Accidentally testing NGDP level targeting

Just a few idle comments to follow up on yesterday’s post.

As we settle into the soft landing, it seems to me that we have, whether by accident or by design, tested market monetarism - the idea that the Fed should aim for a nominal GDP target. And it has passed the test with flying colors.

When Covid first hit, I was afraid that this cycle would be an example of a contraction that nominal GDP targeting could not handle well. Most recessions since WW II have been monetary. Nominal GDP targeting is meant to focus the Fed on a target that is more relevant to nominal cyclical stability than inflation is.

One way it does that is to allow inflation to be a bit counter-cyclical. If the run-rate of GDP growth is 3% real growth plus 2% inflation, then some sort of shock - like a war or a pandemic - that temporarily knocks real growth down to zero will have disruptive effects if we stay on the 2% inflation path. Cyclically, nominal trends are important, and a drop from 5% nominal growth to 2% nominal growth could lead to unemployment and other things we associate with recessions.

But, if the Fed targets nominal GDP, then temporarily inflation will rise to 5%, and the stability in the flow of nominal incomes will help to avoid unnecessary dislocations. Then, when the economy grows back, inflation can fall below the intended trend as real growth catches up.

When Covid originally hit, I thought it would be an example of the limits to nominal GDP targeting. Dislocations happen when markets can’t clear. One way they don’t clear is that falling prices create problems. This was a big deal in 2008, when home prices declined significantly, it created all sorts of knock on problems when homeowners were underwater on their mortgages, etc.

The same thing happens when nominal market wages decline. It doesn’t go over very well if employers go around telling all their existing employees that they have to lower their wages in order to keep hiring new workers. So, wages tend to be sticky and instead employers stop hiring or they lay off workers.

So, my fear was that if real shocks hit a large enough portion of the economy, the optimal monetary policy might be even more aggressive than nominal GDP targeting. In order to prevent disemployment in all the sectors suffering shocks, it might be useful to actually let inflation and nominal GDP both rise above trend, and allow nominal excess in order to bring nominal spending in the affected sectors up to a nondisruptive level.

That might be the case, although, in hindsight, maybe fiscal policy, like checks sent to households and all the industry supports that were initiated, addressed those problems during the Covid shock.

In the initial shock, monetary policy didn’t even come close to even hitting a stable nominal GDP growth shock. In the second quarter of 2020, real GDP declined at an annualized pace of 28% and nominal GDP declined at an annualized pace of 29% (8% on a strict quarter-to-quarter basis).

What would have happened if the Fed had flooded the economy with so much cash that they raised spending in that quarter by another 8%? I think maybe it would have been for the best, on the whole, but I’m sure there would have been a great uproar about 28% annualized inflation.

Yet, in the end, with the help of those fiscal policies, we came through Covid better than any of us might have expected under any monetary framework. So, maybe even an imperfectly underapplied nominal GDP target can still create cyclical stability that is a vast improvement over past experience.

Then, the secondary shock of 2022 came, and we got another chance to accidentally test nominal GDP targeting. As Figure 1 shows, when all was said and done, the Fed really did follow the path of an aggressive nominal GDP path target. By the end of 2021, nominal GDP was back at a 5% trend line dating from 2016 or earlier. And, especially, economic trends in 2022 followed a peculiar nominal GDP path. In fact, the inflation bump in 2022 was exactly the sort of thing we would see with a deep supply shock under a nominal GDP targeting regime.

Figure 1

Initially, after the Covid shock, there was a very aggressive reversal in real GDP, making up about half of the lost ground. As we can see in Figure 2, nominal GDP growth continued to catch up in 2021, through a combination of both high inflation and high real growth. By the end of 2021, above-trend nominal GDP growth had brought the level of NGDP back to a 5% growth track - see Figure 1. (Let me be clear. This is not a policy that the Fed has ever explicitly claimed to follow. They only followed it in practice.)

Figure 2

Then, when nominal GDP growth was back on that trend line, it suddenly pulled back substantially. Nominal GDP growth was throttled back from 14.6% in 4Q 2021 to 6.2% in 1Q 2022. That is a crazy trend shift! And it suggests that Fed policy in late 2021 - with the target interest rate still at zero - was extremely tight on a quarter-to-quarter basis.

But, at the same time, real GDP growth dropped from 7.0% to -2.0%. That is a huge amount of instability. That’s a larger one-quarter shift than in any quarter during the Great Recession. Except for the 2020 shock, you have to go back to 1981 to find a one-quarter shift that large.

An inflation-targeting Fed would have pulled back on growth another 6% in order to bring inflation down to 2%. I just cannot fathom wishing that had been the case.

Now, in practice, the Fed doesn’t currently remotely have or use the tools that could target inflation and growth paths with such specificity in the face of such massive real shocks. But, this makes their success all the more impressive.

You could say that in the 2022 second-phase real shock, the Fed did do what I had originally wondered if they should. They glided nominal GDP growth down quite calmly toward a 5% growth path, but when the real shock hit in the first half of 2022, they let it remain slightly above the 5% growth level trend line, which surely helped to salve the dislocations of that large secondary 2022 supply shock.

Then, just as impressively, in 3Q 2022 when real GDP growth quickly bounced back up from -0.6% to 2.7%, a change in trend of 3.3%, the Fed pushed down nominal GDP growth so that inflation declined from 9.1% to 4.5% - a change in trend of -4.6%, more than countering the positive bounce in real growth.

The Fed has been hitting a nominal GDP growth path pretty closely under conditions that were highly volatile and very difficult.

As I mentioned in yesterday’s post, some influential economists, operating from an interest-rate-targeting and inflation-trend targeting framework, are starting to wonder if the neutral target interest rate is higher now than it was before Covid.

It is. And an important reason for that is that the real growth rate has returned to levels that were common before the Great Recession. . . And an important reason for that is that the Fed has engaged in, de facto, 5% nominal level growth targeting. This greatly reduced the frictions of the business cycle.

As Skanda Amarnath and Preston Mui at Employ America have argued, a strong labor market encourages a high rate of quits and job switching and other types of matching and decision making that increase productivity.

Since I was Sumner-pilled on the value of nominal GDP targeting, I have felt like an underappreciated potential of more cyclically stable monetary policy would be a rise in productivity over the cycle, for the sorts of reasons that Employ America is pointing out.

The economists who use interest rates and inflation as their policy benchmarks will be late to this realization. They are still attributing the decline in inflation to rate hikes in 2022. The Fed just hit a home run and and they're debating whether the Fed swung at a pitch out of the strike zone.

What does this tell us about nominal GDP targeting?

One of the most important advantages of nominal GDP targeting would be removing interest rates from the monetary conversation. The entire financial industry seems to be operating on a framework encouraged by the Fed, that the Fed controls interest rates and interest rates play a strict causal role in borrowing, investing, and spending. They think “ZIRP” (zero interest rate policy) was responsible for all sorts of excesses because supposedly low interest rates encouraged malinvestment and excessive prices in asset markets. But, ZIR was not a P, and interest rates aren’t a good indicator of monetary policy - at all.

Just removing interest rates from the monetary policy conversation would be nice, if only to relieve so many investors of the stress that comes from mistakenly worrying that assets are routinely artificially overpriced.

NGDP targeting would also steer public focus away from inflation. But, I don’t think inflation would ever completely disappear from public focus. Changing prices have marked effects on family budgets. It takes a convoluted macroeconomic model to associate interest rates to monetary policy. It doesn’t take one to associate inflation with monetary policy.

I think our accidental flirtation with NGDP level targeting gives us a window into how nominal NGDP level targeting would play out. It would be a boon for American workers and American economic growth. And, it would lend itself to countercyclical inflation.

When that countercyclical inflation happens, the public will be upset. Really, their economic stresses will have been the result of declining real production. But, they will misinterpret the inflation as the source of their stresses. I doubt we can do much about that. Money illusion is deeply embedded in our perceptions.

But nominal GDP level targeting is so successful at smoothing the business cycle that the good news will outweigh the misperceived bad news. It might always be a Sisyphean task to bring the public along and for the public to appreciate that success. I look forward to a day when our macroeconomists help with that education rather than joining in the misapprehensions.

At least JPow! has provided us with the opportunity to see what is possible. Who will take that opportunity and how soon? Only time will tell, and it appears that it will take a while.