Larry Summers has responded to his critics. My first thought is, yes he is still really smart.
Also Olivier Blanchard also thinks the Covid Relief bill is too generous, and, in particular objects to the $1400 for most Americans. Note that Blanchard is the author of the main citation in my critique of Summers (although I should also have cited Summers on secular stagnation, Delong and Summers on the Keynesian Laffer Curve and Blanchard and Summers on hysteresis).
Summers makes 4 arguments.
Don’t blame me for the fact that the ARRA was too small. Here I think it doesn’t really matter, and that he is mostly right. Certainly he argued frequently for higher deficits.
Fearing Secular stagnation does not imply that one prefers $1.9 Trillion in stimulus to $ 1 Trillion. Indeed, the debate over the right number is almost completely qualitative.
The Fed has trouble fighting inflation without causing recessions. ” Every past significant inflation acceleration has been quickly followed by recession. Tamping down inflation will require allowing unemployment to rise, and engineering a soft landing is difficult:Unemployment has never risen by half a percentage point without then rising by almost two points, or more.”
This argument seems to me to beg the question. Summers assumes that the Fed can’t prevent an increase in inflation or moderate an increase in unemployment. If the Fed has prevented an accelaration of inflation, the success would not show up in Summers’ survey of what happens after inflation accelerates. If the Fed keeps unemployment at the NAIRU then unemployment does not have to increase to fight inflation. I think his point is that, when the Fed loses control and misses its target, then the Fed loses control and misses its target. There is no argument that fiscal stimulus (of any amount ever) makes it more difficult for the Fed to avoid both accelerating inflation and recession. The only clearly demonstrated limit on the Fed’s ability to thread the needle is the zero lower bound. That problem can be avoided with fiscal stimulus.
Also I note that qualifiers “significant” and “quickly” are doing a whole lot of work. The steady increase in inflation from 1964 through 1970 was followed by a recession in 1970 — the longest post world war II expansion which occurred before Summers began working for the Federal Government was mostly a period of accelerating inflation. Also the period of increasing inflation which began in 1976 was followed by a recession in 1980. Here I think the point is that inflation drops during recessions, so one sees peaks when recessions start. That just means that there is a Phillips curve. Mostly, the argument is that the Fed used to mess up more than 40 years ago, so the past 40 years of preventing acceleration of inflation are irrelevant.
The fact that many economists argued that the economy was overheating in the late 90s and just before Covid hit and that inflation would accelerate and then it didn’t seems relevant. Finally, I have never understood why inflation is considered such a horrible thing — I didn’t understand all the fuss in the 1970s and I still don’t. I do know that people hate inflation *and* assume it means increased prices for given wages. Economists note that wages and prices both increase, but also accept the conclusion that inflation is horrible.
But mostly, there is no argument for why the Covid relief bill will make the Fed’s task more difficult.
Summers 4th argument comes in two steps. He responds to the argument that the bill provides “relied to those who need help”.
Much of his argument is based on the assumption that things were fine in 2019 consumer spending by low-income consumers is up“, “cash balances have risen“, “see their incomes rise“. But Summers does not believe that things were fine in 2019 — he believes that income inequality was a huge problem in 2019.
His concluding sentence specifically criticizes the $1400. This, I think, is his arguably valid point. He can think of better uses of the money. However, policy depends on politics, and that exceedingly popular provision will be part of any bill. I guess he contributes to the debate over the income level at which the benefit is to be phased out, but this becomes an argument about a moderate change to about one fourth of total spending mere tens of billions not real money.
I will discuss Blanchard after the jump.
My first thought is that Twitter is evil (but at least Blanchard writes concisely and clearly). I will comment on this thread Blanchard estimates the current output gap and counts up increased spending from wealth accumulated during 2020 and from the December bill and the proposed 1.9 trillion. I do not entirely agree.
He has a generous estimate of the output gap of $ 900 billion. It does seem generous. They key assumption is that back when the unemployment rare was 3.5% the output gap was zero. This seems reasonable.
He assumes people will spend in 2021 half of the wealth they accumulated in 2020 because spending was dangerous for their health. Half is far above any estimate of the marginal effect of increased wealth on spending. The argument must be that the recently accumulated wealth is liquid and people’s appetites to consume are unsatiated. I certainly wouldn’t count on that demand.
He treats the total $ 2.8 T as an increase in demand and then says he is being conservative to use a multiplier of 1. He has presented evidence that the multiplier on Government consumption and investment is 1.5. However, he assumes that a temporary tax cut (or cash transfer) has a similar multiplier — in effect that none of the $2000 will be saved. He also assumes that all paycheck protection loans will be forgiven, so they amount to transfers. Here I see no willingness of macroeconomists to compromise between Ricardian equivalence (the $2000 will have zero effect on aggregate demand) and assuming that the propensity to consume is 1. I think the multiplier is less than 1 (and I basically take my estimates of multipliers from Blanchard and Leigh and ignore everyone else).
There should also be a calculation of the change in spending without stimulus. Here I think the output gap would probably shrink just due to vaccination. This means the $900 billion guess is very generous. However, one important component of demand is spending by state and local governments. Here I think the large provision of the bill is needed to keep it from decreasing. They will have spent their rainy day funds. They do run unexpected deficits when revenue is surprisingly low (as it was in 2020) and spending automatically increases (as in 2020) but they are required to try to avoid deficits when budgeting for the upcoming year. The experience of the 20-teens was of a gradual decline in government employment. Much of the currently debated bill is designed to prevent that from happening again. It is not adding stimulus it is preventing austerity.
Blanchard’s conclusion “Why go there? Why force the Fed to in effect cancel some of the Biden package?” The answer (following Blanchard 2019) is that the Treasury is paying absurdly low interest rates — people are willing to pay the Treasury to keep their real wealth safe. This is a gigantic opportunity. It is a reason to seek a higher public debt. If there were no Covid crisis, I would advocate giving $2000 to every American citizen (better if directed at the non rich — way better but impossible to give it to non citizens). In fact, I did. Blanchard himself presented arguments that support that policy (and said he didn’t mean to propose it but didn’t say why).
Finally Blanchard (who self identified as a socialist back long ago when I assisted his research) argues that he is the really leftist one with his policy proposal
Why couldn’t we finance it partly by an increase in taxes, say an exceptional tax on capital gains, given that the stock market has done so well in the recent past? This would be fair, deliver protection and limit overheating. Wouldn’t this be a better way?
Fine by me, but not really on topic. I don’t think an exceptional tax on capital gains would do much to limit overheating. The question is what is the average marginal propensity to consume weighted by capital gains realized. My guess is that it is very small — people with large capital gains have more money than they have time to spend. The proposal would imply a much smaller budget deficit, but I don’t think it would do much at all to limit overheating.