Winner of the New Statesman SPERI Prize in Political Economy 2016


Sunday, 8 February 2015

The Divine Coincidence in a parallel universe

The Divine Coincidence is the idea that by controlling inflation we also bring the output gap to zero, so we do not need separate targets for both. I talked about this in a recent post, but in writing that I realised I could make my point in another (and perhaps more effective) way. The Divine Coincidence essentially works both ways. So imagine a parallel universe where the monetary authority targeted the output gap, and not inflation. [1] [2] The authority said that by targeting the output gap they also controlled inflation.

Now you may make a practical objection at this point, which is that it is obvious to target inflation rather than the output gap because the latter is so difficult to measure. I think that argument becomes weak once we recognise that by targeting inflation, we are in fact trying to reduce the costs of inflation, and the published inflation rate may be a poor indicator of these costs. In a Woodford type framework, for example, inflation is costly because prices are sticky, so we should focus on those goods (and labour) where prices are sticky. This is one rationale for looking at core inflation, but core inflation is hardly a perfect measure of sticky prices. Arguably our proxies for the output gap, like unemployment, are at least as good at capturing the true costs of a non-zero output gap.

In this parallel universe they too had a Great Recession, and (being parallel and all) their recovery was of a very similar shape to ours. How would the output gap targeting monetary authority in this parallel universe perceive its performance? The story would be one of complete failure. After six years of trying, the output gap had still not been closed. A huge amount of resources had been wasted as a result. In fact, I suspect by now the monetary authority would have said quite explicitly that it just did not have the tools to do its job any more. Furthermore, they probably would have made it clear that one reason they did not have the tools (i.e. why interest rates were still stuck at the Zero Lower Bound) was because of fiscal austerity. If they did not try and blame someone else, they would look utterly incompetent.

I do not think, in our inflation targeting world, the monetary authorities have this view. Instead, based on the limited information I have, and at least outside the Eurozone, they believe performance over the last six years has not been too bad. Inflation was a bit high around 2011, and is maybe a bit low now, but nothing too serious. Yet the data they are looking at is exactly the same as the data in the parallel universe. The only difference is that they are targeting inflation, while in the parallel universe the focus is on the output gap. And by the Divine Coincidence, this difference should not matter!

My parallel universe idea illustrates two points. First, over the last six years, the Divine Coincidence has been distinctly unholy. Second, as a result it is terribly misleading to focus on inflation (and consumer price inflation in particular) rather than the output gap. I suspect in thirty years students will look back on this period with the same disbelief that we look back on the 1930s. How could they have allowed the recession to continue for so long, they will ask, when they had the tools to do much better? Part of the answer will be inflation targeting.
   
[1] It’s not quite that simple, because with either a traditional or New Keynesian Phillips curve, getting inflation to target ensures a zero output gap, but a zero output gap alone does not ensure hitting the inflation target. However I do not think this point is crucial here.

[2] Because of the dual mandate, the US Fed could in principle be in either universe. In practice they seem to focus on inflation.  

32 comments:

  1. Why not target NGDP? It's easier to measure than output gap, likely easier for a central bank to influence and fits nicely with our intuition of what should be on a steady upward path.

    ReplyDelete
  2. I suggest it’s not entirely true to say the BoE has targeted inflation during the recent recession. The BoE has been described as a “closet NGDP targetter”. That is, in the early part of the recession, inflation WAS on the high side, but the BoE didn’t abstain from interest rate cuts as much as we might have expected. That was because they thought much of the inflation was cost push, wasn’t it?

    As to the lack of FISCAL stimulus, that’s explained by macromedia thinking.

    ReplyDelete
  3. "the output gap targeting monetary authority"

    I think there is a little asymmetry in your scenario. A monetary authority has control over the money supply, not over fiscal policy. I'd say that the money supply has a closer relationship to inflation than to output. After all, it is possible to print so much money that at some point, inflation is going to increase without corresponding increases in output, so while there is a relation between money supply and output, they do not completely move in lockstep. The relationship between the monetary authority and inflation is direct, the relationship between the monetary authority and output is indirect, so to say.
    So an output targeting central bank, as we currently understand it, would only have indirect control over output. Wouldn't an output targeting central bank need direct control over, well, output as well? Essentially, it would need fiscal powers to fulfil its mission with the same degree of efficiency as inflation targeting institutions. This central bank would not just be a monetary institution, but something more.
    Still, as far as I know, this mutual relationship that results in the divine coincidence exists, a leads to b and b leads to a, only one way of causality is slower than the other. I think this idea is more important than generally understood, because it basically allows for multiple equilibria. The economy essentially switches from regime to regime. If inflation expectations impact actual inflation and actual inflation creates different expectations, then it becomes much harder for a centrla bank to change trend lines. I mean, right now, Europe is sliding into a deflationary regime and it seems that with each passing month, the necessary measures to changes course need to become more drastic.

    ReplyDelete
    Replies
    1. @ A.S.S.: "I'd say that the money supply has a closer relationship to inflation than to output. " - I agree, but this view is about as popular as advocating dogs and cats as sources of dietary protein at a meeting of the Human Society for Animals. The people in the monetary/Keynesian camps (flip sides of the same coin) believe in such fallacies as inflation is good, the Great Depression was cured by governments, and in money illusion / sticky prices / sticky wages, all of which are in the main false.

      Delete
    2. "A monetary authority has control over the money supply, not over fiscal policy."

      It doesn't have control over the money supply either in a fiat currency economy, there is no "multiplier" of "reserves" into Money M3, it's a myth. Quantitative Easing has proved that. (Acorn)

      Delete
    3. @ Anonymous:

      About QE, I think that's only true when the economy is depressed. And also, if QE does not do the trick, helicopter money might. The central banks have constrained themselves by not using it.

      Delete
    4. "Wouldn't an output targeting central bank need direct control over, well, output as well?"

      They certainly control nominal output. But they don't need direct control over real output any more than they need direct control over interest rates. OMOs should suffice.

      Delete
    5. Helicopter money is FISCAL stimulus. Only the Treasury can create (spend) new fiscal assets into the private sector. The central bank can't, it can only swap one asset for another. For the central bank to drop money out of a helicopter, someone would have to throw collateral back into the helicopter, it operates with a balance sheet, the Treasury, as the currency issuer, does not. (Acorn)

      Delete
    6. @ Anonymous:

      I disagree. Helicopter money does not directly affect the budget of a government. The central bank can literally print money and deposit that money on private accounts, in theory. I think your second sentence actually confirms the argument I made about the institutional deficiency of a monetary authority targeting an output gap.

      Delete
    7. The central bank can't "print money" until the currency issuer, the Treasury, issues "reserves", some of which can be physically "printed" with a picture of the Queen on them; as many as may be required by the commercial banks. The central bank can't increase the total "financial assets" in the economy, only the Treasury (the monopoly currency issuer) can do that.

      The Treasury throwing (spending new money) out of helicopters is a poor form of fiscal stimulus and of increasing the net financial assets in the economy, much better to buy goods and services in the normal manner. Then you could give some unemployed persons jobs, building houses for instance, and "throw" a wage packet at them.

      And yes, it would increase the government budget spend and its budget deficit and it would initially all be saved in the private sector. Eventually it would get spent around the economy, being taxed at each stage, until eventually the government gets all its money back again.

      Delete
    8. @ Anonymous:

      I think you're incorrect.

      http://mainlymacro.blogspot.de/2012/10/what-do-people-mean-by-helicopter-money.html

      Delete
    9. No disrespect to SWL but that is not how a fiat currency banking system operates. May I suggest the following.

      http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1725026

      Acorn

      Delete
    10. I checked a couple of links now and your paper. For example, Nick Rowe says it's basically both monetary and fiscal policy. A paper on Vox also differenciates between different kinds of helicopter drops. It's also being pointed out that central banks don't have the legal authority to conduct fiscal measures. But then that was also my argument above, i.e. central banks targeting output would have to assume more powers.

      Delete
    11. Agreed Alexander, it is worth noting that the government's Treasury and its Central Bank are one and the same operationally. There are no "independent" Central Banks in fiat economies. A point that has been highlighted in the new ESA2010 accounting rules. Such that, Treasury "Gilts" held by the Central Bank have to have the interest paid out on those Gilts, netted out of headline debt interest payments. The government Treasury is actually paying interest to itself, via its own Bank. Likewise, if the Central Bank has swapped Gilts back to Reserves (the QE process), then it is the same as if the Treasury Debt Management Office (DMO), had never issued those Gilts in the first place. (Acorn)

      Delete
  4. In the alternate universe they could reasonably blame fiscal austerity (and the trade deficit in the US), but they could also suggest more powerful policies to their political masters.

    They could target long-term interest rates outright, instead of vague, halfhearted QEs. They could raise the inflation target. Wouldn't they feel pressure to do more given that they were failing so bad. Helicopter drops?

    Bernanke stated outright that all of his Professorial advice to Japan was in order to get out of deflation. That's when you bring out the big guns. He didn't feel they were necessary to combat lowflation or close the output gap quickly. That would bring on too much political reaction from the right wing.

    Yellen is on record saying that 3.5-4 percent wage inflation is "normal." She's been in office a year so give her time she may surprise. "Normal" wage inflation occurs when the output gap is closed. The problem is that they took 6 years plus to get back to normal. Congresspeople and journalists should ask her continuously why it took so long since it's such a waste?

    No doubt she'll say it's because of fiscal austerity and the depth of the financial crisis. When Bernanke was given a softball question about how it takes a long time to recover from financial crises he responded with the deadpan "it takes a long time usually because of policy failure."

    ReplyDelete
    Replies
    1. "In the alternate universe [the Fed] could reasonably blame fiscal austerity (and the trade deficit in the US), but they could also suggest more powerful policies to their political masters."

      I have mixed feelings about Bernanke, but he actually did this consistently.

      In every single one of his many Congressional appearances after the Panic and initial small stimulus, he called for more deficit spending than was being done. He did so very politely and non-forcefully, but he really did make the point over and over and over again.

      Delete
  5. When I was a kid my mischievous buddy said, If we get in trouble, Deny, Deny, Deny.

    When an economy has balance sheet near depression, is at ZLB, with persistent low inflation and under employment and slow wage growth,

    I say Demand, Demand, Demand. Most sane folks knew the solution, you went to Thoma's blog and clicked links. It was fiscal. But politics is like that line from Full Metal Jacket- It's a #### sandwich, and we all gotta take abite.

    ReplyDelete
  6. If your preferred measure of the output gap is unemployment, then both in the UK and the US the output gap is nearly closed. In the US, there may be "disguised unemployment" due to lower labour force participation, this doesn't seem true in the UK.

    What is true is that output is below its pre-crisis trend in both countries. But in the case of the UK in particular, this is due to poor productivity. It's true that NK models can explain temporary falls in productivity during a recession with variable capital utilisation (for example) and "labour hoarding". But it's hard to believe that labour hoarding is still a factor, with firms waiting for years for the recession that never comes. Also, strong job growth in the UK and US in recent years seems inconsistent with labour hoarding.

    Is there another "demand side" explanation for weak productivity? If not, it seems sensible that it comes from the supply side - for example, still weak banks not allocating credit to the most productive firms. If the latter, then this can explain (i) on target inflation; (ii) unemployment approaching pre-crisis averages; and (iii) TFP and output below its previous trend. In this case, I believe that further stimulus won't be effective (if it doesn't deal with the relevant friction).

    I know that you disagree with this line of reasoning. But it would be illuminating to understand how you can reconcile persistently weak productivity with "demand side" recessions as we understand them from New Keynesian models. Thanks.

    ReplyDelete
    Replies
    1. I try and do so here:

      http://mainlymacro.blogspot.co.uk/2014/07/aggregate-demand-and-labour-market.html

      In addition, I think there are clear signs of lots of underutilisation of labour in the UK. Have a look at this:

      https://flipchartfairytales.wordpress.com/2015/02/06/uk-self-employment-success-story-or-basket-case/

      Delete
    2. @anon In the real world, rather than the model world (I don't actually mean to sound derogatory here) the reason for the increase in employment and the low level of productivity is transparent. Zero- or minimal-hours contracts, "self employment", contracting-out, general wage compression.

      Delete
  7. I really like this thought-experiment. We have to take it with a grain of salt, because, strictly speaking, targeting the output gap leaves expected inflation indeterminate. But it still works to make your point. And it works even better in Canada, where the BoC kept inflation pretty close to target (headline fluctuated above and below, but core just followed the trend).

    What I don't understand is why (some/many?) macroeconomists have lowered the bar for inflation targeting. Before the recession, I always thought that if the BoC succeeded in keeping inflation on target, it would also keep the output gap at (roughly) zero. I supported IT *because* I believed that Divine Coincidence was (at least probably, approximately) correct. I changed my mind when the data showed me that Divine Coincidence had failed, in a biggish way. Why didn't all supporters of IT change their minds too? Why did they lower their standards for monetary policy? "Soft bigotry of low expectations" etc.

    ReplyDelete
    Replies
    1. Just noticed your footnote 1, where you already have my "grain of salt" point.

      Delete
  8. Can u please be more special Facebook app Social

    ReplyDelete
  9. The US output gap is more of a power play than an economic problem. Business does not want to empower workers and retaining a certain portion of the output gap ensures less than full employment, thus crippling workers ability to obtain better pay and working conditions. There is no obvious economic reason not to close the output gap via fiscal stimulus. The 5% are doing quite well and have no wish to see their economic and political power diminished in order to benefit labor.

    ReplyDelete
    Replies
    1. i think that the problem is that business is that they know a fair bit about microeconomic decisions but little about marco. Soros would be an exception.

      Delete
  10. "I suspect in thirty years students will look back on this period with the same disbelief that we look back on the 1930s"

    As a Euro Area citizen, my reaction is much fiercer than that. Because in the 1930s, I accept that policymakers knew much less about the economy than today. Today, I see no excuses for the policy failures in the Euro Area. I am not an economist, but I understood very quickly in 2011 that the Euro crisis was not a debt crisis, but a monetary contraction. If I could understand that, then many policymakers should have as well.

    ReplyDelete
  11. I think you are wrong in characterizing the output gap/inflation connection as involving a "Divine Coincidence." In the main, it is definitional. The output gap is zero when inflation neither rises nor falls. In the same way, "potential growth" is defined as the "rate of growth the economy can sustain over time while maintaining a stable inflation rate." It's true that there's a theory of the inflation process in the background: roughly, some form of a Phillips curve. But the fact that controlling inflation and achieving potential output amount to the same thing is not a coincidental feature of our policy framework. It is axiomatic. In my view, at least, that leaves the framework without much comment. If we've learned anything in the last several years, it's that the distinction between cycle and trend is much more slippery than the economics profession has assumed.

    ReplyDelete
  12. I agree with Nick that this is a good post, and I'll try to make Nick's point in a slightly different way. Although I'm skeptical of fiscal stimulus, I think it's fair to say that many mainstream economists think fiscal stimulus is appropriate in a situation like this. That means they think AD has been too low. But if IT leaves AD so low that we need fiscal stimulus, then clearly there is something wrong with IT. And yet I often get the feeling that the "median centrist economist" thinks monetary policy was about right, but fiscal policy was too tight. The widely perceived view that we needed fiscal stimulus is one of the best arguments for switching from IT to NGDP level targeting.

    ReplyDelete
  13. "it is obvious to target inflation rather than the output gap because the latter is so difficult to measure. "

    I'd argue inflation is actually more difficult to measure. Yes, we arrive at various numbers (PCE, CPE, HICP, etc) we synonymize with "inflation," but the very notion we can boil down a multivariate, multidimensional number that is different for every combination of good and consumer into one number for a given polity is extremely problematic and necessarily subject to great imprecision. It's not clear what output "should" but output is at least a real number.

    That said, this is a great piece. The EU counterfactual from the alternate universe is indeed hard to imagine! Targeting the NGDP trend will no doubt seem obvious someday.

    ReplyDelete
  14. I am no economist but I do enjoy reading economics blogs.

    But as I understand it what you are saying about inflation targeting could also be said about debt. In a parallel universe, governments would care about the economic costs (present and future) of debt, not about debt itself. Alas reducing the debt has become a goal in itself - even if that debt reduction effort has huge economic costs and in fact fails to achieve any reduction of the Debt-to-GDP level at all. It has become common wisdom that an unemployment level of say, 20%, in some countries is a necessary evil in order to reduce debt.

    Krugman and others regularly denounce the 19th century vision of economics of many commentators (Say's law most notably). But if those people like ancient economics, shouldn't they go all the way back to barter trade economics ? At least it would be obvious that if a quarter of the population is sitting idle, that's indeed a huge and gratuitous waste of resources.


    ReplyDelete
  15. You can. A job guarantee at a fixed wage allows full employment and price stability by using an employed not unemployed buffer stock.

    ReplyDelete
  16. We don't need to wait 30 years:
    http://www.sciencedirect.com/science/article/pii/S0304393214001007

    ReplyDelete

Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.