Tuesday, November 22, 2011

The FOMC Narrowly Dodges Some Bad Policies: FOMC Minutes, November 1-2

The minutes from the last FOMC meeting are interesting. There was a long discussion of proposals to change Fed policymaking in some fundamental ways. The proposals were:

1. Explicit Inflation Targeting Inflation targeting is standard policy at a number of serious central banks in the world. Why not do it here?
Many participants pointed to the merits of specifying an explicit longer-run inflation goal, but it was noted that such a step could be misperceived as placing greater weight on price stability than on maximum employment; consequently, some suggested that a numerical inflation goal would need to be set forth within a context that clearly underscored the Committee's commitment to fostering both parts of its dual mandate.
I think the people making this argument are less concerned with satisfying the constraints on monetary policy set by Congress, than with using the dual mandate as an excuse to pursue active Keynesian stabilization policy. If the Fed wanted to, it could set explicit inflation targets, and argue convincingly that this kills two birds with one stone. A stable inflation rate also "promotes maximum employment." Indeed, there are even versions of New Keynesian models that can produce that result.

2. Conditional Commitment This part of the discussion seems odd. It starts with this:
As noted in the staff briefing, economic theory and model simulations suggested that a policy strategy involving such commitments could foster better macroeconomic outcomes than a discretionary approach of reoptimizing policy at every meeting, so long as the public understood the central bank's strategy and believed that policymakers would follow through on those commitments.
Note here that the notion is that the change under consideration - making conditional commitments about future policy - involves commitment vs. discretion. Later in the discussion, people are much more explicit about what they mean by conditional commitment:
In this vein, a number of participants expressed support for the possibility of clarifying the conditionality of the Committee's forward guidance about the trajectory of the federal funds rate through setting numerical thresholds for unemployment and inflation that would warrant exceptionally low levels for the policy rate.
The key point here is that, in fact, the proposal involves abandoning commitment. The current commitment involves "reoptimizing policy at each meeting," which is taking into account all the unforeseen circumstances that occurred since the last FOMC meeting. Conditional commitment is bad commitment, as we cannot commit conditionally to a policy response to an event that is not foreseen. Such events can in practice swamp everything else. Committing to a policy of reoptimizing at each meeting is in fact good commitment.

3. Nominal GDP/Price Level Targeting Here's the discussion:
The staff presented model simulations that suggested that nominal GDP targeting could, in principle, be helpful in promoting a stronger economic recovery in a context of longer-run price stability. Other simulations suggested that the single-minded pursuit of a price-level target would not be very effective in fostering maximum sustainable employment; it was noted, however, that price-level targeting where the central bank maintained flexibility to stabilize economic activity over the short term could generate economic outcomes that would be more consistent with the dual mandate.
The key question is what model the staff was using. Most likely it was the FRB US Model. If so, this is bogus. As members of the public, we cannot look at this model, but you can find bits and pieces of it in Fed publications. While there are words in those publications that might make you think this model might have some connection to any macroeconomics done post-1970, I don't think so. Best guess is that the FRB US model looks like the typical expanded IS-LM macroeconometric models developed pre-1970. If so, we can't take it seriously. Who cares if NGDP targeting "works" and price-level targeting does not, in that context? Get serious. See my post on NGDP targeting.

In any case, the FOMC decided not to take action on any of these proposals for now. There are some bad ideas floating around though, so beware.


  1. I don't know why they keep the FRB/US Model secret. What's the point? Also, if the Fed keeps the model secret, why should we care if they say such and such policy works according to the model. It appears from their FOMC minutes that they appear to take this model seriously so I hope it's something better than one of those large-scale econometric models. Also, you would think policy makers that are in the New Keynesian mold would at least use what they consider to be the standard models, something like Christiano-Eichenbaum-Evans or Smets-Wouters, especially since Charles Evans is an FOMC member.

  2. "I hope it's something better than one of those large-scale econometric models"

    It isn't -- it is basically Klein's old model with some weak form of rational expectations.

  3. The fact that the Board clings to its model is a symptom of broader problems with the organizational structure and incentives at the Board in Washington. Within the System, the innovative ideas and serious research are coming from the regional banks. That said, on the modeling front, it's not clear that we are any better off with some version of Christiano/Eichenbaum/Evans or Smets/Wouters. It's hard to see the structure in those things - you could think of them as data description. If put into use at the Board, models like that would no doubt continue to grow until they looked just like, basically, "Klein's old model."

  4. I thought they used a bunch of different models. I would be surprised if one of them was not a variant of CEE/Smets-Wouters.

    It is easy to criticize this model, but as they say it takes a model to beat a model.

    The great achievement of CEE/SW is that the model fits the data well enough that we can think of using it quantitatively. Build your alternative!

  5. "...it takes a model to beat a model."

    It's also true that no model can be better than a bad one.

    "The great achievement of CEE/SW is that the model fits the data well enough that we can think of using it quantitatively."

    Fit is not the criterion for usefulness. Klein-era models also fit the data. VARs fit the data. There is ultimately no more serious structure in CEE/SW than in the Klein-era models.

  6. It remains to spell out what "serious structure" actually means and how to make it quantitatively useful, not only a qualitative toy model to organize academic thinking. Ultimately the Fed must take a quantitatve stance. When do you define things as serious? If it is based on the micro notion of money search but ignores many potentially policy relevant dimensions? If it builds in policy relevant dimensions but puts them in brutally ad hoc as in CEE with some recurrence to micro-foundations, which are arguably not that compelling. I would love to see a blog post on this, I think there is a bit of an academic sidestep here on how to take this stance.

  7. "...qualitative toy model..."

    Don't say that. Before you start fitting the thing to the data, the basic NK model is as simple and abstract as anything I write down. The key current problem with basic NK and the CEE/SW models is that the detail on asset exchange, credit, banking, and the fundamentals of how monetary policy actually works, are essentially nonexistent. That's pretty damning for models that are set up for central banks to use to help formulate policy. Further, the attempts to incorporate some of these features in NK models are pretty feeble. I've put considerable effort into making what we call New Monetarist models more accessible, and applicable to practical policy problems. I can't do everything, though, and I'm not practiced in quantitative methods. It's up to people working in central banks to pick up the ball. The work is out there to use. I'll write more about this.

  8. debt is nominal, not real. As a result, nominal gdp targeting edges out inlation targeting. One of the many reasons to choose a positive inflation target is to minimize the chance of defaltion. too low an inflation target increases the frequency of financial crises. The underlying issue is that since debt is nominal, sufficiently large negative nominal income shocks cause failure and stem financial crises (defaults). Since the volatility of income is a major determinant of bank capital, too low a target increases the volatility of income which increases the capital requirement and lowers potential trend growth. this is aka a "velocity shock" The #1 reason to choose a nominal income target over an inflation target is to minimize the frequency of financial crises. It's no coincidence the "great moderation" was also associated with higher nominal income and higher inflation, thats what the theory says should happen. Now, one could argue that this merely means the inflation target needs to be raised. Sure, thats one way to do it. But the panics in Europe and the U.S. are not one-off events.

    financial crises are more frequent when money supply growth is too tight. The reason is nominal debt. the proximate cause is negative nominal income growth. consequently, nominal income targeting edges out inflation targeting.

  9. Just a follow up: It's hard to reconcile the belief that a small positive rate of inflation is optimal with the belief that nominal income targeting is sub-optimal. Underpinning the belief in a small positive rate of inflation (accepted by the FOMC) is the caustic effects of deflation.

    Any relevant model of the policy choice between nominal income targeting and inflation targeting must generate a small positive rate of inflation as optimal as well (the members of the FOMC accept a small rate of inflation of 1.75% to minimize deflation chances).

    Therefore, the question is not: is inflation targeting superior to nominal income targeting. The question is: under what conditions is 1) a small positive rate of inflation optimal; but 2)inflation targeting is superior.

    Seems to me, any model that generates the conditions for 1) also generates the conditions under which nominal income targeting is superior. So, show us a model of 1 and 2. To generate a model without 1 presents a false choice as its pretty widely accepted that deflation is caustic.

  10. "debt is nominal, not real. As a result, nominal gdp targeting edges out inlation targeting."

    That makes no sense at all.

  11. Just to clarify, though your post is titled "The FOMC Narrowly Dodges Some Bad Policies," your item no. 1 is more accurately a case of "The FOMC Fails to Adopt a Good Policy" (adopting an explicit inflation target), no?

    I agree with the rest of your post.

  12. dwb seems quite insane. Perhaps he is our good friend Greg Ransom in disguise?

  13. "Greg Ransom in disguise"

    No, more likely somebody from the Scott Sumner school of evangelical monetarism - don't worry about anything scientific like models, just have the faith that nominal GDP targeting will save us all!

  14. you'll have to be more creative that that.

    Again, there are a variety of reasons a small positive inflation rate is considered optimal, such as sticky wages to avoid debt-deflation (c.f. http://kcfed.org/PUBLICAT/ECONREV/PDF/2q08billi_kahn.pdf)

    The statement is that any model that generates the conditions whereby a (1) small positive rate of inflation is superior also generates the conditions under which (2) nominal income targeting is superior to inflation targeting. So, show us a model disproving this assertion (1 and not 2).

    Sure, you can deny 1. That's fine. Lotsa papers written on the optimal inflation rate going back 2 centuries (see the KC Fed paper above). BUT the real policy choice is inflation targeting vs nominal income targeting, so any model has to generate real world features that imply that a small positive rate of inflation is good.

    If you are so stumped as to resort to insults, thats perfect, i'll consider my point proven.

  15. Scott Sumner has a post today about the coordinated central bank actions - seems to be of the opinion that this has had substantial effects on asset prices.

  16. "A stable inflation rate also "promotes maximum employment." Indeed, there are even versions of New Keynesian models that can produce that result."

    This strikes me as screwy. Based on models, we are supposed to follow policies that may or may not result in robust economic growth?

    Robust economic growth should be the goal of macroeconomic policies. Please, if you must genuflect to models (or gold), then kneel down before models that target robust economic growth.

    Fighting inflation comes a distant second or third. In a nation that uses global supply chains, and freely imported goods, services, capital, and (until recently) labor, nflation is less a concern than ever. If any good or service becomes too pricey, competition pops up. Globally. And the globe is a cheap place.

    If you want to beat inflation, see Japan. Boy, they whipped inflation there, and the yen is sky-high, an indication of tight money. Of course, their nominal GDP is below the level of 20 years ago, and stock and property markets are off by 80 percent.

  17. The current and peevish fixation on inflation, even inflation in very low single digits, is inexplicable.

    The USA economy did fine from 1982 through 2008 with moderate and varying rates of inflation.

    What makes 2 percent inflation magic? What if more-robust long-term growth is obtained with 3 percent inflation? Should we suffocate growth in order to hit the 2 percent inflation target?

    Add on, that measuring inflation is difficult.

    We should genuflect to a price index that everyone knows is only roughly on target? While people are unemployed? And businesses lose profits? And the DJIA is stuck at 1999 levels? And property markets are down?

    Reminds me of English settlers in Hawaii putting on the woolens in "winter" as it was the "right thing to do." Come Christmas, wrap up!

  18. Why stop at 3% inflation? What's magic about that ? How about 4, 5, 6, 10 %? Wouldn't that be even better?

  19. Anonymous:

    You mke my point for me. The "magic" rate of inflation is any rate.

    The most important aspect of microeconomic policy is robust growth. Who cares if we get robust growth at steady 5 percent inflation?

    Would you prefer feeble growth and deflation, ala Japan?

    This is what is wrong with the right-wing kooks today. They start from the premise that no inflation (even if that could be measured) is the ideal, not from the premise that robust economic growth is the ideal.

    Prices are just a nominal index, that can go up or down. Why fixate on a nominal index?

  20. Interesting post. I happen to think culture trumps even government policies, when it comes economic success. A culture in which honesty and the work ethic are paramount will always beat a corrupt lazy country.

    Sweden will always be a nicer place to live than Nigeria, even if Sweden is icebound and socialist and Nigeria has fertile fields and is capitalist.

    I guess if the government is rotten enough–N. Korea comes to mind–then government can trump culture, but we are talking gong-show governments.

    This comes back to my favorite idea, and that is that in the USA we should print gobs of money. The American culture is freewheeling, risk-taking, boom-loving. Monetary policy that talks about steady two percent inflation trumping all other concerns is death on American culture.

    Americans start business and invest when they think there are going to hit a home run. Boom, baby, boom. No one buys a rundown property in the hopes that it will appreciate at a two percent annual rate. Listening to an anal-gnome pettifog about inflation will kill the typical American businessman.

    A monetary noose around the necks of Americans is death on our culture and incentive-system to invest. We are not Swedes or Japanese or Germans or some wieners from Austria.

    Americans do not play soccer for hours in the hopes of eking out a 2 to 2 tie. We want to hit the home ruin, or spike it in the end zone. The slam dunk at the buzzer.

    Tight money will never work in America—we were better off when counterfeiting was rife in the pioneer days. I wish we had successful counterfeiters today. Lots of ‘em.

  21. Hmm, some interesting ideas from Benjamin. And the idea that money printing boosts growth is testable. Perhaps he could produce a graph for us from historical data showing how the economy grows faster when the money supply grows faster. A simple correlation plot would do the trick. Can you do that for us?
    This evidence would surely convince even Prof Williamson!

  22. Anon-

    Economists have complicated and fragile models pertaining to money supply, but they get gummed up by velocity.

    In general, I feel it is better to run an economy "hot," towards full capacity and moderate inflation, than to run it "cold" towards deflation and ignoring output.

  23. Benjamin: What you need to do is to explain American exceptionalism. Why have so many rich OECD countries adopted inflation targeting but the US sticks resolutely to the dual mandate?

    This despite many pundits suggesting that the dual mandate contributed to the financial crisis of 2008.