Believe those who are seeking the truth. Doubt those who find it. Andre Gide

Monday, April 30, 2012

NGDP Targeting: Some Answers

I want to thank everyone who replied to my previous blog post NGDP Targeting: Some Questions. It will take me some time to digest all of the information sent to me. In the meantime, let me report on a few of the answers I received.

First, some background information. I do not believe that sticky nominal prices or wages matter (at least as far as explaining years of sub par recovery dynamics). I explain why here: The Sticky Price Hypothesis: A Critique. Consequently, Nick Rowe's reply to my post does nothing for me (although I still love the man and his blog!). On the other hand, I am not so sure about "sticky" nominal debt. I am more sympathetic to Evan Koenig's view:
The analysis presented here is completely orthogonal to the literature. It does not involve goods-market or labor-market pricing frictions in any way. As our most severe economic downturns have been characterized by widespread default on financial obligations and disastrous breakdowns or near breakdowns in lending, an analytical framework that puts debt and the distribution of risk at center stage arguably has something to say about optimal policy. 
One of the main proponents of NGDP targeting sent me this article, so I took it to represent a main theoretical justification for NGDP targeting. And indeed, a lot of people seem to be talking about a "debt overhang" problem and a "balance sheet recession." I sort of figured (perhaps incorrectly) that the idea of getting the Fed to commit immediately to (say) a 5% NGDP target was to generate a credible temporary inflation to reverse the effect of the unanticipated and sharp decline in the price-level path (in 2008).  The mechanism people have in mind, I think, is essentially to reduce the real debt burden of debt-constrained households, to get them to start spending, and to increase aggregate demand.

In my previous point, I raised the question of how strong and how desirable this mechanism might be now that we are 3 years out from the 2008 price level shock. Surely, a lot of the debt negotiated prior to the shock has been either reneged, renegotiated, or retired. At the same time, a lot of new debt has presumably been issued under the expectation of the new price-level path (given that people generally believe that the Fed will stick to its 2% inflation target). If the "turnover" rate is high (i.e., if there are large gross flows of debt being created and destroyed), and if the economy remains under "potential" for a long time, then one would have to question the quantitative importance of this mechanism; and also, the desirability of reversing the price-level path.

Well, I have to thank Mark Sadowski for taking the time to dig up some statistics for us. You can refer to the comments section of my previous post for details, but Mark's back of the envelope calculation is summarized here:
At the end of 2011, there was some $13.2 trillion in household debt outstanding. Of that nearly three quarters, or about $9.8 trillion, consisted of home mortgages  
Thus a total of perhaps $5 trillion in debt has been originated/refinanced since the new NGDP trend has been established. Which means that about $8.2 trillion or approximately 64% was negotiated before the new trend was established. 
Assuming that the rate of origination/refinancing is linear (dubious) then it will take a least another five years before all household debt conforms to current NGDP growth expectations. 
So, it seems that there is still a lot of "old" debt out there, negotiated under the old price-level path. But there is also $5 trillion in new debt, negotiated under a new price-level path. And the longer we wait, the more this number will grow. Granted, this problem may have been avoided if the Fed went into the crisis with a credible NGDP target. But this is not the world we live in. What would Scott Sumner do right now? Who is he willing to make angry and why? Scott offers a hint here: Can we confident about the benefits of more NGDP?  
2. But does it still make sense to go back to the pre-2008 trend line? Probably not, recently I’ve been calling on the Fed to go about 1/3 of the way back to that trend line, and then start a new policy trajectory (hopefully explicit in this case.)
In any case, it seems that Scott believes that "more NGDP right now would modestly reduce the unemployment rate."  I confess that I am not entirely sure what mechanism he has in mind here. I really do need to read his 1000 blog posts on the subject one day!

I still have a lot of reading to do before forming an opinion on this subject. There were a lot of really good comments on my post that I haven't mentioned here--I need some time to think them through. Before I sign off though, some of you may be interested in these two links (h/t Prof J):

First, here is George Selgin: Wide off the mark, or, Nonsense about NGDP targeting. This seems like an extreme view, but I think it deserves some attention. 

Second, we have Mark Carney (Governor of the Bank of Canada) speaking here on why he believes a "flexible inflation target" is superior to an NGDP target. 


  1. About the debt renewal. You seem to implicitly assume that any newly negotiated debt is optimal according to long term trends. However, I don't think this is true. They were negotiated while recovery was still underway, so they were reflecting a moving target. So they are only partially better than the old debt. However, a faster recovery would also mean that sooner newly negotiated debt will reflect correct long term trends. Yet another reason for NGDP targeting.

    1. I think thats correct. The assumption being made here is that we have achieved the long term trend, but we have not because unemployment remains high. Home prices are still dropping (see below). Which means that the 36% that has been created (mostly refinances for mortgages by the way) is on average underwater.

      As for the 64%, in terms of mortgages, I seriously doubt there will be much further turnover. The refi eligible population has dwindled. For the vast majority of mortgages outstanding that have NOT refinanced already, its because they are not eligible: negative equity and/or bad credit (due to unemployment) (see pages 92/93 and fig 3/4 in the Nomure presentation i managed to find online). Hard to read the figure, but looks to me like 90% of prime mortgages have a refinance incentive but have not refinanced.

      The rate of home price decline has certainly slowed, but its likely to continue to decline somewhat until employment growth strengthens.

      *sigh* the fact that economists dont understand the housing market 5 years into the balance sheet recession shows you why were are all doomed.

    2. "is on average underwater" -> not necessarily underwater, sorry, home prices are on average lower than when the mortgage was negotiated. if the mortgage was written with 80% LTV, the homeowners might have merely lost equity, depending on how ar they've dropped.

  2. *sigh* the fact that economists dont understand the housing market 5 years into the balance sheet recession shows you why were are all doomed.

    Cheer up, dwb...we are all doomed whether economists understand the housing market or not.

    1. hey, well, my excuse is that you can tell the temperature of the country by the fact that some of the top shows on the discovery and national geographic channel include: Doomsday Preppers, Doomsday Bunkers, American Guns, Sons of Guns, Dual Survival, and Gold Rush. Hmm. And yes, you can even find those doomsday-preparedness recipies on the Food Channel. I would say that it sounds like the proletariat was getting ready to revolt, except that these doomsday bunkers cost upwards of $400,000, and the amount of money people spend on their toys is ridiculous.

      i will be really glad when we are back to normal.

    2. On the plus side, things still look better than they did for Augustine of Hippo, c. 430AD. I think we still have some way to go before we reach that stage...

  3. Carney's speech is interesting - it's a lot of platitudes about this and that, taking credit for Canadian resilience ignoring the fact that the banking system there never became capital constrained, and then a very short and rather thin crux argument - that the quantitative benefits of a price level target are the same as a small reduction in the variance of a CPI target, once we ignore the big disaster and lower bound cases.

    Then, he says that these gains wipe away if there isn't enough credibility in the new regime, thus he is effectively comparing the steady state settled performance of current regime with an agnostic new regime's initial pitfalls. It's an argument from legacy. It's an argument about not leaving your current girlfriend for the charm and pitfalls of a new romance. It's a good argument, but it's not an argument from economic theory.

    And he doesn't even handle NGDP targeting, subsuming it under the broader construct of level targeting and just noting that an NGDp target is unresponsive to secular declines in growth rates. Sure, a fixed NGDP target is. But why wouldn't an NGDP target be revised on the realisation of a secular decline in growth rates? Carney doesn't say.

  4. Carney's speech is interesting but honestly i am not exactly clear what "flexible inflation targeting" means, strictly. I've heard the Taylor tule called flexible inflation targeting (and also heard ngdp targeting equated to flexible inflation targeting as well). Seems to me Australia is the purest form of flex. inf Targeting, where they actually have a 2-3% band as a target. Flexible inflation targeting sounds an awful lot like "we want lots of discretion." Too much discretion is a bad thing. I guess i am either for it or against it depending on what the definition is.

    On Selgins argument I am more sympathetic. But, I do not think that an ngdp target really means ngdp will never deviate in practice. unexpected shocks will still occur. And, as a practical matter, i would expect the Fed to revisit the ngdp target periodically if "trend inflation" was too high (in other words, as a practical not purist matter, i almost see a "conditional ngdp target", that is, the ngdp target most consistent with long run price stability).

    He says:

    That means having a single mandate only, where that mandate calls for the central bank to keep spending stable, and then tolerate as optimal, if it does not actually welcome, those changes in P and y that occur despite that stability.

    Generally I agree that the goal should be for fed policy to achieve to the extent possible the "optimal" flexible price equilibrium. No policy will be perfect (e.g. in the guts of say a NK model where wages are sticky the "optimal control" solution's weight on output/inflation depends on wage stickiness).

    Also, i think its useful to think of the Fed allowing "deviations" around the trend with the expectation that they are borrowed or made up in the future (I think Hoenig's paper makes this point). That, it seems to me, makes it much more likely the CB is accomodating AS shocks and responding to AD shocks (which is the basic idea IMO).