Let’s start with the most obvious example, the annual spike in NGDP around the Christmas shopping season. I hope it’s obvious that the Fed should not try to smooth out this NGDP seasonality with monetary policy.
On the other hand, the fall in NGDP during 2008-09 was clearly undesirable. So what about the coronavirus?
In my view, the Fed should aim for on-track nominal spending one, two and three years out, not in April and May of this year. If virus fears cause many businesses to shut down, and/or if retail spending is temporarily depressed (as seems likely to me), the Fed probably could not and should not try to prevent that, as long as expectations for 2021 NGDP remain on track. Here the Christmas analogy holds.
The further out in the future you go, the more it becomes like 2008-09 and the less it becomes like the spike in NGDP at Christmas. I haven’t given much thought to this issue, so I won’t try to pick a dividing line. But certainly the argument for the stabilization of NGDP growth expectations becomes stronger the further out you look.
I don’t know what’s going to happen to the economy this year, as we’ve never had a shock quite like this except during August 1918 – March 1919, when there was a very short (7 month) recession and a very quick recovery before and during the Spanish flu epidemic. Friedman and Schwartz don’t even mention the flu, instead pointing to the end of WWI. (A similar short recession occurred in 1945). And this current epidemic will likely be much less severe than the one in 1919. On the other hand, we are far more risk averse today, so the economic effects could easily be far larger.
Here’s something to look for. Suppose we have a mini-recession; would that tell us anything about the business cycle? I think it would. Previously I speculated that our total lack of mini-recessions (at least since WWII) is due to the fact that business cycles have been caused by monetary policy failures, where policy reacted too slowly to the onset of recession. This failure was sort of built into how the Fed operates—interest rate targeting with inertia, and also a lack of level targeting. If the first “real” recession of my lifetime (except arguably 1974) is also a mini-recession, that would tend to provide support for the hypothesis that our lack of mini-recessions reflects the nature of demand shocks and monetary policy in the US.
PS. When I say “real recession” I don’t mean “actual” recession, I mean supply-side recession.
PPS. I define a mini-recession as a period where unemployment rises by a total of less than 2% during the contraction and early recovery, but by more than 0.8%.
Here’s a picture from the Spanish flu:
READER COMMENTS
Thaomas
Mar 2 2020 at 6:29pm
A supply side shock, if it can be identified, would mean that the “maximum” employment that the Fed is legally obliged to seek would fall and so they would be correct not to adjust their instruments to attempt to prevent that fall. There would be no reason to abandon their PL target and it would be wise to announce a re-commitment to achieving it and explain that because of the fall in productive capacity, inflation should be expected to accelerate in the short term, but this does not mean that the long term target for the price level has changed.
Michael Sandifer
Mar 2 2020 at 11:18pm
Scott,
This raises a lot of questions in my mind. First, you now favor discretion for Fed officials at times, at least for “real” recessions? Or, would you just have a more nuanced rule than NGDP level targeting?
Second, how do you target beyond this year, without trying to keep NGDP on trend in the meantime? You called for a significant cut in the Fed Funds rate target range recently. Yes, most of the effects will be felt in future years, but NGDP growth would be boosted this year too. Do you suggest that the Fed merely rely on forward guidance? If so, how would you deliver that forward guidance? Have you changed your mind on wanting an immediate rate cut?
Third, theoretically, how do you avoid more unemployment than is necessary if you allow NGDP to fall below trend this year? I understand that real GDP potential will temporarily shrink, and some job losses, or at least temporary layoffs and cuts in hours may be inevitable, but what of the “seondary depression” that Hayek wrote about? Why let labor compensation rise against NGDP, compounding the real problem?
What am I getting wrong here?
Nick
Mar 3 2020 at 5:06am
Ill have a stab at this, will be good to test my own understanding of Scott’s policy.
the policy goal is level targeting of NGDP, with an NGDP growth of circa 4%/year.
you change policy (changing expectations of current and future money supply) to keep market expectations of future NGDP on that theoretical line.
Hence you respond to shocks which the market interprets as deviating longer term growth from the line, and you don’t respond to shocks that don’t.
jj
Mar 3 2020 at 12:21pm
Nick, it looks to me like your understanding is correct but I’d suggest one correction to your wording, since it looks like you’re making a distinction in terminology between “policy” and “policy goal” which I’ve never seen done before. I believe the terms are used interchangeably.
I think Scott would say that you don’t change your policy of stable NGDP, rather you take the required actions to ensure you achieve your policy. The specific actions could be things like open market operations, helicopter drop, etc.
Scott Sumner
Mar 3 2020 at 2:04pm
Even after today, I still want an immediate rate cut at next week’s meeting, at least 50 basis points.
Michael Sandifer
Mar 3 2020 at 2:05am
Scott,
And as I think more about your post, I think it’s very inappropriate to compare a negative supply shock to seasonal changes in NGDP growth. Those seasonal changes are already baked in expectations-wise, unless they’re more or less extreme than expected. But, in that case then, you obviously have a supply-shock and monetary policy should respond accordingly.
I think it’s best to keep policy close to neutral, which is what NGDP level targeting does.
Scott Sumner
Mar 3 2020 at 2:02pm
No, I haven’t changed my views at all. I’ve always favored targeting expected NGDP one or two years in the future. The Fed cannot and should not try to control every little blip in NGDP.
Alan Goldhammer
Mar 3 2020 at 12:37pm
So the Fed does the rate cut this AM and the stock market seems not to care. I guess we now have a ‘real’ NGDP targeting experiment underway now.
Scott Sumner
Mar 3 2020 at 2:01pm
Alan, No, we don’t have an NGDP experiment underway, and the stock market rose sharply Friday on expectations of the rate cut. That’s how markets react. The Fed is not targeting NGDP.
Alan Goldhammer
Mar 3 2020 at 4:59pm
A rate cut does not help address global supply chain issues which I commented on to an earlier post. It’s likely my economics lack of understanding. Given we are almost at zero interest rates and there is little room to go unless the Fed wants to go negative, what is the solution. If we enter a COVID-19 based recession, are there any policy options for the Fed?
Over the weekend I listened to Barry Ritholtz interview Danielle DiMartino Booth who has been highly critical of the Fed. Podcast is HERE. Some of the parts of the discussion were wrong and Ritholz, who is no fool, pushed back. Krugman weighed in on this as well today in the NY Times, also wondering if the Fed is capable of doing much of anything. Tough questions and maybe not many good choices.
Scott Sumner
Mar 3 2020 at 6:11pm
(I meant Monday, not Friday, in my previous remark.)
There’s no limit to how much the Fed can boost NGDP, unless it runs out of ink and paper to print money.
You are right that the Fed should not try to address “supply chain issues”, they should try to keep monetary policy stable by adjusting their target interest rate when the equilibrium interest rate changes. The goal is stable growth in NGDP over a one or two year horizon.
Michael Sandifer
Mar 4 2020 at 12:23pm
Scott,
Okay, so your policy ideas haven’t changed, but don’t the rate cuts have effects that apply to the current year? Granted, any increased inflation might come with variable lags, but real GDP boosts should be quick.
Also, I don’t agree even with the idea that monetary policy shouldn’t adjust to accommodate current year seasonal variations in NGDP growth. I recall a presentation by George Selgin in which he showed data from periods of relatively free banking in which the money supply naturally expanded even on a daily granular basis in response to spikes in demand for money during harvest seasons, for example. I hope I can find that presentation again, and hopefully I’m not confusing money supply changes with other liquidity needs.
Scott Sumner
Mar 4 2020 at 1:06pm
Michael, You are mixing up two issues. Yes, the money supply should adjust to accommodate seasonal needs for currency. But no, the money supply should not adjust to prevent seasonal fluctuations in NGDP. That was not George’s argument.
I agree that a cut in rates is appropriate right now, and I agree that this would help output during this year. I don’t think we need to have enough monetary stimulus to prevent a fall in NGDP for a few months due to the effects of the virus.
Michael Sandfer
Mar 5 2020 at 11:50am
Yes, I Selgin was referring to currency in circulation, not NGDP or the monetary base.
Michael Sandifer
Mar 4 2020 at 12:31pm
Also, in my mind, I draw a sharp distinction between temporary reductions in RGDP potential and permanent ones. I don’t see anything wrong with more stimulus this year to help smooth out effects on employment, even if those effects would be minor otherwise, and even if the stimulus would be of relatively slight effectiveness.
This is very different than say, an asteroid hitting the western US that wipes out half of the country’s capital stock and labor supply, in which case NGDP level targeting would lead to a doubling of NGDP with no where near a doubling of real output versus the no stimulus counter-factual.
Michael Sandifer
Mar 4 2020 at 12:38pm
In my last comment above, I should have referred to a doubling of the price level, not NGDP.
Michael Sandifer
Mar 5 2020 at 2:44am
I looked it up and it was the currency in circulation that varied seasonally in the Canadian banking example Selgin cited. Changes in currency in circulation have almost no correlation with changes in NGDP, so I’m guessing that doesn’t matter.
Comments are closed.