the absence of any further aggregate shocks, in an NK model the output gap (which is well defined within that framework) will disappear, but the inflation rate will be higher under the second policy.

To have it make sense, within the NK paradigm, we need to add to Schembri’s policy rule some notion of a neutral policy rate, along with a description of how the Bank responds to inflation. In its October Monetary Policy Report, it’s stated:

“The neutral nominal policy interest rate is defined as the real rate consistent with output remaining sustainably at its potential and with inflation at target, on an ongoing basis, plus 2 percent for inflation. It is a medium- to long-term equilibrium concept. For Canada, the economic projection is based on an assumption that the neutral rate is at the midpoint of the estimated range of 1.75 to 2.75 percent. This range was last reassessed in the April 2021 Report.”

So, as far as we know, the Bank still thinks that, roughly, a neutral rate of 2.25% is about right. The press release from the October 27 meeting sees slack disappearing sometime during the period April-September 2022, with inflation returning to 2% by the end of 2022 in the Bank’s projections. But the Bank’s forward guidance says that increases in the policy rate don’t happen until the slack goes away. So it’s hard to see how the forward guidance, the Bank’s own forecast, and the framework described in the Bank’s literature are all consistent, as the Bank seems to be planning for a policy rate far below 2.25% at the end of 2022.

A key development in Bank policy has been the steady reduction in the rate of purchases of long-maturity Government of Canada bonds. At the October meeting, the Bank announced that it was entering a “reinvestment” phase, meaning that, as the Bank’s bonds mature, they will be replaced in the Bank’s asset portfolio. Therefore, for the time being, the nominal value of bonds held by the bank will be held constant. In the press conference following the meeting, Governor Tiff Macklem appeared to signal that reinvestment would end at some point, but the Bank has not been clear about what is to follow. In particular, it would be useful to know whether the bank intends to have a large balance sheet (as in the United States) over the long term, or if it intends to revert to the system in place in Canada prior to the pandemic.

Here are the key entries on the asset side of the Bank’s balance sheet:

And here are the liabilities:

Part of what’s interesting here is that the Bank’s bond purchases have been offset by reductions in its holdings of Treasury bills and its lending through the repo market. So, on net, total assets are lower than earlier in the year, but average maturity has increased. On the liabilities side, total settlement balances (reserves) are down from early 2021, currency outstanding is up, and federal government balances held with the Bank are about the same.

So, if one thinks that quantitative easing (QE) works through the size of the balance sheet, or the quantity of reserve balances held by the banks, then the Bank has been backing off QE for some time, even before the Bank formerly announced the end of net positive asset purchases at the October meeting. I tend to think that QE has little or no positive economic effects, and that its only role in this context is to delay increases in the policy rate. That is, the standard central bank playbook now seems to be that QE programs are wound down before policy rate increases can happen. So, in contrast to the Fed, which will not be finished with asset purchases until mid-2022, the Bank has more policy options in the short term.

So, let’s look at the macro data. Real GDP in Canada looks like this:

The recovery in real GDP has been more swift than in a typical recession, but real GDP is currently 5.1% below the 2010-2020 trend. The labour market data looks better, though.

The labour force participation rate is about where it was prior to the pandemic, and the employment rate is almost there.

The unemployment rate, currently at 6.7%, is higher than it was in Janaury 2020 (when it was 5.6%), but the same as in March 2017 or October 2005, which we wouldn’t characterize as bad times. Also, as in the U.S., the vacancy rate is high. The vacancy rate in Canada in 2021Q2 was 4.6%, higher than the 3.0% recorded in 2019Q4, indeed a higher vacancy rate than in the entire sample, which goes back to 2015.

So, I think we could characterize the Canadian labour market as tight, in the aggregate (we’ll discuss sectoral issues later). In this sense, there is no slack. This looks like an economy with low productivity — because of pandemic restrictions on activity, fear of contagion, and disruption in production, transportation, and retail activity — and it’s basically producing all it can.

An interesting feature of the pandemic is that it’s a sectoral shock, which I’ve illustrated here by looking at some components of GDP.

In the figure I’ve normalized each time series to 100 in 2018Q1. So, residential investment and consumption of durables are much higher than pre-pandemic, while total business investment is lower, and consumption of services is much lower.

In the press release following the October Governing Council meeting, it’s stated:

“… the economy continues to require considerable monetary policy support.”

But, according to how the Bank thinks about the problem, that support is being provided to sectors that are most responsive to interest rates, at the short and long ends. Including residential construction, and consumer durables, which apparently are booming, i.e. not in a state of “ongoing excess capacity,” as the October press release states as the reason for continuing accommodative policy.

Finally, of key importance of course, is the recent behavior of inflation. Here, it’s key to emphasize that the Bank’s current agreement with the federal government states:

“…the Bank will continue to conduct monetary policy aimed at keeping inflation, as measured by the total consumer price index (CPI), at 2 per cent, with a control range of 1 to 3 per cent around this target.”

Inflation is currently at 4.7% (October), year-over-year, which of course is well outside the 1% to 3% range. Even if the Bank were pursuing some type of makeup strategy, it would have made up more than the undershoot in 2020, with average inflation above 2%, even if we go back to 2018.

So, this inflation behaviour was not what was envisioned when the Governing Council wrote down its forward guidance. For example, here’s the Bank’s inflation projection from the July 2020 Monetary Policy Report:

So, note in particular that actual inflation behavior in 2021 is outside even the shaded area, which appears to be a confidence interval. So, the Bank’s mandate is inflation control, as specified in its agreement with the federal government. Inflation is well outside the range specified in that agreement. But the only change in the Bank’s policy has been to move up by a few months the vaguely specified date at which it might consider increases in its policy rate. Note in particular that the reductions in asset purchases were hardwired — these reductions were set in motion long before high inflation was on the map. That is, apparently the Bank decided, almost a year ago, that purchases would fall in steps of $1 billion per week, at every other Governing Council meeting.

But, the Bank’s story (as with the Fed) is that the increase in the inflation rate is temporary. That’s what’s in their forecast (from the October Monetary Policy Report):

Note here that they have the inflation rate peaking in 2021Q4, then declining to close to 2% by 2020Q4. So, I don’t have any good reasons to argue with that, but I don’t think that forecast (or anyone else’s) is worth much right now.

As I’ve argued, there do not exist good reasons for continuing with low interest rates. Though the economy is still adjusting to the sectoral disruption caused by the pandemic, the labour market is tight, the aggregate economy is operating at capacity as far as we know and, of course, inflation is well above target. In the short term there is little harm in increasing the overnight interest rate target in 25-basis-point steps — this won’t make the economy fall off a cliff. Sooner or later, the policy rate needs to go up to a level that is conducive to supporting 2% inflation over the long run, and given current conditions, it’s appropriate to start moving ASAP.

First, how did we get here? In August 2020 the FOMC unveiled a new “Statement on Longer-Run Goals and Policy Strategy.” For our purposes there are three important elements in that Statement. The first is:

“Owing in part to the proximity of interest rates to the effective lower bound [because the neutral interest rate is low], the Committee judges that downward risks to employment and inflation have increased.”

So, the key problem the Fed was focused on was the low real real rate of interest. Low real interest rates imply that the average overnight nominal interest rate consistent with the Fed’s 2% inflation target must be lower, implying that, given typical stabilization policy, the Fed will encounter the zero lower bound (or some effective lower bound) on nominal interest rates on a more frequent basis. The FOMC thinks that zero-lower-bound encounters tend to cause low inflation and low real GDP growth — a notion supported by the New Keynesian academic literature on sticky-price zero-lower-bound economics.

So, this is our first problem. A key issue the FOMC anticipated dealing with — inflation below target — is not the major problem that the FOMC needs to deal with now, which is inflation well above its 2% target.

Second, the Statement says:

“…the Committee’s policy decisions must be informed by assessments of the shortfalls of employment from its maximum level, recognizing that such assessments are necessarily uncertain and subject to revision. The Committee considers a wide range of indicators in making these assessments.”

So, this says is that maximum employment is a goal, but how we measure it is something we can make up as we go along. Third,

“…the Committee seeks to achieve inflation that averages 2 percent over time, and therefore judges that, following periods when inflation has been running persistently below 2 percent, appropriate monetary policy will likely aim to achieve inflation moderately above 2 percent for some time.”

This is a quasi-average-inflation-targeting approach which, in contrast to straightforward inflation targeting (what the Bank of Canada does, for example), takes past misses of the inflation target seriously, and attempts to make up for those misses. But, the Statement leaves a lot out. How much of the past will the FOMC consider when it thinks about making up for past misses? How long is the future makeup period supposed to be? And, perhaps most importantly, given current conditions, how come the FOMC did not specify what would happen if there were inflation target misses on the high side?

So what’s going on in the data? Let’s look at prices first.

Here, red is the headline CPI, blue is the PCE deflator (what the FOMC claims to care about), and green is a 2% trend from January 2020. Everything is normalized to 100 in January 2020. Recall that the FOMC’s Statement of Goals says that, if inflation runs below target, then the Fed is supposed to target inflation “moderately” above target. In the chart, average inflation was below target until March of this year, but by March, the Fed had made up for the undershooting in 2020. In fact, average PCE inflation has been 5.8% per year since February of this year, and will come in at 5.3% year-over-year in December 2021 if the inflation average is the same for the remaining 3 months of the year as in the preceeding 9. So, this is hardly “moderate,” I think, and the undershooting was made up long ago, if we take our base period as January 2020. It appears that the Fed is willing to tolerate inflation well above target, in a manner inconsistent with what was written down in the FOMC’s goals statement.

But what if we use an earlier base period? This is what we get with a base period set at January 2018.

So, note here that the Fed has more than made up for inflation target undershooting going back almost 4 years. But the Fed’s policy rate (the interest rate on reserves) is still at 0.15%, and the FOMC has just committed to a start in reducing the rate of its asset purchases — a process that will not end until the middle of 2022. Given the ideas of FOMC members about inflation control, how do they justify emergency settings for policy instruments given their own Statement of goals and what we’re seeing in the inflation data?

But maybe there’s something in FOMC statements that we can use to make sense of this. From the November 3 statement:

“The Committee seeks to achieve maximum employment and inflation at the rate of 2 percent over the longer run. With inflation having run persistently below this longer-run goal, the Committee will aim to achieve inflation moderately above 2 percent for some time so that inflation averages 2 percent over time and longer‑term inflation expectations remain well anchored at 2 percent.”

But, that’s not right. As we’ve shown, inflation has not been running persistently below 2%, even if we go back almost 4 years. However, maybe the FOMC is worried about the second part of its mandate — the maximum employment goal. After all, the November 3 FOMC statement also says:

“The Committee expects to maintain an accommodative stance of monetary policy until these outcomes [maximum employment and inflation at 2%] are achieved.”

So, how are we doing on the labor market front? First, real GDP has made a rapid recovery, relative to previous recessions.

Though higher than it was in 2019Q4, real GDP is still 2.8% below the post-financial crisis trend. And in the labor market, the labor force participation rate and employment/population ratio are very low, relative to early 2020.

So, those indicators might signal a shortfall relative to maximum employment. However, here’s what we see when we look at other labor market indicators:

The unemployment rate is currently at 4.6%, which is the same is August 2006 or in February 2017, neither of which we would characterize as bad times. More strikingly, the job openings rate is at an all-time high (at least since we’ve been measuring it). For emphasis, here’s a standard measure of labor market tightness, the ratio of job openings to unemployment:

That’s also at an all-time high. So are we at maximum employment? We could argue yes. Why labor force participation and the employment-population ratio are low is something of a puzzle, particularly since we don’t observe this in Canada, where most labor market indicators are close to where they were in early 2020. But one could make a case that low labor force participation is a choice — maybe primarily due to the pandemic. There are plenty of jobs available, and this economy is more or less producing all it can, given the circumstances.

So, unfortunately, this episode illustrates the limits of forward guidance — at least forward guidance that commits the central bank to future actions without anyone thinking through all the possible future contingencies. The FOMC seems implicitly committed to a policy of tapering — reducing asset purchases — before there are any increases in the overnight interest rate target. That just creates foot-dragging into mid-2022, when action now is appropriate, given the FOMC’s own principles, as written down in its Goals Statement and FOMC meeting statements.

If you’ve read other things I’ve written, for example this piece in the St. Louis Fed Review (and there’s more where that came from), then you’ll know that I think the best way to do inflation control is in terms of Fisherian logic. High nominal interest rates for a long time imply high inflation, and low nominal interest rates for a long time imply low inflation. Just ask people who live in Japan. If a central bank is persistently undershooting its inflation target, it needs to find excuses to raise its overnight nominal interest target.

So, according to this logic, shouldn’t the Fed just hold tight with the policy rate at 0.15% and wait for inflation to come down? No, because inflation control is about bringing the central bank’s overnight nominal interest rate back to a level consistent with its long-run inflation target, after any period where the policy rate has been moved around for stabilization policy reasons. For example, the Fed’s response to the pandemic had little to do with inflation control. The Fed thought, correctly, that emergency action was necessary in early 2020, and reduced its policy rate essentially to zero (I’ll leave aside QE, which I think is ineffective). But once that was done, the game is to find excuses to move toward “normal,” which in this case could be a policy rate of about 2%, supposing the long-run real interest rate is about 0%. That is, interest rate hikes should take place in the context of good news. If you’re looking at the right things currently, the real economy is giving off good news. The inflation data is bad news, in a sense, but good news in terms of pressure to tighten.

I think it’s still likely that the burst of inflation we’re seeing is temporary — in fact, that we’ll be seeing inflation below the 2% target next year, in year-over-year terms, if policy settings remain where they are. As I saw Ricardo Reis discussing on Twitter recently, it doesn’t matter much what your theory of inflation is right now. You could be a hard-core Phillips curve type or even an old school monetarist and think that it’s time to tighten. The Fed should get on with it.

I’ll repeat Figure 9 from Part I for easy reference:

Figure 9:

The last thing to note about Figure 9 is that, in early 2018, the interest rate differentials you see before then disappear, coincident with the phasing out of Fed asset purchases. Recall that this is also the time that ON-RRP activity drops dramatically, but not to zero. So, for the most part, when the Fed is no longer purchasing assets, the ON-RRP facility is not needed to the same extent to prop up the overnight repo rate from below.

But that’s not the end of the story:

Figure 10:

Figure 10 shows the repo rate (blue, that’s SOFR), IOER (green), fed funds rate (purple), and 4-week T-bill rate (red), from January 2019 until the pandemic hits. Three of those interest rates track each other fairly closely over this period, but what stands out is the behavior of the repo rate. Remember the end-of-month downward spikes in the fed funds rate in Figure 9? Well, in Figure 10, those are end-of-month upward spikes in the repo rate, which start happening in early 2019, as reserve balances are falling.

Then, on September 17, 2019 (perhaps surprisingly, not a month end), a large shock to the overnight market (more on that in what follows) resulted in the spike you see in interest rates in Figure 10, with the largest effect on the repo rate. Again, that’s not supposed to happen in a floor system, theoretically. Relative to what was going on in 2016-2018, the stress is on the other side of the repo market, appearing as upward pressure on SOFR rather than the downward pressure that the NY Fed was accustomed to mitigating with the ON-RRP facility. So, perhaps unsurprisingly, the Fed began intervening by lending in the repo market, which we saw in Figure 5 in Part I. The Fed intervened by lending on the repo market at IOER which, as you can see in Figure 10, served to essentially peg the fed funds rate, SOFR, and the 4-week T-bill rate to IOER. But of course what was happening was that the Fed was varying the quantity of repos on its balance sheet day-to-day to peg SOFR, just as it would pre-financial crisis. You may be getting the idea that this isn’t really a floor system.

Regarding the September 2019 event, this note written by Board of Governors economists is very interesting. Posted in February 2020, it makes the case that a number of special factors contributed to the spike in overnight rates in September 2019, and discusses the Fed’s response. Essentially, the special factors cited in the note read like the usual laundry list of shocks - unanticipated and anticipated - that the Fed normally has to deal with, through daily intervention, when a corridor system is in operation. Two things are surprising here. The first is that the note gives you the impression that the factors that led the Fed to intervene in September 2019 were temporary, but of course the repo intervention continued at similar levels into the pandemic, as you can see in Figure 5 in part I. The second surprise is that there’s no mention of this episode being abnormal in terms of how anyone - in or outside the Fed - thought floor systems should work. The Fed was saying: “no problem, everything is fine here - and pay no attention to the man behind the curtain.”

Finally, we’ll look at interest rates in the recent past:

Figure 11:

This is data for the same four interest rates as in Figure 10, but for the pandemic period. In Figure 11, there’s first a period of turmoil, and then things settle down from fall 2020 until early in 2021. That period of calm, when SOFR, the fed funds rate, and the 4-week T-bill rate are pegged at one or two basis points (mostly) below IOER, is a period when there was no intervention by the Fed on either side of the repo market, as we saw in Figure 5 - essentially the only time that occurs in the whole period from December 2015 (liftoff) to now. But, in early 2021, when the volume started to increase on the Fed’s ON-RRP facility, there was downward pressure on overnight rates, with SOFR and the T-bill rate effectively pegged by the ON-RRP rate.

So, we’re now back to the regime of 2016-2018, when the Fed was making large asset purchases, the stock of reserves was increasing, and ON-RRP intervention was pegging the overnight rate. Except now it’s more extreme, as ON-RRP volume is getting very high.

Here’s another interesting tidbit. In Jay Powell’s June 16, 2021 press conference, Michael Derby asked a question about the high takeup on the ON-RRP facility. Powell answered:

“So on the facility, we think it's doing its job. We think the reverse repo facility is doing what it's supposed to do, which is to provide a floor under money market rates and keep the federal funds rate well within its range. So we're not concerned with it. You have an unusual situation where the Treasury general account is shrinking and bill supply is shrinking. And so there's downward pressure. We're buying assets, there's downward pressure on short-term rates, and that facility is doing what we think it's supposed to do.”

So, three things. First, the facility is doing what it’s supposed to do alright, but the Fed claims the floor is simpler than operating a corridor, while effectively conducting intervention as if operating a corridor. Second, the business about keeping the fed funds rate in its range is baloney. There are two interest rates that matter in this system, the way the Fed operates it. Those rates are the ON-RRP rate and the rate at which the Fed is willing to auction repos. Those rates bound SOFR from above and below. There can be unusual circumstances where the Fed is both borrowing through ON-RRP and lending by way of repos. But typically whichever facility is operative - the ON-RRP facility or repo auctions - is going to determine what SOFR is pegged at, the ON-RRP rate, or IOER, which is the rate at which repos are auctioned. The fed funds rate is not bounded by the ON-RRP rate on the low side, and IOER on the high side, except to the extent that repos and fed funds borrowing and lending are close substitutes. Third, in his comments, Powell is following a pattern that tends to show up in Fed discussions of anomalies in the operation of its floor system. That is, blame the Treasury. That’s inappropriate I think, as the Fed is quite capable of setting things up so that Treasury actions don’t impinge on monetary policy.

What’s Going On?

The frictionless version of my alternative model in Figure 4 of part I clearly falls short of explaining what is going on, as it tells us that, given an outstanding stock of reserves, IOER should peg the overnight rate. I’ll modify Figure 4, and then discuss what potential frictions are limiting arbitrage in the overnight market.

Given our examination of the data, it seems the Fed’s floor system works in two modes. In the first mode, in the absence of Fed intervention in the repo market, markets clear with IOER > overnight rate.

In periods when the Fed’s ON-RRP facility was operational (seemingly with a relatively high volume), the situation without ON-RRP intervention would look like Figure 12, if there were no ON-RRP intervention. Reserves are willingly held by banks with IOER = R*, but the overnight market clears with the overnight rate R** < R*. So, you would think that that unexploited profit opportunities exist here. For example why don’t banks without reserve accounts that are lending at the overnight rate R** obtain a better deal by lending to banks that can hold reserves at the rate R*? Typically, the answer would be that there are costs for a regulated bank, in addition to the interest payment they might pay to a non-bank lender, to acquiring funds and holding reserves. For example, deposit insurance premia apply to all assets, not just risky assets, and capital requirements, for example the supplementary leverage ratio (SLR), also apply to all assets. The flip side of this is that R* > R** reflects a desire by banks to unload their reserves - the Fed has to pay banks a premium above overnight rates (which as we’ve seen are pegging T-bill rates) to get banks to hold the reserves.

But, in a situation like Figure 12, the Fed may not like the result, as it wants to peg the overnight rate higher than R**. What to do? That’s where the ON-RRP facility comes in:

So note, comparing Figure 13 and Figure 3 from Part I, that the mechanics of how the Fed’s intervention works to peg the overnight rate are the same as in a corridor system, as effectively this is a corridor system, in the sense that IOER is not determining the overnight rate - that’s determined by how the Fed intervenes in the repo market. But what’s the difference between Figure 13 and Figure 3? Of course, its that the Fed has a large stock of interest-bearing reserve liabilities that are confined to the banking system. The reserves want to get out of the banking system, as reflected by R** > R*, and the ON-RRP facility gives the reserves an escape valve.

Currently, it looks like what happens, roughly, is that $1 of extra asset purchases goes into ON-RRP liabilities, not reserves. The current situation looks like a case where the banking system has a glut of reserves, and marginal asset purchases by the Fed are funded with reverse repos. But why wouldn’t the Fed just set the ON-RRP rate equal to the interest rate on reserves, following the logic of how they intervene on the other side of the market, where the Fed lends in the repo market at IOER? Well what might happen in Figure 13 if the Fed pegs the overnight rate at R** = R* is that all the reserves turn into reverse repos. You would think the Fed would like this, as they could talk about policy in terms of a single interest rate target instead of a range, but I’m guessing the issue is that they’re wedded to the idea that stuffing reserves into the banking system actually does something useful. They learned about the money multiplier in money and banking class? Who knows.

Next, let’s think about what was going on during the period January 2019 to early 2020 when overnight rates were being pushed above IOER.

In Figure 14, with the absence of Fed intervention in the repo market, the overnight credit market clears with the overnight rate R** above IOER, which is set at R*. We’ve covered what could be going on in terms of bank regulation and limits to arbitrage when R** < R*, but what’s happening here? Why are banks willing to accept a low return on holding reserves when it seems it would be more profitable to lend in the overnight market?

In order to substitute from reserves to overnight repos in the short run, a bank would have to sell some liquid assets. Principally, those liquid assets are reserves and Treasuries. A bank can get rid of reserves, but only by selling them to another financial institution with a reserve account. But, basically, the banking system can only shuffle the reserves around - it can’t jettison the reserves, except through the the ON-RRP facility, and that’s inoperative in this case. What about Treasuries? First, if all banks had binding liquidity coverage ratio (LCR) constraints, then in some sense the Treasuries are locked up too. LCR requires that banks hold a specified ratio of liquid assets to total net cash flows. The constraint can be satisfied with other than high-quality liquid assets (HQLA - reserves and Treasuries), but the weights on other assets in calculating the ratio are lower. Second, if the overnight rate is high, so is the T-bill rate, so banks get a high return on T-bills and may choose not to sell T-bills to lend in the repo market.

So, there seem to be good reasons to think there are significant arbitrage frictions present that could cause R** > R*, and that’s certainly what we observed in September 2019. What’s the solution? The Fed can intervene by lending in the repo market.

Probably you can predict what’s coming next. In Figure 15, the Fed intervenes by acquiring overnight repos, and pegs the overnight rate at R*, which is also IOER. First, the overnight market intervention is exactly the same as in Figure 2, which models what the Fed was up to in overnight markets pre-financial crisis. Second, what’s perhaps most important is that IOER is not pegging anything. For example, in Figure 15, the Fed could peg the overnight rate below IOER if it wanted to. This is a system where IOER is not a floor, and it’s not a ceiling either. Again, this is just Fed intervention working in the usual way, but with a big stock of reserves outstanding.

A proposal that came from within the Fed well before September 2019, was that the Fed set up a “standing repo facility.” Such a facility is not unusual, for example the Bank of Canada has an overnight standing repo facility, which appears intended to supplement lending through the equivalent of the U.S. discount window, at the equivalent of the discount rate. But the proposed Fed standing repo facility appears to be directed at making the floor system work better, not at making the discount window work better. It seems that people in the Fed have convinced themselves that the problem that created the events of September 2019 is that banks like reserves too much, that Treasury securities are not liquid enough, and thus the demand for reserves is higher than it should be. So, according to this argument, there’s a problem that can be mitigated through a standing repo facility that allows banks to conveniently unload Treasuries, along with other assets, depending on what the facility accepts as collateral.

So, the standing repo facility seems harmless enough, but the problem people say it will correct is, I think, not a problem. The actual problem is that the reserves are illiquid. Sure, reserves are daylight settlement balances among financial institutions with reserve accounts, but they are not acceptable means of payment for any economic entity without a reserve account. Reserves are locked up in the banking system and, in the situation depicted in Figure 14, the reserves cannot escape. For the banking system as a whole, reserves are illiquid. So, the problem is the floor system, not the “illiquidity” of Treasury securities.

What’s This Mean For QE?

So, I think we have determined that any claim that the Fed’s floor system eases fulfillment of the FOMC’s policy rate directive is a bogus claim. Most of the time, the Fed is basically intervening in the same way it did before the advent of the floor system. The only differences are that: (i) sometimes it intervenes through reverse repos rather than repos; (ii) the NY Fed is pegging SOFR, not the fed funds rate.

Where does that leave us? The “floor” system is about QE, that is turning Treasury securities and MBS issued by GSEs (government sponsored enterprises) into reserves. That is, the Fed has decided that it wants to get into the business of debt management. Purchases of long-maturity Treasuries by the Fed reduces the quantity of those securities that are publicly held, and increases the stock of reserves held in the banking system. Purchases of MBS do nothing to the stock of Treasuries, but reduce the stock of MBS held by the public and increase reserves.

Ignoring the MBS part of the process, and focusing on consolidated government (Fed and Treasury) liabilities, here’s what’s happened since the beginning of 2019.

Figure 16

So, in debt management terms, Figure 16 is what we should care about (though there may be issues of on-the-run and off-the-run Treasuries to be concerned with) - it’s the net effect of Treasury and Fed actions that matters, not the Treasury securities held as assets by the Fed. In Figure 1, the total is Treasury notes and bonds plus Treasury bills, plus TIPS, plus reserves. All Treasury security quantities are what is held by the public, that is subtracting securities held by the Fed from securities issued by the Treasury. Notes and bonds are Treasuries with maturities greater than one year, and bills are T-bills, which have maturities of a year or less. The scale on the vertical axis runs from 0 to $20 trillion, so 1 is $10 trillion.

In Figure 16, note the runup in consolidated government debt during the pandemic. But given the collective actions of the Treasury and the Fed, notes and bonds are at about the level they were in early 2020, and bills and reserves have roughly doubled in both cases. It’s hard to see it in the picture, but TIPS have declined as a fraction of the total. In January 2019, bills and reserves accounted for 26.8% of consolidated government debt, notes and bonds for 64.4%, and TIPS for 8.8%. By May 2021, this had changed to 40.2% for bills and reserves, 53.5% for notes and bonds, and 6.3% for TIPS.

So, consolidated government debt is on average of shorter maturity, and much more of it is reserves locked in the banking system. In January 2019, 11.1% of consolidated government debt was reserves, and in May 2021 that number was 19.7%. The Fed claims that turning Treasuries and MBS into reserves is a good thing - it’s “accommodation,” apparently, and this should continue for a long time, according to the FOMC. But what’s actually going on here? The Fed is a financial intermediary that, like any financial intermediary, transforms assets. That’s it’s main function. But why are we supposed to be better off with less Treasuries and MBS in our possession, and more reserves loaded into the banking system?

To the extent that there is any science in debt management (and it seems this area is in need of much more economic research), one of the principles is serving the market. Debt managers spend time talking to financial market participants to find out what they want. More ten years? Fewer 3-month T-bills? So presumably the Treasury had something in mind when it issued the notes and bonds that the Fed is holding in its asset portfolio. Maybe the market participants who make use of 10-year Treasuries aren’t happy with the situation. Indeed, there are anecdotes about unhappy market participants in the midst of QE episodes. For example, in 2020 the Bank of Canada reduced its asset purchases from $5 billion per week to $4 billion. Though the Bank didn’t say so, that seems to have been a response to complaints from market participants about shortages of government debt.

Further, if you think about what’s going on currently, you may be suspicious that marginal asset purchases by the Fed are doing absolutely nothing. It seems that, roughly, each additional dollar of asset purchases shows up as an additional dollar of ON-RRP balances at the Fed, with reserves unchanged. So, at the margin, the Fed is converting long-maturity safe assets into repos. But any financial entity not subject to banking regulation can do that - the Fed has no natural advantage in that activity, so it should have no effect. It’s perhaps astonishing that the Fed has itself tied up in knots about when it is going to even talk about reducing the volume in an asset purchase program that is having no discernible effect.

As well, from our analysis above, it seems there are good reasons to doubt that reserves are a useful form of consolidated government debt, not only relative to T-bills, but relative to any maturity of Treasuries. Reserves are safe assets that can be held only by a closely regulated set of financial institutions, and those regulations mean that this is an inefficient use of safe consolidated government debt. The regulations are there for good reasons, and reserves are indeed useful for intraday transactions. But, as far as I can tell, overnight reserves serve no useful purpose that cannot be accomplished more efficiently with Treasury debt.

Given recent events in the US overnight credit market, this is a good time to re-assess the Fed’s current approach to implementing monetary policy. This is a topic that some people think is just a matter of some technical details, but I think it is much more important than that. Thinking about this helps us understand quantitative easing (QE) and its effects, and I think reveals important aspects of how conventional monetary policy works.

The Poole Model, and How the Fed Uses it to Think About Implementation

The people at the NY Fed have a particular view of how monetary policy implementation works, with and without a large central bank balance sheet. You can find this theoretical apparatus in any number of notes and staff papers written in the Federal Reserve System. Here’s how it goes. We’re supposed to think of monetary policy implementation - fulfilling the directive of the FOMC to hit a target for the fed funds rate - in terms of a partial equilibrium model of the market for overnight reserves. Looks like this:

So, the narrative that typically goes with Figure 1 is that, before the financial crisis, when the Fed ran what’s typically considered a “corridor” system, each day the people running the Open Market Desk at the NY Fed would need to figure out what the demand curve for reserves looked like, then inject the required quantity of reserves into the market. In the figure that quantity of reserves is m*, which then pegs the fed funds rate at R*, provided of course that the NY Fed people are good at predicting the demand for reserves, based on the day, the month, and inside information they may have had about disturbances in the overnight market. In a corridor regime, the interest rate on reserves, or in this case the interest rate on excess reserves - IOER in Figure 1 - is set below the interest rate target R*. For example, before the financial crisis, IOER = 0 in the United States.

But, according to this thinking, the corridor system involves implementation with “scarce” reserves. Alternatively, the Fed could run a “floor” system, which in this modeling framework means that IOER puts a floor under the fed funds rate. If reserves are sufficiently “abundant,” that is m > m**, then through arbitrage in the overnight market the fed funds rate is pegged at IOER. This is where the view that the floor system makes policy implementation “easy” comes from. Figure 1 makes you think that the corridor system is hard, requiring a lot of work to forecast reserve demand and then, with inevitable errors in forecasting, there is some variability in the market fed funds rate from day to day. A floor system is “easy,” as all that has to be done is to administratively set IOER, and arbitrage looks after the rest. With the floor, all the Fed has to be seriously concerned with is what m** is, apparently. Figure that out, and you’re done.

After the financial crisis, there seemed to be a view at the Fed that the floor system was temporary. The Fed would intervene with unconventional asset purchases, making the floor approach a necessity - the Fed wasn’t thinking about simplicity, their unconventional large-scale asset purchases were going to result in m > m**. So, the Fed moved up the introduction of the payment of interest on reserves to fall 2008, so that tightening could happen in the future, when appropriate, by increasing IOER. Otherwise, the Fed would be stuck at a zero fed funds rate, until sufficient assets matured or were sold, or the demand for currency grew sufficiently.

But in 2017 the Fed decided that it would continue indefinitely with its floor system. Why? Well at that point, the FOMC had convinced itself that the floor system made life easier. Indeed, that was part of what Bill Dudley, then the President of the New York Fed articulated in a speech in April 2018 (see also George Selgin’s piece relating to Dudley’s speech). Dudley said:

“In my view, the case for retaining the current floor system is very compelling for a number of reasons.  First, it is operationally much less complex than a corridor system.  In the current regime, the setting of IOER is largely sufficient to maintain the federal funds rate within the FOMC’s target range, as we have seen over the past few years. In contrast, a corridor system requires forecasting the many exogenous factors that affect the amount of bank reserves outstanding, and then engaging in open market operations on a near daily basis to keep reserves at a level consistent with the FOMC’s target range. This task would likely be more difficult now because of greater fluctuation in these exogenous factors relative to when the corridor regime was last in place.”

So, what’s wrong with the Poole model in Figure 1, as a tool for understanding how the Fed implements policy:

1. Reserve balances have a special purpose, in that they’re used, intraday, as settlement balances by financial institutions with reserve accounts. But, though the volume of transactions that occurs over Fedwire in a day is enormous, those transactions can be carried out with a very small quantity of total reserve balances, that is the velocity of intraday reserves was humongous when the corridor system was in place. But overnight excess reserves were essentially zero and, since the Fed has now removed reserve requirements (they did it on March 26, 2020) if we had a corridor system today, it would run with essentially zero overnight reserve balances. In Canada, for example, the pre-March 2020 floor system typically always had overnight reserve balances of essentially zero, in a system with no reserve requirements. So, it doesn’t make much sense to be modeling a demand for overnight reserves when overnight reserves are zero. That is, under a corridor system, IOER is set below the target overnight rate so that banks don’t want to hold the reserves overnight. Basically, it makes more sense to think about Fed intervention in a corridor system in terms of the overnight credit market.

2. It’s actually not easy for the Fed to run a floor system. In fact, as I intend to show you, the Fed’s floor system actually requires active intervention on most days, and in some cases much more intervention than was the case pre-financial crisis.

3. Part of what is weird about the Figure 1 model is that it does not capture what the Fed actually does in a corridor system to peg the fed funds rate. The Fed did not intervene directly in the fed funds market pre-2008, it intervened by way of the repo market, varying secured lending in the repo market day-to-day. The idea is that for financial institutions with reserve accounts - a reserve account is of course required to participate in the fed funds market - overnight repos and participation in the fed funds market are close substitutes. Though the fed funds market is unsecured, the repo market is secured (the borrower posts collateral), and the repo market has more participants and higher trading volume. But, the idea was that, by intervening in the repo market, the Fed could guide the fed funds rate - albeit without observing the fed funds rate (not actually a rate, but a whole distribution of interest rates on any given day) while the intervention was happening. So, if the repo market was so important for intervention, why aren’t we thinking in terms of an overnight credit market instead of a market for reserves? And note that the repo market also plays an important role in the floor system, which we’ll get to.

An Alternative Approach

I’ll construct a different model. This will be quite crude. It’s going to leave out general equilibrium implications - for example how actions by the Fed matter for macro aggregates such as inflation, employment, and real GDP. But, if we want to think about Fed intervention in the daily overnight market, the approach I’ll take seems OK. Perhaps more important is that we’re going to leave out details of the the structure of overnight markets - over-the-counter and intermediated trades - as well as the market for Treasuries, and I’m not going to be explicit about bank regulation, which is going to be important here. So, some of this is suggestive, and it’s going to be more informative to work this out formally, with the details. That’s for later.

Corridor System

First, I’ll try to capture how the Fed pegged the overnight rate prior to the financial crisis. I’m going to ignore the difference between repos and fed funds lending and borrowing, and just refer generally to the “overnight rate” in an overnight credit market in which all financial institutions can participate. Figure 2 shows a supply curve and demand curve, capturing the behavior of overnight market participants. The trick for the Fed in this corridor regime was to assure that excess reserves were essentially zero overnight, by making sure that the overnight market cleared above the zero nominal interest rate on reserves, and hopefully close to the target overnight rate, given by the FOMC directive. So, the Fed did this by purchasing outright a quantity of (mostly) Treasury securities that would be just short of backing the quantity of currency demand at market rates, and then fine-tuning the supply of overnight credit by essentially supplying a perfectly elastic quantity of repos at the target overnight rate.

In terms of Figure 2, the demand curve (and supply curve too) would be shifting day-to-day, due to shocks to the market, and then repo lending would vary each day to peg the overnight interest rate. For example, in Figure 2, the target overnight rate is R*, and the Fed is supplying q* - q** repos. One important point is that the Fed could have made life easier for itself if it had targeted a repo rate rather than the fed funds rate. This could have been done by auctioning funds at R*, with uptake determined endogenously, much like with the overnight repurchase agreement facility currently in place at the New York Fed.

That is, arguing that a floor system is far easier to run than a corridor system, as for example Bill Dudley did in the passage from his 2018 speech above, is a false comparison. The New York Fed could change its approach by targeting a repo rate - the secured overnight financing rate (SOFR) for example - in a corridor system, and this would be very straightforward.

Another approach to pegging the overnight rate in a corridor system is to use reverse repos to fine tune policy each day. In this case, the Fed purchases a quantity of government securities that exceeds currency demand at market interest rates, then makes up the difference with reverse repos, which are then varied day-to-day to accommodate shifts in demand and supply.

In Figure 3, the target overnight rate is R*, and on the day depicted the Fed intervenes by auctioning off q* - q** reverse repos at the New York Fed’s ON-RRP auction. Essentially, there’s not much to choose between repo intervention and reverse repo intervention - the Fed is just on a different side of the market in the two cases. One could even imagine a setup where, on some days, the Fed has repos outstanding, and on other days has reverse repos outstanding. But, in general, reverse repos by the Fed push up the overnight rate and repos push down the overnight rate

Floor System

I’ll start by assuming a frictionless world. Then, we’ll look at features of the data since the Fed’s floor system has been in place, which clearly indicate that this floor system does not work as a frictionless theory tells us it should. Then I’ll discuss what seem to be the key frictions at work, and we’ll sketch out how the alternative theory addresses what is going on in the data.

Now we’ll be dealing with two markets. In the left panel of Figure 4 is the market for reserves. That’s an asset market. There’s an inelastic supply of reserves, and banks have an upward sloping demand curve for reserves, given that the variable in the vertical axis in the left panel is IOER. In the right panel of Figure 4 is the overnight credit market, as before. So that’s the flow of overnight credit on a given day - upward sloping supply curve and downward sloping demand curve, with the overnight interest rate on the vertical axis. Again, we’ll ignore the distinction between secured and unsecured lending in the overnight market, but in terms of empirical counterparts, it’s most useful to think of the overnight rate in the model as the repo rate in reality.

First, we have to address what actually pushes the system into floor mode from corridor mode. To see this, go back to Figure 2, and recall that it’s the repo facility at the Fed that pushes the overnight rate down. So, in the figure, suppose that the Fed reduces R* to IOER through repo intervention (so the flat part of the supply curve is at IOER). And suppose that, given R* = IOER, the demand for currency is exactly backed by the assets in the Fed’s portfolio (assuming no Fed capital). That’s the tipping point. Given those settings, if the Fed buys more assets, given R* = IOER, then that requires an increase in reserves.

Next, we want to understand what’s going on in Figure 4. At this stage we’re assuming that there are no frictions that inhibit arbitrage in the overnight market, so in equilibrium the overnight rate is R*, which is equal to IOER. And, given that interest rate configuration, the markets for reserves and overnight credit clear. The quantity of reserves is m*, and the volume of overnight credit is q*.

So, suppose in Figure 4 that the Fed proceeds to buy more assets (this is QE of course), and the supply of reserves increases to m**. This puts upward pressure on the interest rate on reserves, as banks want to receive higher remuneration to hold the extra reserves. But of course the Fed has set IOER at R*. So something else has to adjust. Here I’ll take some liberties. It’s a bit hard to do this with supply and demand curves, without having worked out a more serious model, but here goes. Since banks will be holding more reserves in equilibrium, which are overnight assets, this should diminish the demand for overnight credit. As well, if the Fed purchased Treasuries to increase reserves, that could reduce the demand for overnight credit as there is less safe collateral available. The supply of credit also falls, as banks are now lending to the Fed by way of reserves rather than lending overnight to other financial institutions. Further, the demand for reserves rises, as arbitrage in part works through non-banks that can’t hold reserves lending to banks which hold reserves. The market equilibrates when both markets clear. Reserves rise to m**, and the volume of overnight credit falls to q**.

How’s the Floor Been Working?

The problem, either with the Poole model in Figure 1, or my alternative frictionless model in Figure 2, is that neither is consistent with what we’ve been observing since the financial crisis, with the floor in operation in the U.S.

I’ll start with observed quantities:

Figure 5:

Figure 5 shows repo activity on the Fed’s balance sheet, reverse repo activity, and the quantity of reserve balances. The latter is on a different scale (right hand) so you can see what’s going on. In the reserves time series, there are several important events:

1. Successive rounds of QE that increase reserve balances after the financial crisis, with QE phased out by late 2014.

2. Reinvestment continued after late 2014, whereby maturing assets were replaced in the Fed’s portfolio continued. Reinvestment was phased out by October 2017.

3. After reinvestment stopped, reserves declined until fall 2019, when asset purchases recommenced.

4. Then the pandemic happened, QE started again, and reserves are now close to $4 trillion, about double the previous peak in 2015.

A key takeaway from Figure 5 is the active intervention that occurred both before and after liftoff (the beginning of conventional policy rate tightening) in December 2015. You can see that ON-RRP intervention (red in Figure 5) began before liftoff. It’s clear the Fed people understood that the floor system wouldn’t quite work, and that ON-RRP intervention would be required for confident targeting of overnight interest rates. But ON-RRP volume declines significantly beginning in late 2017, and then sits at a relatively low level from early 2018 until the pandemic hits in Spring 2020. During that period, another thing happens, which is that Fed intervention begins in a major way on the other side of the repo market. In fall 2019, the Fed began intervening in a big way by lending in the repo market (blue in Figure 5). That continued into the pandemic, and then Fed repos went to zero in mid-2020, and stayed there. Then, the Fed began intervening on the borrowing side of the repo market in early 2021. Since then, ON-RRP volume has steadily increased, and has exceeded $800 billion - a massive amount for Fed repo market intervention historically - in late June 2021.

So, in either the Poole model in Figure 1, or my alternative model, with no frictions, in Figure 4, intervention in the overnight repo market - on either side of the market - should not be necessary. The whole idea behind “easy” floor systems is that, with a large stock of excess reserves outstanding, all the Fed should have to do is set IOER, and that should peg overnight interest rates.

So, what’s going on? As a start in figuring that out, note the following regularities in Figure 1. First, the drop in ON-RRP activity in early 2018 is associated with the phasing out of asset purchases by the Fed. That is, the flow of asset purchases by the Fed seems to be important. Maybe the stock of assets held by the Fed is important too, or it matters what quantities of Treasuries by maturity are held by the public. You can see that in what happens in Fall 2019. In September 2019, Fed asset purchases are still zero, but the stock of reserves becomes sufficiently low at that point that the Fed feels a need to start intervening by doing repo lending every day. Then, by mid-2021, Fed asset purchases have continued at such a high volume, and the stock of reserves is so large that the Fed has to intervene massively through its ON-RRP facility to hold up overnight rates.

Another key element in the story is what was happening with the Treasury’s General Account (TGA), which is another Fed liability:

Figure 6:

Every central bank has to worry about where the central government holds its cash. That is, tax revenue does not come to the government as a constant flow, the government does not spend in a constant flow, and government debt is issued in bunches. So, the federal government needs a cash buffer - a deposit balance that is held somewhere. But if that balance is held with the central bank, that’s going to interfere with monetary policy. Why? If the TGA balance goes down by a dollar, that increases reserves held in the private sector by a dollar, so that works just like an open market purchase of government debt. There are different ways for a central bank to deal with this. In the US before the financial crisis, standard practice was to park the Treasury’s cash in private financial intermediaries, with the TGA balance close to zero at all times. In Canada, when the Bank of Canada ran a corridor system, part of daily BoC intervention involved auctioning off government of Canada cash balances in the repo market.

But, if you think that the floor system works like the Poole model in Figure 1, you would think the TGA balance was irrelevant, as long as it didn’t grow to absorb most of the stock of reserve balances. That seems to have been the view at the Fed and the Treasury post-financial crisis, as the TGA balance grew by a large amount post-2008, as you can see in Figure 6. Then, during the pandemic, the TGA cash balance buffer became enormous by historical standards, peaking at about $1.8 trillion in mid-2020, in Figure 6. Why the Treasury felt it needed such a large buffer is not clear. But, note that the decline in the TGA balance beginning in early 2001 seems to have helped trigger the large increase in ON-RRP volume, through the increase in reserve balances that you can see in Figure 6. So, the large TGA balance during the pandemic presumably just postponed the increase in ON-RRP uptake that would have otherwise occurred earlier.

Finally, even in the period from early 2018 to early 2020, there is still significant activity in the ON-RRP facility, if you look at it on a different scale:

Figure 7:

Compare that to the intervention in the repo market that happened before the financial crisis, in 2007:

Figure 8:

The latter is weekly data. Note that Figure 8 shows us the weekly variability needed in the Fed’s corridor system to peg the fed funds rate, and Figure 7 shows us daily intervention that occurred under the floor system, when supposedly such intervention was unnecessary.

To get a clearer picture of what is going on, we need to look at overnight market interest rates as well. Here’s data for the period after liftoff, 2016-2018:

Figure 9:

I’m not showing you a repo rate, as we don’t have easily accessible data that’s equivalent to SOFR (secured overnight financing rate - now a standard repo rate measure) for that period. Instead, Figure 7 shows IOER (green), the fed funds rate (blue), and the 4-week T-bill rate (red). Over a good part of that period, at least until early 2018, the ON-RRP rate, i.e. the rate at which ON-RRP funds are auctioned off, was set at 25 basis points below IOER. A typical pattern, until early 2018 was IOER > fed funds rate > T-bill rate. If we think of the 4-week T-bill rate as a proxy for the overnight repo rate, effectively the ON-RRP rate was pegging the the overnight repo rate, from liftoff until early 2018. That is, it looks as if the Fed were targeting the repo rate, and the public language about fed funds rate targeting within a band was all smoke and mirrors. For example, note the downward spike in the fed funds rate, at the end of each month, until early 2018. Those spikes are due to some financial institutions making end-of-month balance sheet adjustments, which affected the overnight market - a predictable disturbance that would be offset by intervention in a corridor system, and which the floor system is supposed to correct on autopilot.

[Continued in Part II]

In his press conference, Powell was doing essentially everything an intelligent lawyer could do to articulate monetary policy to the public. He sometimes reads directly from prepared material, he’s clearly well-briefed, he’s worked hard to understand what’s going on, and he doesn’t deviate from the script. People are not directing animosity (any more than is usual) toward the Fed, and part of the reason for that is that Powell does a good job of projecting genuine empathy. That said, I think there are flaws in how the FOMC thinks about policy, and Powell’s lack of economics training - and a general move toward non-economists on the committee - doesn’t help in that regard.

To get up to speed, the forward guidance in the FOMC statement refers to the future path for the policy rate, and asset purchases. First, the policy rate:

“[The FOMC] expects it will be appropriate to maintain this target range [0-0.25% for the fed funds rate] until labor market conditions have reached levels consistent with the Committee's assessments of maximum employment and inflation has risen to 2 percent and is on track to moderately exceed 2 percent for some time.”

Next, asset purchases:

“…the Federal Reserve will continue to increase its holdings of Treasury securities by at least $80 billion per month and of agency mortgage‑backed securities by at least $40 billion per month until substantial further progress has been made toward the Committee's maximum employment and price stability goals.”

Some of the questions Powell got in the press conference were aimed at getting him to be more specific about what those words mean. And the words in the FOMC statement aren’t written off the cuff - they’re carefully chosen, and discussed extensively by FOMC members. So, we should expect the meaning to be clear, right? But, since the FOMC only “expects” to move the policy rate in a way consistent with its current guidance, how is the guidance supposed to help us? What’s a level for labor market conditions consistent with maximum employment, and what conditions are we talking about? What’s “moderately” mean, in specific quantitative terms? How long is “some time?” How could I measure “substantial,” and what would it mean to “moderately exceed” 2% inflation?

Pressed on some of these things, Powell repeated the words in the statement, or referred back to the “Statement on Longer-Run Goals and Monetary Policy Strategy.” He insisted that the forward guidance is “outcome-based.” To me, that would mean that the forward guidance specifies future actions contingent on observable outcomes - that is, outcomes that we all can observe. Clearly, the forward guidance does nothing of the kind. The outcomes are in the minds of FOMC members, and the actions triggered by the outcomes are also something of a mystery. So, a good question would be why the FOMC bothered to write down any forward guidance at all, if what they wrote down has so little content.

But did Powell say anything in the press conference that might help us understand how the FOMC is likely to behave in the future? A couple of questions came up to do with what might happen if high inflation - more than moderately-exceeding 2% for some time, I guess - were to rear its ugly head. In particular, what if the economy is not at maximum employment, but inflation exceeds 2% by a large amount. Powell insisted that “we understand our job,” and that the Fed would “use its tools” to control excessive inflation, referring to the “Statement of Longer-Run Goals…” which deals with what the FOMC should do if the two parts of the dual mandate are in conflict. Problem is the “Statement of Longer-Run Goals…” is in conflict with the current FOMC statement, which says the policy rate stays where it is as long as we’re below “maximum employment.”

An overriding theme in the press conference is that, according to Powell, we have a long way to go, both in terms of seeing “substantial further progress,” (which would trigger tapering of asset purchases) or “maximum employment.” Indeed, the dot plots from the last FOMC projections that were issued suggest that FOMC members think they won’t be increasing the policy rate until early 2024. But this is no ordinary recession. Some sectors are already doing quite well - housing in particular:

Currently, housing starts are at their highest level since the financial crisis. Presumably the housing sector is a primary target of the Fed’s QE program, as the FOMC sees it. So why does that sector need continuing stimulus? You might think the Fed should be warming us up for a phasing-out in their asset purchase program.

As well, job openings are plentiful:

In this chart, I’ve plotted the ratio of job openings to the number of unemployed - a standard measure of labor market tightness. So tightness is well down from early 2020, but at about the same level as in 2016 or just prior to the financial crisis. So that’s not a depressed job market.

Though Powell didn’t want to be nailed down about how he might measure “maximum employment,” I think he’s focused on getting back to some sort of long-run trend. For example:

That chart isn’t great, but I think you can tell which is actual real GDP, and which is the 2.3% growth path (average rate of growth, 2010Q1 to 2021Q1). So according to the chart, we don’t have that far to go. Real GDP is currently 3.6% below the trend. The shortfall relative to trend is larger, though, if we look at employment:

That is, employment is currently 7.1% lower than the 1.6% growth trend. That difference (relative to the real GDP chart) seems to be due to the fact that the large drops in employment occurred in low-productivity sectors (essentially by definition).

So, this looks to me like an economy that’s being held back primarily by COVID. That seems obvious of course, but the point is (again) that this is no ordinary recession. As Powell pointed out, neither households nor banks are financially stressed, as they normally would be when we’re pulling out of a recession. So, once people are vaccinated, the summer and fall start to look pretty good. We could have all the bugs worked out by early next year. The caveats have to do with what is going on in the rest of the world where, in some cases, vaccination on a large scale is just starting, and with the unpredictable nature of coronaviruses.

With respect to inflation control, my concerns were stated, for the most part, in my previous post. The Fed’s new quasi-average-inflation-targeting approach is based on the idea that, if a central bank keeps its policy rate low for a sufficiently long time, then inflation will surely go up. Unfortunately, that doesn’t happen, in theory or practice. It’s belief in that idea that has caused any number of central banks to undershoot their inflation targets. So, unless ideas change on the FOMC, the most likely future is one where inflation and interest rates stay low. That’s OK, but it won’t make Fed officials look like they know what they’re doing.

Given Wednesday’s monetary policy decision by the BoC, and the publication of April’s Monetary Policy Report, it’s a useful time to take stock of what the BoC is up to.

In some ways, the BoC is as transparent as any central bank but, particularly relative to the Fed, we know little about what goes on in the discussion that leads up to a meeting of the Governing Council, where decisions are made 8 times per year. The Governing Council (the Governor, Senior Deputy Governor, and four Deputy Governors) meets in secret, no vote is taken on the decision, and the Governor can basically do what they want. So, if something goes wrong you know who to blame - the guy at the top of the page, Governor Tiff Macklem.

Taking into account what we’ve had to deal with since a year ago, the Canadian economy is in good shape. This is from the April Monetary Policy Report:

Fourth quarter real GDP was an upside surprise, which caused the BoC to revise up its forecast, as reported in its previous Monetary Policy Report. That upward revision is consistent with what private forecasters are saying, so there is considerably more optimism about future macroeconomic activity than was the case a couple of months ago. Eyeballing the chart, the BoC thinks that real GDP will be back on trend in the latter half of 2022. Keep that time horizon in mind, as that will be important for how the Bank has framed its forward guidance. The monthly data strengthens the case for optimism. Here’s employment data from the Monetary Policy Report:

There’s still dispersion in employment across groups, for example employment of low-wage workers remains low due to continuing COVID restrictions, but employment growth is generally good and, again, has surprised on the upside. And of course, particularly since the BoC has a headline CPI inflation target of 2% (in a 1%-3% range), we have to pay attention to inflation. Again, from the Monetary Policy Report:

Just prior to the pandemic, inflation had been pretty much on target. As in the US, inflation fell in Spring 2020, with attendant measurement issues that tended to temporarily bias inflation down. Measured inflation is currently at the bottom of the 1% to 3% range (though I see that the March number came in at 2.2%), but we’re about to move into a period where we’re calculating year-over-year inflation from a low base in Spring 2020, and we’re also seeing large increases in commodity prices. As a result inflation should be temporarily high - perhaps toward the top of the 1% to 3% range - until the fall. That’s the BoC’s take anyway, and I agree with it. Where we may disagree is with respect to what might come later.

Unconventional monetary policy is new to the BoC, but they have been fairly aggressive about it since March 2020.

QE in Canada has consisted primarily of purchases of long-maturity Government of Canada bonds, initially at a pace of $5 billion a week, subsequently reduced to $4 billion per week (seemingly because this was having disruptive effects in financial markets), and then to $3 billion per week in the policy decision yesterday. The latter appears to be the beginning of a tapering-off in bond purchases, though in the press conference Macklim wouldn’t confirm that.

An interesting part of the intervention, evident in the last chart, was substantial repo market intervention (some was term repos), on the lending side, presumably intended as a broad-based lending operation, though BoC officials never said much about it, focusing on QE. But note that, since the middle of last year, there has not been much change in the size of the Bank’s balance sheet, as net bond purchases have been roughly matched by a declines in repo lending, and in the Bank’s holdings of Government of Canada T-bills. In the last few weeks, though, balance sheet size has come down. On net, the average maturity of the Bank’s asset portfolio has increased substantially.

On the liabilities side of the balance sheet:

In this chart, the large-scale asset purchases are reflected in increases in liabilities, primarily reserves (deposits of financial institutions with the BoC), and government deposits (basically the government’s chequing account with the BoC). Presumably the federal government thought it important to have a buffer of deposits with the Bank during the pandemic. Those balances have fallen over time, though they’re not back to normal levels yet. Note that, everything else held constant, a reduction in government deposits will increase reserves one-for-one, which in part explains why reserves have continued to increase since mid-2020. However, in recent weeks, the declines in the Bank’s short-term assets have reduced reserves as well.

To get a full picture, so we can evaluate the effects of QE, we’ll look at interest rates. Short-term rates first:

In March of 2020, the BoC reduced its target for the overnight rate to 0.25%, where it has been since. Given the large-scale asset purchases, the Bank has been running a floor system. That is, in theory the interest rate on reserves (BoC deposits), at 0.25%, should determine the overnight rate, through arbitrage in overnight credit markets. But you can see that has not happened. Since late last year, the gap between the interest rate on overnight central bank balances and the overnight rate has increased to about 10 basis points, as shown in the chart. And the margin between the interest rate on reserves and the three-month T-bill rate is even larger, at close to 20 basis points at times. The Fed’s floor system, in operation since the financial crisis, is notorious for friction, and lack of arbitrage, but one might have thought that wasn’t the case in Canada, given the BoC’s prior experience with a floor system in the period from Spring 2009 to Spring 2010.

This chart shows the same interest rates for the earlier period. Typically, the overnight rate was closer to the interest rate on reserves during that period than is the case now. The overnight rate was at most one or two basis points on either side of the interest rate on reserves. And the 3-month T-bill rate was usually lower than the overnight rate, but at most only 10 basis points lower than the interest rate on reserves.

A standard narrative that you hear in US monetary policy circles is that floor-system deviations of the overnight rate from the interest rate on reserves will go away with sufficient reserves in the system. But here, the interest rate differentials have become larger as the size of the BoC’s balance sheet has increased, suggesting that large-scale asset purchases actually make overnight markets less efficient. How could that happen? Basically, the Bank is taking good collateral (government debt) out of financial markets, and replacing it with a less-useful asset, Bank of Canada deposits. Note in particular that financial institutions require a premium of up to 20 basis points to hold reserves rather than Treasury bills. Further, if the BoC thinks that their control over short-term interest rates could be better under a floor system, they should think again.

Let’s turn to long bond yields:

As in the U.S., there were sharp declines in bond yields with the onset of the pandemic and the drop in the policy rate, but in the figure the 30-year and 10-year bond yields have increased about 100 basis points from the middle of last year. Indeed, these bond yields are more or less back where they were before the pandemic. Unless you want to make the case that the inflation premia implicit in current bond yields are higher than in January 2020, say, it’s hard to see how QE did anything. The primary purpose of QE, as BoC officials explain it, is to reduce long bond yields. This is supposed to be “stimulus,” a word that Macklem used frequently in his Wednesday press conference to describe the Bank’s QE program does.

So, to me, the QE program seems a problematic component of the Bank’s policy. The world’s central bankers continue to defend QE as an effective policy, but there’s little or no evidence that it helps in terms of a central bank’s ultimate goals, and it could do harm to the functioning of overnight credit markets. Given the potential issues with QE, and the experimental nature of the program in a Canadian context, you might expect a greater effort on the part of BoC officials to explain how they think it should work, and to provide evidence that it is, in fact, working. Here’s what the Monetary Policy Report says:

“The cumulative stock, the flow and the composition of quantitative easing (QE) purchases continue to put downward pressure on borrowing rates, supporting economic activity. QE also reinforces monetary stimulus provided by the Bank’s forward guidance. This guidance has committed to holding the policy interest rate at the effective lower bound until economic slack is absorbed, so the inflation target is sustainably achieved.”

Here, I was hoping for some information on exactly how QE should work, if at all, and what characteristics of the central bank’s balance sheet - the size of the balance sheet, the first derivative of the size of the balance sheet, the maturity of asset purchases, what’s happening on the liabilities side - are going to matter for the effects of large-scale asset purchases. The above statement just says that anything could matter, and there’s no explanation, as far as I can tell, of how or why. In the press conference, Macklem just repeated what’s in the Monetary Policy Report.

The above statement links QE to the BoC’s forward guidance, so we should look at that. This is from the last paragraph of Wednesday’s press release, and I’ve included the last sentence of the previous paragraph.

“…As slack is absorbed, inflation should return to 2 per cent on a sustained basis some time in the second half of 2022. "

“Even as economic prospects improve, the Governing Council judges that there is still considerable excess capacity, and the recovery continues to require extraordinary monetary policy support. We remain committed to holding the policy interest rate at the effective lower bound until economic slack is absorbed so that the 2 percent inflation target is sustainably achieved. Based on the Bank’s latest projection, this is now expected to happen some time in the second half of 2022. The Bank is continuing its QE program to reinforce this commitment and keep interest rates low across the yield curve. Decisions regarding further adjustments to the pace of net purchases will be guided by Governing Council’s ongoing assessment of the strength and durability of the recovery. We will continue to provide the appropriate degree of monetary policy stimulus to support the recovery and achieve the inflation objective.”

At first glance, it’s hard to tell whether this is state-based or calendar-based forward guidance. State based says: “we will do x if we see y.” Calendar based says: “we will do x after z months have passed.” At the press conference, Macklem insisted that the guidance was “outcome-based,” which is the former. But based on what state or outcome? Saying the BoC will keep the policy rate at 0.25% until “economic slack is absorbed,” isn’t much help, as we can’t observe economic slack. Further, Macklem wouldn’t be pinned down in the press conference as to what slack measures he might be looking at - a wide, non-specific array, apparently. But I think you can infer what he’s talking about from the first chart. Late 2022 is when the Bank’s forecast for real GDP catches up to the post-financial crisis trend, extrapolated from fourth quarter 2019. And they’re hoping that inflation is on track in late 2022, as that’s when they’ll start to tighten. Why couldn’t the Governor just say that, and keep it simple?

Here are some issues with the policy, and things that could make life difficult for the Governing Council, and the Governor in particular:

1. The BoC claims it is using QE to keep interest rates “low across the yield curve,” to provide stimulus. Presumably, in good part, they want to stimulate the housing market with low mortgage interest rates. But the housing market is booming, and apparently in no need of stimulus. Here’s the latest data on housing starts:

   

2. I don’t think this recovery will look typical. Some sectors - including housing - are already doing well, and the ones that are currently doing badly will come back in a hurry once more people are vaccinated and restrictions are lifted. The caveat is that there will be parts of the world that will be doing badly for some time. But the BoC could be surprised again on the upside, moving the reckoning date forward from late 2022 to something more like early 2022.

3. The Fed seems intent on staying at the zero lower bound into 2024, so carrying out the BoC’s forward guidance likely implies a significant departure from US monetary policy.

4. The longer the BoC keeps the policy rate low, the greater the risk of Japanese syndrome - a low policy rate and low inflation, indefinitely. For example, what if the BoC forecast is correct, in terms of real GDP growth, but we get to late 2022 and inflation is at 1%? The Bank will then have a hard time justifying tightening, but they won’t be able to sustain 2% inflation without getting back to what the Bank considers the “neutral rate.”

We’ll start with current time series for year-over-year percentage changes in the headline pce deflator and the consumer price index. The former, of course, is the Federal Open Market Committee’s inflation measure. Given what’s laid out in the FOMC’s “Statement of Longer-Term…” that’s the measure of inflation that the FOMC targets at 2% (more or less).

But, perhaps more importantly, the pandemic has brought with it large changes in relative prices.

The pandemic is primarily a sectoral shock. Some sectors (restaurants, bars, hotels) were partially or entirely shut down (through government restrictions and private choice), while others (grocery stores, garden centers) experienced an increase in activity. And the relative prices of things consumed at home went up while the relative prices of things consumed away from home went down. On net, measured inflation went down substantially during Spring 2020, and then increased, as shown in the first chart. But, at about 1.5%, inflation (according to the Fed’s targeted measure, pce inflation) is still below target.

But where is inflation going in the future? If you sample opinions on this you’ll find a lot of dispersion, and each group seems to have its own unique concern. What are the key issues?

Measurement

Those large relative price swings shown in the second chart represent trouble for price level measures, particularly the consumer price index, which is a weighted average of prices with the weights fixed. So, for example, the CPI exaggerated the importance of prices in shut-down sectors, and therefore biased measured inflation downward in the Spring of 2020. The good news is that, in the second chart, most of the dispersion in relative prices has gone away, and part of what remains is due to secular changes in relative prices that had started before the pandemic. But, into the fall, year-over-year inflation measures will be upward-biased, given the temporarily low base in Spring 2020, and reversals in the Spring 2020 relative price movements.

Commodity Prices

Commodity prices fell at the onset of the pandemic, and have recently increased substantially.

Here, I’m including a long time series, to show that we’ve seen large commodity price swings before. Indeed, the decrease and increase in 2008-2011 were much larger than what we have seen since early 2020. It’s useful to plot commodity price inflation against pce inflation:

So, if you look at the scales on the axes in the last chart, and fit an eyeball regression line, it looks like pce inflation goes up 4 or 5 percentage points for every 100 percentage points in commodity price inflation. This is quite crude, but should relieve at least some of your concern if you think that the recent runup in commodity prices will inevitably lead to a big surge in consumer price inflation. Further, an idea driving concern about commodity prices is the “wage-price spiral,” according to which input prices increase prices of final goods and services, which induce workers to demand higher wages, increasing prices, etc. One of the useful insights Friedman left us with is that “inflation is always and everywhere a monetary phenomenon.” I think that’s true, and in this instance what it means is that a wage-price spiral can be put out of its misery by the central bank (possibly by doing absolutely nothing in terms of course-correction).

Fiscal Policy

This has created some conflict, even among the otherwise-unified elder Cambridge MA types. For example, Larry Summers and Olivier Blanchard think that the ARP package will potentially cause inflation to explode. Paul Krugman doesn’t. What are people thinking about here? With Summers and Blanchard, this seems more or less an IS/LM/Phillips curve story. According to them, we may need stimulus, but not something this big. Too much stimulus produces “overheating,” apparently. That is, large deficits increase demand, but too much demand means the unemployment rate gets too low, and according to Phillips curve logic, inflation goes up. I know that there are stories about deficits during wars - the Korean War and the Viet Nam War, for example - generating inflation, but I went looking for these effects in the data and didn’t find much.

The data in the chart is for 1960-2020. I know this is crude, but I don’t see a negative correlation there. We could also look for the Phillips curve - in a similarly crude fashion - in the recent data.

What’s in the chart is unemployment rate vs. pce inflation rate, for 2009-present. Here I’m supposing that a Phillips curve person would think that inflation expectations were anchored at 2% over this period, and that the natural rate of unemployment was roughly constant. Maybe what we see in the chart is what people mean when they say the Phillips curve is flat. But, if there’s a consensus view in the Phillips curve community that the Phillips curve is flat, and you think the unemployment rate causes inflation, it seems there’s no reason to get bent out of shape about high deficits causing high inflation in the immediate future.

The other story I’ve heard about fiscal policy and inflation is a fiscal theory of the price level (FTPL) narrative. In the FTPL, the government debt is priced like any asset in basic asset pricing theory - the real value of the government debt is the present value of the expected, and appropriately discounted, payoffs on the government debt. For government debt, the relevant payoffs are future primary government surpluses. So if the government is planning a persistent increase in primary deficits, that should cause a reduction in the real value of the government debt or, given the nominal government debt, an increase in the price level. A problem with this is that what’s relevant is the entire future path for the primary primary government surplus, which of course is something we can’t observe - we’re dealing with beliefs about the future. We could deal with this by doing something analogous to how monetarists approached the inflation problem. Monetarists would typically argue that the demand for money is stable, and therefore we can predict inflation from the path for the money supply. Similarly, if the present value of anticipated future primary government surpluses is stable, this implies stable backing for the government debt, and the price level should move in proportion to the nominal government debt. Let’s try that.

The chart shows US government debt as a percentage of GDP. Under the assumptions I’ve been discussing, we should see the analogue of constant money velocity, that is a constant ratio of nominal debt to GDP, which of course is not what’s in the chart.

Another illustration of this point is in Japanese data from 1995 to the present.

From 1995 to 2019, Japanese government debt increased from under 100% of GDP to more than 230% of GDP, but the consumer price index was essentially flat. So, clearly it’s possible to have large expansions in government debt outstanding with no consequences for inflation.

There is plenty of theory and evidence to suggest that the relationship between fiscal policy and monetary policy is important. And there are plenty of examples of off-the-rails fiscal policy leading to high inflation, if not hyperinflation (see for example, Tom Sargent’s “The Ends of Four Big Inflations”). But, I think the people who gave the central bank the job of controlling inflation had the right idea. Fiscal policy can indeed create enough dysfunction that inflation runs out of control, but under the relatively “normal” conditions I envision for the immediate future, I think it’s possible for the Fed to meet its 2% inflation target on a consistent basis. Whether it will is another matter.

Inflation Control

What should give you the most confidence that inflation will stay low, after we get through the fall, is what the FOMC has told us about its policy rule. Prior to the pandemic, an idea took root in monetary policy circles that the policy tightening that had occurred between December 2015 through December 2018 had been overly aggressive. I don’t think this makes any sense. First, this was a long-delayed “normalization” of the policy rate, which had been in the planning stage since 2011. Second, relative to historical experience, this tightening cycle was more drawn-out, with a smaller increase in the policy rate. For example, it consisted of a hike of about 230 basis points in the fed funds rate over 3 years, whereas in the June 2004 to August 2006 period, the fed funds rate increased 425 basis points over a little more than two years. Finally, judged by objective measures of performance, the FOMC was successful, in that, by late 2018, the unemployment rate had fallen to 3.9%, and inflation was essentially at target, at 1.9%. So what was there to complain about?

This notion of a policy “error” helped propel a change in the FOMC’s “Statement of Longer-Run Goals…”. The change in policy amounts to some kind of quasi-inflation-averaging approach:

“…following periods when inflation has been running persistently below 2 percent, appropriate monetary policy will likely aim to achieve inflation moderately above 2 percent for some time.”

And, here’s another important change in the longer-run goals statement:

“…the Committee’s policy decisions must be informed by assessments of the shortfalls of employment from its maximum level…”

That’s important, as it essentially specifies the second part of the dual mandate as an unattainable goal. That is, there will perpetually be shortfalls of employment from its maximum level - this isn’t some “natural” growth path around which the economy fluctuates.

So, what does that mean for actual policy decisions? The key forward guidance, from the statement for the last FOMC meeting in March reads:

“The Committee decided to keep the target range for the federal funds rate at 0 to 1/4 percent and expects it will be appropriate to maintain this target range until labor market conditions have reached levels consistent with the Committee's assessments of maximum employment and inflation has risen to 2 percent and is on track to moderately exceed 2 percent for some time.”

What do FOMC participants think this means for the future path of the target fed funds rate? That’s in the dot plot from the last FOMC meeting:

These are projections for each FOMC participant of the fed funds rate at the end of 2021, 2022, 2023, and in the “longer run.” So, even 32 months hence, a majority of FOMC participants is thinking that it won’t be a wise idea to be increasing the fed funds rate target from where it is now. That is, most of them think that the US economy will either be short of maximum employment, or on the low side of 2% inflation, going out 32 months. I think that majority assessment is correct. First, the FOMC can eternally argue that the US economy is short of maximum employment, which is unobservable. Second, given the projected policy, inflation less than 2% will be self-confirming.

What’s that about? I first encountered this idea when Jim Bullard was talking about it it in 2010. Jim knew about a paper by Benhabib, Schmitt-Grohe, and Uribe on “The Perils of Taylor Rules.” Basically, the idea is that, in standard macro/money models in which we know the global dynamics, a sufficiently aggressive Taylor rule (satisfying the Taylor principle) will have poor properties. That is, a Taylor-rule central banker with an inflation target, who observes inflation below target will aggressively cut the central bank’s nominal interest rate target. Unfortunately, inflation dynamics don’t work as the central banker had hoped, since they forgot to take account of Fisher effects. So, with a decline in the nominal interest rate, inflation actually falls, which leads to further rate cuts, and lower inflation. This process comes to rest at the effective lower bound, where the central banker is stuck, forever expecting a rise in inflation that never happens, and perennially undershooting the inflation target.

For me, that idea was a revelation, as it helped make sense out of some puzzling observations, in particular widespread inflation-target undershooting post-financial crisis, and 26 years (now) of average inflation at about zero in Japan. What’s that mean for the United States, going forward? The fed funds rate target has already been effectively zero for over a year now and, as I discussed above, the FOMC has locked in a view of inflation control which is likely to keep the fed funds rate target at zero going out at least into 2024. And theory and evidence tell us that persistently low nominal interest rates induce low inflation.

Conclusions? Inflation will be a little higher, maybe even above the 2% target, as measured by the pce deflator, over the next few months. But going into the fall, when I expect we’re going to be close to being back on a long-term real GDP growth trend, inflation will start to come in below the 2% target, and that’s what we’ll see for a long time, barring some dramatic change in the FOMC’s policy regime. In terms of macroeconomic performance, that will be fine, except that policymakers will at some point get agitated about below-target inflation. We like it when our central bankers just relax and do their work, in a way that’s boring to watch.