A model

Blanchard and Gali (2008) incorporate unemployment into the standard model. They have some interesting findings, but this one stuck out:
Unemployment under inflation targeting
This is the response of the unemployment rate over time to a 1% decrease in productivity of the economy (a “real” shock) WHEN the Fed has a inflation-only target (i.e. it doesn’t care about unemployment). Blue is the response in the model when its calibrated to look like the American economy and red is the response when its calibrated to look like Europe. Unemployment keeps increasing after the shock for a couple periods and then gradually declines. Look familiar?

Here’s the same graph but for when the Fed has the *optimal* policy of targeting a weighted average of inflation and unemployment:
Unemployment under optimal policy
Unemployment still jumps but it doesn’t have the hump shape. More importantly, notice how little the unemployment increases when the Fed is following optimal, or best, policy… almost an order of magnitude difference in the response of unemployment.

What does this mean? This paper gives us two ways to interpret what happened since last Fall. One, there was a small to medium sized real shock but because the Fed cares too much about inflation, unemployment sky rocketed. Reality looked (and looks) like the first graph. Or two, the Fed is following optimal policy but there was a huge real shock. Our reality is more like the second graph, but amplified.

Ironically, those making a bunch of noise about banking regulation, centering the blame for the recession on the financial sector, are arguing for the second graph. My impression is that those people are also more likely to be agitating for more aggregate demand policy. The second graph has the Fed acting optimally, i.e. there’s no need for further stimulus because the Fed is doing everything necessary. If you buy the logic of this paper, however, you can’t have it both ways.

PS – This model also has a positive response of inflation to the real shock. Given we saw a negative response last Fall there’s still room for a monetary shock in the story.

25 thoughts on “A model”

  1. Forgive me if my undergraduate macro is jumping to conclusions here, but doesn’t this reconcile the Kling-Sumner divide?

    The scenario you’re describing is a real shock, so the economy must perform Kling’s “recalculation”. However, poor monetary policy actually causes most of the damage, a la Sumner.

    I am assuming here that targeting a blend of inflation and unemployment is closer to Sumner’s recommendation of targeting NGDP expectations than targeting inflation alone.

  2. Could be Kevin. I’m not sure about the last sentence you write, though. Maybe Prof. Sumner will stop by. I tend to think Sumner means the Fed made a mistake (actual policy deviated from their desired policy) when he says they had bad policy, not necessarily that they had a sub-optimal policy target.

  3. I don’t think anyone would claim that the Fed or for that matter the ECB have lost all interest in unemployment (before or during current events) so that’s not what’s happening here, I guess.

    Any model in which there are “adjustment costs” or some other mechanism that prevents some agents from changing their behavior will generate larger, hump shape-ier responses to shocks. A hump shaped IRF is neither sufficient nor necessary for a good model, IMHO.

    P.S. Generic disapproval comments on “the standard model”.

  4. AC, I think the problem for the interpretation of these events using the second graph is that it would take one hell of a real shock (~10%) to get us to the levels of unemployment we’re seeing. Also, I place the time of the shock at about September last year. Unemployment has been humped shaped since then (it was only 6.2% in September ’08 and its only now reaching its peak… we hope).

    BTW, the Taylor rule that approximates the optimal policy in this paper suggests interest rates should have decreased 6.6% since the beginning of the recession. They’ve decreased about 5%. If you just look at August through December of last year, interest rates should have dropped 8.2%. Most of that would have been a reaction to falls in inflation. They actually only fell 2% in that period and I don’t think QE-like policy started until winter. If you just look at this year, interest rates should have fallen less than .5% and all of that would have been a reaction to increasing unemployment.

    This makes the episode look more like a mistake rather than the problem being sub-optimal policy targets (i.e. too little concern for unemployment).

  5. Well, Sumner’s knock on the Fed in the Fall of 2008 is that they used interest rates as an indicator of whether money was qualitatively loose or tight, rather than NGDP expectation. My guess is that this laser-like focus on interest rates is a result of targeting just inflation.

    If you also look at the change in unemployment, you’re implicitly examining whether AD is rising or falling. And I think that Sumner’s assertion is that people believe that AD today is driven a lot by expectations of future NGDP.

    I agree that it would be nice if he dropped by and explained this result in his framework. Because resolving the differences between his and Kling’s perspectives is a goal of mine. Both make a lot of sense to me.

  6. Push, I’m not sure I understand your latest reply. I was not trying to suggest that we use the 2nd graph rather than the 1st one to make sense of things. I was saying that the assumption that gets us from the 2nd one to the 1st must be patently false (pure, mechanical, unsmoothed inflation hawkishness is not what the Fed or ECB are about).

    Re: the numbers you mentioned, here’s the St. Louis take on things: http://research.stlouisfed.org/publications/mt/page10.pdf

    I’m surprised to see you pushing a NK take on things since I take the consensus to be that these models have very little, if anything, to say about recent events. Personally, I agree with CKM that in order to get these models to be relevant for policy-making, too many compromises and too much nonsense have been accepted.

    Maybe it’s just that my brain is marinating in the “wrong” departments :-)

  7. AC, I’m not pushing the view… I’m having a conversation conditional on the view being correct.

    Also, I’ve observed that for a framework to be taken seriously by policy makers, the word “Keynes” has to be somewhere in its title.

    Kevin, my take is that in December 2007 a Great Recalculation (out of finance and construction into health and education or whatever) pushed us into a mild to moderate recession. Then something else happened in Fall of 2008. If nothing else, Sumner should be given credit for the observation that there was something qualitatively different about Fall 2008. He is probably right it was a money shock, but in any case it was something big. Explanations that have the chickens of the financial mess just coming home to roost don’t work for me… People, if not the headline writers, knew — for years! — what was wrong in finance. Why the sudden cliff dive 14-15 months ago?

  8. I agree with your story and I think it’s Sumner’s as well. In general, I’m inclined to believe that there are several different kinds of downturns in order of frequency:

    – real shock, followed by expansionary monetary policy = mild recession due to real recalculation

    – monetary shock, followed by expansionary monetary policy = mild recession due to perturbation from equilibrium

    – real shock, followed by contractionary monetary policy = severe recession due to positive feedback–this is what we are just (hopefully) coming out of

    – monetary shock, followed by contractionary monetary policy = ? (don’t know if this has ever happened)

    Obviously, the policy problem is identifying a priori when you have qualitatively contractionary monetary policy.

  9. push- so what happened in September 2008? There must be some explanation, no? The Fed was not raising interest rates or contracting the money supply. (Though I recognize that monetary policy could be “tight” in a Sumnerian sense).

    Who are these people that knew what was wrong with finance? The massive growth in the financial sector up to that point doesn’t make it seem that people having a big impact on the market were seeing many problems in the financial sector.

    The major cliff dive happened because of a major financial crisis in Lehman brothers bankrupcy, Merryl Lynch’s sale, and AIG’s problems, with a stock market crash. See Reinhart-Rogoff for many historical instances of financial crises leading to major output drops. There’s really nothing surprising about it.

  10. I exchanged brief emails with Sumner. The money line was that he agrees with my assessment of unemployment + inflation being closer to his model, “I’ve always thought NGDP targeting is similar to other hybrids that combine prices with a real variable.” He also said he may do a post on this if he can find the time.

  11. gabe, how do you measure “the massive growth in the financial sector up to [Fall 2008]”? Employment started declining in early-mid 2007 in the finance sector (also in construction)… almost a year before employment started declining everywhere else.

    I haven’t read the RR book yet. Do they deal with endogeniety? Which way does causation run between “real” and financial crises?

  12. push- I was not aware that employment started declining in 2007, though that seems reasonable. So growth in the financial sector up to 2007 then. Just look at financial profits as a fraction of total corporate profits, or income for employees in the financial sector, etc.

    RR look at the aftermath of financial crises, so in a granger causality sense, it would be causal, though maybe not given expectations of crises. Is Granger causality just a post-hoc ergo-proctor hoc fallcy though? Tobin 1970 though so.

  13. Hmmm… I wonder if they deal with money shocks. An interesting story would be that large real shocks are followed by a lethargic money policy response which is the lynch pin of the financial crisis which may or may not have further effects on the real economy. In the simplest case, it doesn’t and the “effects of the financial crisis” are really just effects of the money shock. I guess I’m just convincing myself that the RR book should be higher on my to-read pile…

    Most analyses treat shocks as orthogonal (“here’s the IRF for a money shock, here’s one for a real shock”), but its likely that money shocks are correlated with real shocks (because the Fed doesn’t respond strong enough to the real shock).

  14. Huh… if monetary policy is bad at responding to real shocks, that better be endogenous and not in the shocks.

    Or maybe they just need some time to figure out there was a shock?

  15. No Socratic method, please!

    If monetary shocks are unexpected deviations of the instrument from what is set by the monetary authority’s rule or best response function and if those are correlated with stuff going on then you’ve probably misspecified the rule or problem of the monetary authority?

  16. Looking at pg. 30, the coefficient on inflation and unemployment deviation are 5 and -0.8 respectively. Add rho to those terms, and you get i_t. So if inflation is 1% and unemployment is 6% higher than average, then the interest rate should be rho+0.2%! The baseline unemployment for the US is 5%. So if rho is 2% like the benchmark Taylor Rule, that means fed funds rates over 2%, with 11% unemployment. I can’t accept that.

    Beyond using productivity shocks are the main driver, why are real wages rigid? I can see nominal rigidities more easily than real rigidities. Is this some kind of COLA wage increases? Also, would we expect for nominal wages to fall as fast as prices under deflation?

    Also, why is the increase in inflation what we want to consider? The levels are what the paper looks at. It doesn’t seem that expectations of inflation are high either. Do we really think that an increase from 0% to 1% inflation is a reason to raise interest rates by 5% in the USA? If you believe GB we should.

  17. Since it fits in this thread, I want to address your question of why we need a mechanism for unemployment lags. As a baseline, we know that when output and inflation are declining, AD stimulus is called for. We know that if we’re at long-run equilibrium, then stimulus will just cause inflation. The question is: is lagging unemployment an equilibrium outcome. Can we be at potential output when there’s above natural rate unemployment?

    Let me give a far fetched example. Suppose we wake up tomorrow and find that the entire economy consists of goo manufacturers. Now goo making is an intricate production process… goo specialist have to go to special goo schools to learn to make goo. Ok. We wake up and we’re all unemployed. 100% unemployment rate AND the economy is producing at its maximum. Also, as more and more people get geared up to make goo, GDP is rising but unemployment is much, much above its “natural” rate.

    There are two types of mechanisms causing unemployment to lag: ones that are responsive to monetary policy and ones that are not. In the goo example, no matter how stimulus you pour into the economy, all of it will go to inflation… People can’t be made to learn faster. Now, for money policy to help with unemployment in an environment of increasing output and increasing inflation, it has to reduce the lagginess of unemployment. In the goo example, it would have to make people learn faster how to make goo (unlikely, right?).

    Personally, my dissertation is about domain specific human capital and so the goo story resonates with me. But that just may be because its my lamppost. What mechanism for laggy unemployment makes it less laggy with additional AD stimulus?

  18. Sept 08 market correction was caused by aa number of factors. 1) counter party risk in the commercial paper market driven mostly by transparency of the assets being secured. 2.) CDO platforms which issued the notes for the commerical paper market used synthetic structures 3.) differing regulatory enviroments created regulatory arbitrage conditions which exacerbated the above situation. As a trader/banker i like many had long suspected structural problems. But due to the segementation of jobs and deal structures its hard for anyone person to see it all. So we all know that markets tend to self reenforce, and when you have a system risk based failure based upon the factors mentioned you get alot of downward presure. The commercial paper market which is the principle short term cash flow market has not recovered. The EU has made great progress and is much further ahead than the US on clearing and settlement platforms to address this issue. The US has done very little here. So what does any of this have to do with JOBS!. Simple, the US economy is driven by small business, they have no access to credit. So job growth is not likely in the near future. How do you get GDP to grow if the backbone lacks capital? You dont. Problem is simple to solve. Take 500B and push it into the small company market and you will see growth.

  19. What am I missing? Assume you tack an “error term” at the end of your Taylor rule or your money process is AR(1) and it has an error term. There’s a case to be made that those shocks be orthogonal to everything else and for you to model explicitly any linkages rather than put some link in the covariance matrix of the shocks.

    Re: the goo story, that sounds like a flexible price setup. If we’re retooling while some prices are wrong, “stimulus” might cause more than just inflation?

  20. RP, how do you measure prices and quantities in the “credit market”? You’d think rates would rise if there was a shift to the left of the supply curve. I get data from the Fed. Am I looking at the wrong prices?

  21. “here’s a case to be made that those shocks be orthogonal to everything else”

    My complaint is this is an assumption in most pencil sharpened analyses, but the stories people tell often have the Fed “overreacting” to a shock.

  22. Pushmedia- prices in CDO markets are base upon “mark to market” rules. Here is the problem. Many platforms have a high degree of independance in pricing, because so few instruments maybe compariable. Now this was a very significant issue in the market correction last year. Platforms that only had a few comparibles but one sold a lower level, then everyone would get marked down, this was a negative cycle reinforcing itself. Prices are found in exhcanges not the fed.

Comments are closed.