Saturday, December 20, 2014

Deflation links

Commenter Zack sent the following Paul Krugman links and quotes, which deserve promotion from the comments section.

"But deflation is a huge risk — and getting out of a deflationary trap is very, very hard. We truly are flirting with disaster."
http://krugman.blogs.nytimes.com/2009/02/04/about-that-deflation-risk/

"So we're really heading into Japanese-style deflation territory"
 http://krugman.blogs.nytimes.com/2009/07/02/smells-like-deflation/

 "So tell me why we aren’t looking at a very large risk of getting into a deflationary trap, in which falling prices make consumers and businesses even less willing to spend." http://krugman.blogs.nytimes.com/2009/01/10/risks-of-deflation-wonkish-but-important/

 "But the risk that America will turn into Japan — that we’ll face years of deflation and stagnation — seems, if anything, to be rising."
http://www.nytimes.com/2009/05/04/opinion/04krugman.html

"What I take from this is that deflation isn’t some distant possibility — it’s already here by some measures, not far off by others."
http://krugman.blogs.nytimes.com/2010/07/11/trending-toward-deflation/

"Worst of all is the possibility that the economy will, as it did in the ’30s, end up stuck in a prolonged deflationary trap."
http://www.nytimes.com/2009/02/06/opinion/06krugman.html?partner=permalink&exprod=permalink&_r=0

As we know, it didn't turn out that way. We have had positive inflation for 6 years.

Why does this matter? Normally, it doesn't and it shouldn't.

To repeat points made earlier, economics should be science, not witchraft. We do not say "the witch doctor said it would rain, and it did!," and follow him for a while. At a minimum, we measure a forecaster's ability by collecting all his or her forecasts, not just the good ones -- or the bad ones. More deeply, personal prognostication is a nearly useless test of economic models. Prognostication mixes judgement, opinion, forecasts of what politicians will do and what shocks will hit the economy, along with economic  models, in ways that tell you little about the models. If a climate scientist tells you he thinks it will rain this weekend and it's sunny instead, we do not say "well, climate science is bunk." You can only "test" a model once it is written down in a way that anyone operating the model can agree what its prediction is.  Finally, there are a lot of other shocks hitting the economy; a forecaster that was right 60% of the time would be a genius in this business, so one blown forecast is meaningless. If anyone else had written these, they could reasonable respond "it's still a danger, we only avoided it by the Fed's QE and huge deficits."

But I don't write endless posts excoriating "inflationistas" for the lack of their largely mischaracterized inflation forecasts, crowing about how I'm always right about everything,  damning others for failing to learn from evidence, and Bulverizing (look it up, here too, a great word) about their evil motives. So a few look-in-the-mirror-why-don't-you quotes are appropriate.  I've been too lazy to look up these quotes, so I thank Zack for doing it.

This also will matter Monday -- I have a piece coming out that mentions the failure of the widespread "deflationary spiral" forecast. These quotes offer a nice documentation.

By the way, yes, at the time I warned of the risk of inflation, and that didn't happen either. I was quite clear it was a risk not a forecast. California has a risk of earthquakes. And the failure to see one in six years does not prove geologists are all mendacious morons. There are precedents for the inflation risk. Reinhart and Rogoff pointed to quite a few cases in which after a "quiet period," banking crises are followed by sovereign debt crises or defaults.

As I see it, that risk remains, though it has declined a lot. The reason it declined is that our government, and the European governments, kept their eyes on long-term budget issues. Our Administration has from the beginning always promised long-term budget repair, despite Keynesian theory that says if you want to stimulate, you keep quiet about future taxes or spending reductions. No point in waking up the Ricardian genie. You might complain the Administration wasn't serious enough, but they were always saying there would be a long term plan, and bond markets evidently bought it. Many in Congress too have had their eyes on long-term budgets, and long-term fiscal solvency depends if anything more on Congress than on this Administration.

The Europeans have gone through several rounds of "austerity," despite Keynesian and especially Krugmanian excoriation. (True, they started with counterproductive high-marginal-tax austerity, but Europe seems quickly to have learned that less spending and structural reform are a better path.) They came darn close. If Italy and Spain had defaulted, we would likely be having a different conversation today.

Doing so, our and Europe's governments persuaded bond markets that the currency and debt are safe -- maybe even too safe - -and avoided, for now, the sovereign debt fate Reinhart and Rogoff warned about.  If I erred in overestimating the inflationary risk, I erred in underestimating the fiscal sobriety of all our governments, and I overestimated the extent to which they believed the Keynesian advice to ignore, default, devalue, or inflate away debt. That's why no earthquake in 6 years hasn't changed all that much my views on the "model," in this case of underlying causes of inflation.

On the "spiral" or "vortex." Perhaps there are a few true-blue Keynesians reading who can help me out. I understand the idea that deflation leads to high real rates leads to lower demand leads to lower output leads to more deflation. ("Understand" ≠ "Agree".) I don't see how the model ever predicts this to end. Is there some clear "and it bottoms out when x y z?" In my model, it bottoms out when you hit the top of the present value Laffer curve, and future taxes cannot hope to pay back the deflationary increase in the real value of the debt.  But I don't know where it ends in the standard old-Keynesian model. Positive eigenvalues are positive eigenvalues. Maybe some wealth effect of government bonds (Another way to put "my model")? But why doesn't that stop the spiral in the first place? Is it right to characterize the model's prediction as an endless spiral to zero? If not why not?

Update: Commenter JZ sends the following link, and it's only fair to include it.

" ... back when the crisis started, I did expect to see deflation, Japanese style, if it went on for an extended period. I was wrong ... "
http://krugman.blogs.nytimes.com/2013/03/05/why-dont-we-have-deflation

24 comments:




  1. Zack's diligence in retrieving theses citations would be praiseworthy, and we might have a discussion, except they're all about five years old. (And of course none displays anything reasonably described as "religion-based" "screaming.") The fact is, Paul Krugman has changed his views and has admitted he was mistaken, tried to pinpoint where he went wrong, and done this without claiming he was only referring to some vague possible future risks.

    http://www.businessinsider.com/big-things-paul-krugman-got-wrong-economy-2014-11

    http://krugman.blogs.nytimes.com/2013/04/13/missing-deflation/

    There's even been a tentative move along these line in this blog.

    http://johnhcochrane.blogspot.com/2014/12/policy-penance.html

    Let's hope we see more of it from all sides.

    ReplyDelete
    Replies
    1. And that justifies him calling others evil or mendacious idiots for being wrong on other things?

      Delete
    2. 1. Just so everyone is clear, the "religion-based" and "screaming" quotes are not from me. They're from "Anonymous." I never endorsed those words and I didn't attempt to speculate about what Krugman's motives are or aren't.

      2. "Zack's diligence in retrieving theses citations would be praiseworthy, and we might have a discussion, except they're all about five years old."

      I don't even know what your point is here. Outside of time-travel, how else are we supposed to measure the accuracy of statements about the future, other than waiting for a period of time to see what the actual results are? How's this: I think the Chicago Cubs will win the World Series sometime in the next few years. Obviously I can't be proven right or wrong right now.

      3. Your comment in the previous post (again responding to Anonymous):

      "Yikes! Kindly cite an instance or two of Paul Krugman's "repeated" "religion-based" "screaming[s]" that "'deflation is coming.'" Pure fantasy. Call it Krugman derangement dementia. Props to the grumpy one for (all attempts at) dodging this."

      All I did was post a few links showing that, yes, he did repeatedly warn of a deflationary disaster. That's it. Others can discuss whether his motives are "religion-based" or not. Again, I don't speculate about this stuff.

      4. See the paragraphs starting with "But I don't write endless posts excoriating..." I think that's the whole point here.

      Delete
    3. Professor Cochrane's larger points about communication and ambiguity are solid, but he, and Zack, misread the posts. The provided articles do not forecast deflation. They variously claim disinflationary trends, and the recursive dangers of deflation. The forecast is in the assertions about deflationary effects.

      Delete
    4. Why do you think I misread him when all I did was post a few links to his own columns? And if comments like "so we're really heading into Japanese-style deflation territory," "deflation isn't some distant possibility- it's already here by some measures, not far off by others", and "the risk that America will turn into Japan- that we'll face years of deflation and stagnation- seems, if anything, to be rising" are not considered forecasts, then none of Cochrane's (and others) previous concerns about inflation can possibly be either.

      Scroll down to JZ's comment- Krugman himself correctly admitted that he expected deflation and it didn't happen. The issue here isn't that the fact that he was wrong. As this post says, "Why does this matter? Normally, it doesn't and it shouldn't." I think Prof. Cochrane's point is very clear here and in previous posts on this topic.

      Delete
  2. Krugman is Krugman, and in fact Japan and Europe do face deflationary quagmires, while the United States perhaps dodged the bullet on the strength of QE.

    That said, in Cochrane's last post had a very interesting chart showing that at any point in time in the last 30 years economists (as a group) have predicted higher interest rates (and thus inflation). Instead, we have had a 30-year secular decline in inflation and interest rates.

    And today, there are still bodies of economists who predict inflation or hyperinflation right around the corner.

    There is a bias in the economics profession, deeply entrenched, to foretell pending higher inflation. The question is, why?

    ReplyDelete
  3. I think the spiral would never end, so society has sticky wages to reduce the risk/ speed of deflation.

    ReplyDelete
    Replies
    1. To clarify, I know that sticky wages are supposed to be the problem according to textbook Keynesianism or your recent treatments, but I seem to recall Krugman pointing out that adding sticky wages back into say his model with eggertson, gives the result closer to the reality that we observed. I'm sure if I cared to dig it up I'd find Keynes agreeing..

      Delete
    2. I heard the "sticky wages" part too. But, are you allowed to say you're and your model are always right if you make a forecast, it doesn't come out, then you add sticky wages to your model to explain why it doesn't come out? Also "sticky wages" are pretty much a Deus-ex-machina here. What are the rules of "sticky wages"? Aggregate nominal wages may never fall? The wages of an individual in an individual job may never fall? It makes a big difference!

      Delete
    3. "But, are you allowed to say you're and your model are always right if you make a forecast, it doesn't come out, then you add sticky wages to your model to explain why it doesn't come out?"

      You are right, this is a prime example of ad-hoc New Keynesianism.

      In old Keynesian economics, sticky wages are irrelevant. Keynesian policies in during the 1930s (the most important historical case study we have) worked in countries that were remarkably price variable. Sticky prices are not the reason Keynes argued that you can have a deflationary trap. Farmer is only one person who continually reminds us of this.

      Something Keynes did talk about was economic decline. This seems to have been repackaged as "secular stagnation". This requires proper historical analysis before we start talking about models. The recovery from the Great Depression was complicated by WWII events, so it is not helpful as a case study of their long term impact. However, economists may have to learn, as historians already have, that economies do not always grow and you cannot always expect multiplier effects from macro-policy.

      Delete
  4. This comment has been removed by the author.

    ReplyDelete
    Replies
    1. http://andolfatto.blogspot.com/2009/02/on-krugman-barro-boneheads-and-keynes.html?m=1

      Delete
  5. How many people have deferred gasoline purchases in anticipation that oil prices have further to fall? Where's the deflationary spiral at the pump?

    ReplyDelete
    Replies
    1. Paul doesn't make the distinction between durable / non-durable goods.

      Delete
  6. Mr. Cochrane:

    Many of those comments from Krugman date back to 2010 or earlier. As the data shows, we were losing jobs each month until about mid-2010 so the risk was that we could fall into deflationary spiral. You and other anti-stimulus types warned about runaway inflation. Well - that didn't happen either did it? We also got about as close a natural experiment for the impact of fiscal stimulus as possible in many European countries where an initial recovery was followed by another downturn. If your claim is that fiscal expansion does not result in income/job-growth, then by logical consistency, fiscal contraction shouldn't result in a slow-down either. Yet we did get a slowdown in Europe. The most recent example is Japan where they increased sales tax again and lo and behold, their growth numbers fell again. What other real-world proof do we need? Interest rates were zero/near-zero so that didn't change. Structural reforms by definition take a few years to play out. The only thing that changed in Japan was the sales tax cut and the roll-back of the cut. Surely "science" should adapt its models to real-world observations.

    ReplyDelete
  7. Dr. Cochrane,

    We understand that you dislike Krugman's blunt style and his liberal politics. In the interest of fairness and balance, however, would you please add the following link to your collection of Krugman's posts above?

    " ... back when the crisis started, I did expect to see deflation, Japanese style, if it went on for an extended period. I was wrong ... "

    http://krugman.blogs.nytimes.com/2013/03/05/why-dont-we-have-deflation

    ReplyDelete
    Replies
    1. Here's Paul Krugman on Art Laffer:

      "Recently some conference planners tried to recruit me for an event in which I would be presenting the alternative view to the main experts — Arthur Laffer and Stephen Moore. This would be the Art Laffer who among other things warned about soaring inflation and interest rates thanks to the rapid growth in the monetary base (ask the Swiss)"

      This was just a few days ago. Art Laffer did, in fact, admit he was wrong. I guess it doesn't matter to Paul, so why does Paul admitting he was wrong excuse him?

      Delete
    2. Oh by the way I missed the even meatier paragraph under it:

      "Obviously these “experts” appeal to the political prejudices of a business audience, but taking their advice would have cost you a lot of money. So why isn’t their popularity dented by the repeated pratfalls? Are they, also, in the entertainment business?"

      Delete
    3. Here's something else you missed, Unnamed One:

      Paul Krugman: In Praise of Art Laffer

      http://krugman.blogs.nytimes.com/2014/01/03/in-praise-of-art-laffer/

      Delete
  8. John,

    I am not a Keynesian but:

    "On the spiral or vortex. Perhaps there are a few true-blue Keynesians reading who can help me out. I understand the idea that deflation leads to high real rates leads to lower demand leads to lower output leads to more deflation. (Understand ≠ Agree.) I don't see how the model ever predicts this to end. Is there some clear and it bottoms out when x y z?"

    First in your standard Keynesian model change all agents from infinite lifetime to fixed lifetime. Second, change all goods from durable to non-durable.

    ReplyDelete
  9. "In my model, it bottoms out when you hit the top of the present value Laffer curve, and future taxes cannot hope to pay back the deflationary increase in the real value of the debt."

    This is non-sense. The present non-discounted value of all future tax revenue is infinite. Apply any discounting factor to it that you like (real or nominal) and it's still infinite. The only way you get to a finite value for future tax revenue is:

    1. Create a date in the future at which time the government in question ceases to exist (insurrection, Armageddon, foreign invader, etc.)
    2. Change the means by which a tax liability is settled

    ReplyDelete
  10. Please see the graphs here. Your belief is that the first graph always happens while the second never happens, but they both happened. You should try to figure out what you were wrong about, learn, and then explain by saying, "I was wrong to think that the second graph is impossible. I now see that it occurred because _______."

    ReplyDelete
    Replies
    1. The second graph only shows that increases in the Consumer Price Index have been relatively low in recent years compared to M1. When exactly did he claim this was "impossible" and "never happens"?

      Delete

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.