Larry Summers: ‘I think Keynes mistitled his book’
Last week, my column was about the lessons that the Obama administration’s “all-star Keynesians” learned when they unexpectedly had to face down the exact sort of crisis John Maynard Keynes and his successors had described. My conversation with Larry Summers, however, included much more of interest than I was able to include in the article. So I asked Summers for permission to transcribe and edit our interview and post it in full. Here it is:
Ezra Klein: You were in the uncommon position of actually having to deal with the exact sort of generational economic crisis that Keynes described in his writings. What did you learn?
Larry Summers: I think Keynes mistitled his book. The correct title was “A Specific Theory of Collapsing Employment, Interest and Money.” What his book really was about was the proper understanding of the convulsive downturns to which free market economies are intermittently prone. And in that, Keynesians have been vindicated. The core Keynesian lesson, which is that most of the time economies are self stabilizing, but two or three times a century, they become self-destabilizing, where a worse financial system creates a worse economy creates a worse financial system, has been central to this episode.
A second broad Keynesian lesson, which was a deep Keynes conviction, is that macroeconomics policy is about filling in the gaps, not smoothing the cycle. Prevailing economic theories employed in the the ’80s and ’90s held that policy could aspire to smoothing out the cycle, but for every trough in output that was avoided you would shave off a peak level of output. The belief was that macroeconomics was about reducing the variability of output over time, not raising its average level. Keynes on the other hand focused on raising the average level of output through time by raising demand. Our current predicament makes the relevance of his ideas apparent. No one supposes that the output losses suffered by the American economy over the last five years will ever be made up.
A final lesson is the overwhelming importance of demand in major business cycle fluctuations. Unemployment rates are at super-high levels for all industries and demographic groups, while at the same time, utilization in factories is low and houses and commercial office buildings are sitting empty. Employers report low vacancy rates by historical standards and minimal efforts at recruiting. That tells you this is about demand, not some kind of structural factor in which there are mismatches between the kinds of workers available and the kind employers are seeking.
EK: But whether Keynesians were right, it doesn’t look to me like they’ve been vindicated. It looks to me like Keynesian is under attack.
LS: Yes, and the attacks echo things that were said in the late 1930s before military mobilization lifted the American economy out depression. There are two ways to understand the backing away from Keynesian ideas. In the first speech I gave and almost every subsequent speech I gave, I included the sentence that “the central irony of financial crises is that they’re caused by too much borrowing, too much confidence and too much spending and they’re solved by more confidence, more borrowing and more spending.” That is profoundly counterintuitive. It makes it difficult to persuade people of the need for more fiscal policy. That’s one element.
The second element is that people see economic issues through moral frames and people think there’s an extent to which recessions are punishment for sins — mainly sins of excess — and you don’t expiate sins by binges. So there’s a kind of moral counterintuitiveness that has made it difficult for the public and for political figures to accept stimulus. You can see this in the past, too. John Kennedy’s economic advisers told him after his 1961 inauguration that the country needed a tax cut and he told them they may be right, but he had just given an inaugural address about bearing any burden and paying any price and so he explained to them that he could not turn around and propose a tax cut.
EK: But ultimately, I guess, he did. Which brings us to the policy response. Even if it’s not going to be particularly popular, what have we learned about stopping these crises?
LS: It’s not so much what we have learned as what we have recalled. We have been reminded that, as former Mexican President Ernesto Zedillo said, “Markets overreact so policy must overreact” Because crises have elements of a vicious cycle. It is important that policy responses be prompt and strong so as to turn vicious cycles into virtuous circles.
We have also been reminded that stabilizing the economy cannot be entirely given over to monetary policy because you might find yourself in a liquidity trap where you can’t lower interest rates much. Therefore there has been a substantial renewal of interest in discretionary fiscal policy. Going forward, I think you’ll see more attention given to the automatic-stabilizer element of fiscal policy, particularly if we stay in a low-inflation society where interest rates are coming close to hitting the zero bound.
The other part of fiscal policy is that we could do much more to have plans for spending quickly if the occasion required it. No one has a strong incentives to really have a shovel-ready project. Why would you get the final zoning approval to do an expansion of your house unless you had the money for it? But one could imagine much more contingent planning across the government to be in a position to spend on infrastructure and maintenance and repair quickly.
EK: What about ideas to simply speed up the regulatory process during downturns?
LS: This could be very important. Harvard built a whole football stadium in the 1890s with less than two years from the idea’s conception to the stadium opening. That kind of thing does not happen today. My attitude is we probably slow things down unnecessarily with our various approval processes. So I’m sympathetic speeding the regulatory process during recessions. But I have to be honest and say my enthusiasm for reducing regulatory processes in recessions is in significant part a reflection of a more general concern about regulatory delays.
EK: And even if we did get that done, I remember there was an ongoing concern that building things, though productive, isn’t always the quickest way to get money into people’s hands.
LS: One other thing that slowed it down was there was a general political urge to do things that were substantial legacy of achievements rather than just things that were reasonably productive infusions into the economy. Do you want to start on a high-speed rail network in America or do you want to fix bridges and schools that are crumbling? You can start maintenance very quickly, but it doesn’t leave behind anything you can name or tell your children about.
One of the dynamics of the discussion over the stimulus was whether the idea was to spend money in not-inefficient ways in as large and rapid a scale as possible, as the macroeconomists wanted, or was the objective to take advantage of this moment to redeem the progressive hour, or was this the hour apart from it being a progressive hour to do things that would constitute substantial renewal of America? Those were the three values being compromised around. You could get more spending more quickly if you stayed focused on number one.
EK: How about the other side of countercyclical policy? It seems to me that one reason that the politics of deficit-spending are tough in recessions is that we don’t always make a show of fiscal discipline during expansions, so it’s not necessarily credible to go to the people and say we need to just binge for a little while. They think we’re bingeing all the time.
LS: Every time I talked about this in the 1990s, I used the phrase “reload the fiscal cannon.” That’s my version of fix the roof while the sun is shining. You need to get yourself into a position where there’s room for a countercyclical response to economic downturn. That was the tragedy of the 2001-2008 period. Having put the Federal government in a strong position, we dissipated that position even in years when the economy was growing reasonably.
So I think more prudence in the good times, more automatic stabilizers, more contingent capacity to spend or cut taxes quickly. Another whole set of ideas that we looked briefly at in the context of the stimulus was giving people some form of debit card. If you want to give relief to people that maximizes their capacity to spend, give them a debit card that expires in six months.
EK: That seems like a good idea to me. Why didn’t we do it?
LS: When I first heard that idea, I loved it. Then I thought about it more and decided I only liked it. It occurs to you that if you give Ezra a $500 card he’ll spend that, but on things he would have bought otherwise and he can save the money he’d have otherwise used for those purchases, so there’s little net impact. In the event in 2009, the question was moot because the government did not have the capacity to implement such a plan quickly. I hope standby planning with respect to such ideas is underway.
EK: Overall, could we have done more?
LS: There are multiple aspects of this question. First, the administration proposed considerably more than Congress was willing to enact. Taking account of the addition of the AMT, the Administration’s Recovery Act proposals were cut back by about 20% in the process of passing Congress by very narrow margins.
Second, the President’s economic team advised that there was essentially no danger of excessive fiscal stimulus in 2009. I joked at the time that worrying about overdoing fiscal policy was like my losing too much weight and becoming anorexic — a conceivable possibility, but very far from the dominant risk. The President’s political advisers — rightly in my view — constrained the initial stimulus proposals to avoid sticker shock and rejection or great delay on Congress’s part.
Third, politics aside there were difficulties in moving spending rapidly in 2009. So-called shovel-ready projects often were not in fact ready to go. Almost everyone close to the process feels that Joe Biden and his team did a very good job of moving the stimulus money through the system and, as a consequence, money moved more or less on the schedule we projected in 2009. They would be the first to say that it would not have been possible to move vastly more money into quick trigger infrastructure projects. Of course it would have been possible to increase the tax cuts or assistance to state and local governments, but there were severe political limits in both those areas.
Fourth, we believed in the winter of 2009 that if, as seemed likely, more stimulus would ultimately be required, it could be passed through the Congress using the unemployment insurance extension for 2010 as a vehicle. This view proved incorrect. The administration proposed and the House passed in the fall of 2009 a substantial further program with respect to unemployment insurance, job incentives and infrastructure. Unfortunately it did not pass the Senate and the focus has shifted very much towards deficit reduction rather than job creation. It is fortunate that the President was able last fall to lead an effort to pair extended tax cuts with payroll tax reductions — without that stimulus we might be looking at a double dip today.