Thursday, February 6, 2014

Phillips curve, RIP


From February 6 Wall Street Journal. So much macro discussion presumes that inflation is always and everywhere driven by booming economies, it's worth remembering rather striking contrary evidence.

I'm looking for a good graphic, but this last week's failure of emerging market central banks to put a dent in their currencies by raising interest rates is also an important reminder of the limits of monetary policy. As surely Argentina's and Venezuela's troubles are beyond anything a central bank can fix by any finite interest rate.

27 comments:

  1. I thought the Phillips Curve was dead after the Presidential Lecture at the 1968 American Economics Association annual meeting... At least it should have been.

    ReplyDelete
  2. Hyperinflation is a market revolt. Turkey raised their OVERNIGHT rate to 12% and the lira responded for the life span of a mayfly.

    “The crisis takes a much longer time coming than you think, and then it happens much faster than you would have thought.” -- Dornbusch

    ReplyDelete
  3. You cannot apply one-size-fits all model to emerging market (periphery) and industrialised countries (core). Periphery countries have to be concerned with exchange rate and capital flight instability and have much less control over their domestic economies with MP; industrialised countries, especially ones whose currencies underwrite the international monetary system have a lot less to worry about.

    I would suggest you read Marx, this gives a lot of good insight into this. I would also strongly suggest a review of interwar or late Bretton Woods exchange rate instability focussing on good historical analysis, rather than model.

    ReplyDelete
    Replies
    1. I really don't think Marx is important here.

      Delete
    2. This is a troll...

      Delete
    3. The core and periphery are Marxian concepts in political science and political economy, although Marx himself, as far as I know did not that terminology himself. It is basically consistent with his idea of permanent divergence, inequalities, disequilibrium and crisis which at a basic level is in contrast with neo-liberal views of equilibriating, win-win outcomes. The core/periphery concept is related to a (skewed) dependency relationship that exists between core countries and peripheral ones - again in contrast to neo-liberal arguments that stress interdependence between units (eg states) in an international system.

      The distinction between emerging markets and large industrialised countries is arguably also important when talking about the limitations of MP.

      Delete
  4. The Phillips Curve never really existed; it was an artifact of clever statistical manipulation. From McCullough, McGeary and Harrison (Canadian Journal of Economics, 2008, p. 1410, footnote 3):

    Phillips did not run a regression on his 53 observations for unemployment and inflation data.
    He binned the data and then fit a curve to the resulting averages within each bin. This is evident when one considers the ‘points’ that Phillips used to describe his ‘curve’; with U
    (unemployment) on the abscissa, the bins are defined by 0-2, 2-3, 3-4, 4-5, 5-7, and 7-11: the bins containing 6, 12, 5, 11, and 9 observations, respectively.

    Why such unequal class widths (with such differing numbers of observations in the classes)? Other choices for intervals did not support his theory. For example, the intervals 0-2, 2-3, 3-4, 4-6, 6-8, and 8-11 show a positive relation between wage inflation and unemployment for high unemployment.

    All this is well described in Wulwick (1996). What is also little known is that Phillips never intended this work as serious scholarship; it was just some casual work he did over a weekend at the suggestion of a colleague, and he was embarrassed by it. See the papers by Phillips in Leesom (2000) for further discussion of this point.

    ReplyDelete
  5. Surely you should use compare the cyclical component of both series rather than comparing in aggregate? I.e. estimate the spectra

    ReplyDelete
  6. There's probably some different expectations at work. Nobody thinks inflation in Venezuela and Argentina is about to slow down any time soon, since it would require some very painful money supply contraction and anything resembling austerity is unpopular after the dollar peg debacle. So you've got a situation where inflation is steadily high and increasing - I'm not surprised that real GDP is tanking. Any real growth (which did happen for a few years in the mid-2000s) is getting eaten up.

    That's different from a situation where you've got low single-digit inflation and there's no doubt that your central bank will throttle down the money supply if inflation starts to grow too large (as in the US).

    ReplyDelete
  7. "So much macro discussion presumes that inflation is always and everywhere driven by booming economies, it's worth remembering rather striking contrary evidence."

    Inflation is driven by declining productivity.

    ReplyDelete
    Replies
    1. How would you explain extreme deflation? For example as occurred during the interwar period?

      Delete
    2. High leverage is what causes extreme deflation.

      Delete
    3. "High leverage is what causes extreme deflation."

      Plausible. At a national level are you suggesting that it is deflation that arises from an excessive debt/GNP ratio (if an upper bound on national debt levels exists)?

      Delete
    4. Anonymous,

      "How would you explain extreme deflation? For example as occurred during the interwar period?"

      Increasing productivity. See:

      http://en.wikipedia.org/wiki/Great_Depression#Productivity_shock

      Productivity shock

      "It cannot be emphasized too strongly that the [productivity, output and employment] trends we are describing are long-time trends and were thoroughly evident prior to 1929. These trends are in nowise the result of the present depression, nor are they the result of the World War. On the contrary, the present depression is a collapse resulting from these long-term trends."

      —M. King Hubbert

      The first three decades of the 20th century saw economic output surge with electrification, mass production and motorized farm machinery, and because of the rapid growth in productivity there was a lot of excess production capacity and the work week was being reduced.

      The dramatic rise in productivity of major industries in the U. S. and the effects of productivity on output, wages and the work week are discussed by Spurgeon Bell in his book Productivity, Wages, and National Income (1940).

      Delete
    5. Efth,

      "High leverage is what causes extreme deflation."

      No. High leverage causes a legal bind when it is combined with extreme deflation. Debt contracts are protected by a legal system, but they are protected on a nominal basis. However, individuals / private enterprise faces real borrowing costs.

      High leverage in itself doesn't cause a problem.
      Extreme deflation in itself doesn't cause a problem.
      High leverage and extreme deflation creates a severe legal problem.

      There are a couple ways to deal with the legal problem of extreme deflation and high leverage:

      1. Limit leverage - that means either scrap the capital markets altogether (government issued money) or rely on more equity and less debt
      2. Accept higher inflation with higher leverage (at the expense of productivity)
      3. Change the legal system to protect real instead of nominal contracts

      Delete
  8. Professor Cochrane,

    You said: "So much macro discussion presumes that inflation is always and everywhere driven by booming economies, it's worth remembering rather striking contrary evidence."

    I wonder what macro discussion you are talking about. Most macro discussion presumes that inflation above an anchored value is always and everywhere driven by economies with excess 'demand'. This can be caused by higher than average NGDP growth, which might be associated with a booming economy, or caused by average NGDP growth in the face of shrinking RGDP growth, which would not be associated with a booming economy.

    ReplyDelete
  9. You should keep in mind there's substantial literature on unifying the wage and phillips curves. Looking cross-sectionally doesn't tell you anything about the with-in country phillips curve relationship.

    ReplyDelete
  10. Friends,

    In a recent event Tim Hartford spoke about Philip and the curve -

    http://www.cato.org/multimedia/events/undercover-economist-strikes-back-how-run-or-ruin-economy

    - Avi

    ReplyDelete
  11. This is a classic example of Austrian Business Cycle Theory. Argentina has increased the money supply 300% since 2008. Venezuela is even worse, increasing the supply of money by over 600% in the same time period. This of course, produced short-term gains, and had the Krugmans and Yglesias types dancing in the streets. But, as night must follow day, bust follows boom. Tim Hartford gives great empirical evidence for the ABCT in the paper listed above by Avi.

    ReplyDelete
  12. I have to say, I am disappointed by this post, Prof Cochrane. I don't think that any serious macroeconomist, of any school of thought, would ever believe that inflation is a a phenomenon that can only happen due to excess demand over supply, as you're suggesting above.

    The recent experience in Zimbabwe, not to mention the inflations of Latin America of the 80s and 90s put rest any such notions and I've not seen anyone claiming otherwise, be it in academic papers or in blogs, anywhere in the last 20 years, at least.

    In that sense, I'd be most interested in seeing who has claimed otherwise and in what context.

    Of course, the distance between these two examples, Venezuela and Argentina, and most other economies is vast and the level of economic mismanagement that they've experienced, on so many levels, is so abhorrent that I don't see them as good examples of what might happen in other economies, particularly more developed ones such as the US. It would require changes in economic policy on unprecedented scale and I don't see there being any significant risk of this happening in the EU, USA or even the vast majority of developing countries; for the most part, they've learned their lesson.

    I do like your final point though and I think these are clear, if extreme examples, of the limits of monetary policy.

    ReplyDelete
    Replies
    1. "that inflation is a a phenomenon that can only happen due to excess demand over supply, as you're suggesting above."

      I'm trying to find where he said that.

      Inflation happens because of expectation of inflation.

      Delete
    2. "So much macro discussion presumes that inflation is always and everywhere driven by booming economies, it's worth remembering rather striking contrary evidence."

      That's my inference of what he meant by the above phrase, but perhaps he meant something else?

      Delete
    3. I think his statement makes sense. He is too smart to be wrong about this. Actually he said the opposite of what you said he did imo.

      Delete
    4. "Inflation happens because of expectation of inflation."

      This is a very unhelpful remark. Are you telling me that the Arabs raised the price of an essential commodity in the 1970s because they expected inflation?

      And when your remark is true, it is important to ask what sets off inflationary expectations. For example it could be political instability, the breakdown of supply chains...

      Delete
    5. "Are you telling me that the Arabs raised the price of an essential commodity in the 1970s because they expected inflation"

      No, but when the price was raised, inflation increased because of inflation expectations. Why do you think inflation did not occur when the price of oil in 2007 rose higher than in the 1970s? It did not increase because inflation expectations because of the oil price increase were low. These inflation expectations were dampened by conviction from the government to flood the market with reserves. Inflation is outcome of mass psychology. No expectations of higher prices, no inflation.

      Delete
    6. "No, but when the price was raised, inflation increased because of inflation expectations."

      I think the causation is not that clear. People's expectations of higher prices followed actual higher prices in this case. So in this case it followed the real event. A few countries had a monopoly over production in the 1970s. It took a while to get alternative petroleum sources into operation and make them feasible. This changed with the fall of the Eastern Bloc in the 1990s and new oil supply and the development of new production technologies since the 1970s. There was an important weakening of the OPEC cartel and oligopolistic price setting and therefore its overall effect. You can also add the government commitment in there - for one thing it had the reserves!

      Delete
  13. anti inflation: hold down interest but forget rise in consumption over norm, Increased money supply ineffective that could be said of the US. Were's the monkey?

    ReplyDelete

Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.