Thursday, January 1, 2015

Historical Fiction

Paul Krugman has an interesting perspective on the history of economic thought. According to him, the Volcker disinflation played out exactly as Keynesians thought it would. Cutting to the chase:
So were Keynesian economists feeling amazed and dismayed by the events of the 1980s? On the contrary, they were feeling pretty smug: disinflation had played out exactly the way the models in their textbooks said it should.
That's not the way I remember it. But maybe my memory is bad, so I thought I would do some research to check out these claims. We'll set the wayback machine for 1978, when Arthur Okun wrote "Efficient Disinflationary Policies." Okun set out to obtain a good measure of the costs of disinflation, and consulted what he thought were some of the best minds in the profession at the time:
You should recognize at least some of the names of these estimators-of-Phillips-curves. Some of them are still big shots. What was the conclusion from that work?
So, if we look only at the 1981-82 recession which, according to the NBER, is defined by a peak in 1981Q3 and a trough in 1982Q4, the drop in the quarterly PCE deflator was about 3.6%, or an average of 2.9% per year. By Okun's estimate, this should have resulted in a 29% drop in GDP per year. What actually occurred was a drop in real GDP during the recession of about 2.5%, or an average of about 2% per year. If Okun was feeling pretty smug [infeasible actually (see the first comment below), but imagine he were still alive then] about all this by the mid-1980s, I'm not sure why, given that he was off by an order of magnitude.

There's more to Okun's "Efficient Disinflationary Policies." After coming to the conclusion that using conventional policy to address the inflation problem would be too costly, he suggested how we could do that in a more efficient manner, by using "the direct influence of public policy on costs." What he appeared to have in mind were subsidies to firms - subsidies lower costs, which lower prices, so there would be less inflation, he reasoned. Obviously Okun's idea was not highly influential. Around the same time (this is from 1977), you would find James Tobin in roughly the same ballpark. Like Okun, Tobin ponders the costs of disinflation, and judges that they are just too high if conventional means are used. He says, "the way out, the only way out, is incomes policy," and concludes:
For the uninitiated, "incomes policy" means wage and price controls. Tobin is saying that the cost of using wage and price controls to control inflation is small, with the welfare distortions measured by "Harberger triangles," but the cost of using conventional policy to control inflation is a potentially very large output gap. As with Okun's policy recommendation, Tobin's went by the wayside.

So, now we know how some of the established macroeconomic big shots of 1978 were thinking. What about the upstarts? In early 1981 (the dates on the working papers are May 1981) Tom Sargent wrote a couple of papers which address directly the costs of disinflation, and how we should think about the policy problem. These are "The Ends of Four Big Inflations," and "Stopping Moderate Inflations: The Methods of Poincare and Thatcher." In "Four Big Inflations," he tells us about the consensus view of deflation at the time - i.e. the view of Okun and Tobin as outlined above:
Further, he supports that in a footnote:
Then, Sargent describes an alternative view:
While much of Okun's and Tobin's papers referenced above read like a foreign language now, I think most people will find that paragraph of Sargent's very familiar. Today, I don't think many people would find it objectionable, but at the time this sort of thinking was getting serious pushback from some big shots. Sargent quotes one of them in a footnote:
In contrast to Okun, who is quite pleased to supply an estimate of the cost of disinflation, in terms of lost GDP, Sargent doesn't think he has a serious model that he can take off the shelf to address the quantitative implications of disinflation. But, his two papers provide nice historical examples of disinflations (in some cases the ends of hyperinflations) that are quick, and that seem to conform to his ideas.

So, who were the winners and losers from this episode? Probably that's the wrong question to be asking. When science progresses, we all win. Science does not progress when individual scientists see it as a loss if their ideas are superseded, and try to prevent that loss from happening by denigrating new ideas. Careful re-examination of the Volcker disinflation might lead to the conclusion that Volcker should have done something different, but I think the consensus view among economists is that the Volcker disinflation was necessary, and whatever output was lost in the process was outweighed by the benefits of low inflation that we have enjoyed for the last 30 years. The idea that we should control inflation through tax/subsidy policy or wage/price controls gets no traction in the 21st century. It's now widely accepted that inflation control is the province of central banks, though we owe that principally to the Old Monetarists rather than the 1970s macro revolutionaries. But those revolutionaries gave us much of the framework we now use for addressing monetary policy problems: (i) we want to think in terms of policy regimes rather than policy actions; (ii) commitment is important; (iii) well-understood policy rules are important; (iv) people are forward-looking. All of those ideas are now common currency in central banking circles. People who succeeded as macroeconomists in the post-1980 world absorbed the ideas of Sargent and his contemporaries and worked with them - you can certainly see that in Woodford's work, for example.

Are these long-gone controversies which have no relevance for current policy issues? Certainly not. In the messy world of monetary policy-making, we can still find people who want to think about monetary policy in terms of actions rather than state-contingent policy rules, or who want to base policy decisions on things akin to an estimate of the slope of the Phillips curve. People with the resolve of Paul Volcker are unusual, and institutional commitment can be difficult when there are many people with disparate ideas weighing in on a policy decision.

One way to think about the history of macroeconomic thought is as a series of battles. These people thought x, those people thought y. There was a fight between x and y, and y won. Then z came along, tried to beat up y, but y kicked z's butt, etc. That may sound very exciting - people find stories about personal animosity intriguing. But the reality is perhaps not so exciting. We have a set of ideas that we currently find useful, some of which we can trace directly to x, y, and z. Other ideas have morphed in various ways through the work of many people, and through public discussion, so that it's hard to give credit to some unique originator. And we're all "rational expectationists" now. Perhaps it's best to remember Samuelson for the positive things he gave us - the Foundations of Economic Analysis, the overlapping generations model, for example - rather than his carping about "rational expectationists." As well, if you want to know what people really think, look at what they do, not what they say. For example, when Paul Krugman wants a
... core insight that changed ...[his] ... mind about monetary policy in a liquidity trap (and is useful for fiscal policy too),
what does he do? He takes a Lucas cash-in-advance model off the shelf - a model with forward-looking optimizing economic agents having rational expectations - and uses it to learn something. He's a rational expectationist too!

21 comments:

  1. Re " If Okun was feeling pretty smug about all this by the mid-1980s," Okun died in March 1980 (http://en.wikipedia.org/wiki/Arthur_Melvin_Okun) so he didn't have a chance to check his forecast against the record or make a mea culpa.

    ReplyDelete
  2. Sargent has a number of more recent papers that explore this period. Here are a couple examples, possibly not the best ones:

    https://files.nyu.edu/ts43/public/research/inflat21.pdf
    https://files.nyu.edu/ts43/public/research/ccs13.pdf

    ReplyDelete
    Replies
    1. By the way, Dave's an excellent example of the macroeconomists who lived through the transition to modern macro, and succeeded by absorbing new ideas and building on them, which wasn't always easy. Unless you were at Minnesota, Chicago, or Carnegie-Mellon, you had to do this indirectly, and work a lot of stuff out for yourself.

      Delete
  3. Good post! But even if Okun's numbers were way off, shouldn't the early 80s recessions teach us that actions have some importance, and that disinflations are not costless in terms of output?

    ReplyDelete
    Replies
    1. But doing so certainly wouldn't make us Keynesians. The issue is if all actions or only unanticipated ones have importance.

      Delete
    2. "...actions have some importance..."

      But it looks like a regime change, don't you think?

      "...disinflations are not costless in terms of output..."

      I tend to agree, but it's actually not obvious what you attribute that recession to. And if it's monetary policy, the propagation may not work through sticky wages and prices.

      Delete
    3. Agree all round.

      Actually I suspect that Fed actions have real effects because they are signals that a regime change is real. That was the implication of a decomposition I saw by a Bank of Canada guy in 2013 showing that the rounds of asset purchases in QE-infinity each caused immediate jumps in rates that decreased as the policy went on.

      But why are the signals costly? For that, we have to know the propagation mechanism - whether it's sticky wages/prices, or something else.

      Delete
  4. "But those revolutionaries gave us much of the framework we now use for addressing monetary policy problems: (i) we want to think in terms of policy regimes rather than policy actions; (ii) commitment is important; (iii) well-understood policy rules are important; (iv) people are forward-looking. All of those ideas are now common currency in central banking circles. People who succeeded as macroeconomists in the post-1980 world absorbed the ideas of Sargent and his contemporaries and worked with them - you can certainly see that in Woodford's work, for example."

    In many ways this is a continuation of the previous discussion, and the reply to your previous post “Where is the multiplier?” by Anonymous (Jan 1, 2015 at 10.11 AM). Most of these things you have identified have been debated for the last 200 years, they were key arguments put forward for maintaining the Gold Standard, and are unfortunately likely to be debated for the next 200. Even though they have been unreservedly debunked by historians time and time again, people keep bringing these old arguments back up again. Sargent did not give us anything of substance new; his final sentence on the extract you provide remains highly controversial and is just a repeat of those old arguments. But what he did was dress it up in maths and some implausible assumptions that are consistent with this ideology. (The forward looking arguments are weak and largely irrelevant. When I look at applied mathematicians like Sargent it seems to me the support what-ever ideology gives a more “sophisticated” mathematical model, which is even worse than choosing a model on the basis that it supports an ideology!!! This is probably explains Krugman’s observation (if he is correct) that some of these die hard monetarists were later miraculously transformed into RBC card carriers.) And this is reflected in their approach to history. Samuelson, who ironically inadvertently paved the way for this indulgent nonsense by gutting the General Theory and forcing it into a mathematical Walrasian model, however, is spot on in the quote you give regarding the historical evidence. When the dust settles, surely the overwhelming majority of people who will not put the stabilisation of inflation from the 1980s down to the Volcker shock. The lasting stabilisation is more likely to be put down to being a result of factors that were occurring in the world economy at the time (relating to oil and commodity prices, and then later the entry of China and the ex Soviet bloc into the international system which altered the relationship between capital and labour and the prices of manufactures and commodities throughout the world and which lasted for decades). Endorsing the policy measures taken during this episode would be like trying to endorse pre-General Theory policies taken during the Gold Standard which were all about policy credibility and commitment. Like Samuelson, most people now, both policy economists who have real economies to run, and certainly historians, do not.

    ReplyDelete
    Replies
    1. Be quiet John, your nonsense is not needed here. Krugman will supply enough nonsense for all of us.

      Delete
  5. “It would be a grave mistake for the Fed to commit to conduct monetary policy according to a mathematical rule,”

    Good stuff.

    http://blogs.wsj.com/economics/2015/01/04/minneapolis-feds-kocherlakota-opposes-policy-rule-to-decide-u-s-monetary-policy/?mod=marketbeat

    ReplyDelete
    Replies
    1. I'm guessing that you think that's inconsistent with what Sargent was saying in 1981. That's not correct.

      Delete
  6. Suppose someone held the following view in 1980 or so: "Inflation only *seems* to have a momentum of its own; it is actually the long-term government policy of running persistently high deficits and creating money at high rates which imparts momentum to inflation." What would such an economist think of the U.S. economic record in the last decade or so, with its "persistently high deficits," "high rates" of money creation, but not much inflation?

    ReplyDelete
    Replies
    1. They might say: "That's pretty interesting. Once we thought the trap central banks tended to fall into was in creating too much inflation by persistently exploiting some temporary (and perhaps only perceived) tradeoff between inflation and unemployment. Now the trap seems to be that central banks create too little inflation under the false impression that persistently low nominal interest rates will create inflation."

      Delete
  7. Lot of nonsense in this Sargent paper. "Persistently running large deficits" does not automatically lead to higher inflation, it depends on how the monetary agency behaves. And Krugman is spot-on about the calibration part, the New Classical are unscientific as they ignore empirical evidence.

    ReplyDelete
    Replies
    1. "...the New Classical are unscientific as they ignore empirical evidence."

      Who are those New Classical anyway?

      Delete
    2. Guys who do e.g. predict hyperinflation and don't acknowledge their mistake after their model failed empirically.

      Delete
  8. Those two papers you quote by Tom Sargent were simply brilliant and yet live in obscurity in the working papers series of the Federal Nank of Minneapolis.

    They have much to teach today's microeconomics despite been written in 1981.

    It is also an excellent analysis of the contemporary policies of Margaret Thatcher.

    ReplyDelete
  9. "It is also an excellent analysis of the contemporary policies of Margaret Thatcher."

    Lol. Instead of reading your "holy books" you Austrian nutcases really gotta look at some DATA from time to time to reconnect with reality.

    ReplyDelete
  10. Okun said a drop of "10% of a year's GNP", not 10% per year, as Williamson claims. That is, Okun's 10% of GDP (the "sacrifice ratio") could be spread out over many years.

    In hindsight, we know that Okun was pretty accurate. For example, Ball (NBER #4306, 1993) found that the Volcker disinflation had a sacrifice ratio of about 7%, fairly close to Okun's estimate. So the episode cost the US over 50% of a year's GDP.

    ReplyDelete
  11. If you want to see how all those outdated ideas about inflation fail just look at Argentina, we have it all, price controls, wage controls, subsidies (for all...costs and utilities)...please send us a couple of Sargent's books for our ministrer of economics....I expect he might find them useful.

    ReplyDelete