Friday, March 6, 2015

We're out of our ever loving minds.

I was listening to NPR this morning, and talk, of course, was of the employment report, most importantly as a signal about when the Fed would raise interest rates.  This is an example of one of my rules:  If people are debating it on NPR, we have already failed.

How bizarre is it that the health of the entire economy hinges on this one price that we set by committee?  Of course we are constantly on a knife's edge, fearing economic dislocation.  That's what happens when you set the price of something by committee.  It will be wrong, because that's not what a price is.  A price isn't something you set.  A price is something you observe.

Sure, it looks like prices are set.  It also looks like the sun goes around the earth.

So, we have this one thing, currency, that is provided by government monopoly.  It doesn't have to be.  It hasn't always been.  But it is.  And we manage the production of that currency by targeting the price of very short term credit.  We don't have to do it that way.  The connection between that price and the stability of the currency isn't even particularly easy to track or manage.  But we try.

Do we need to live in a world where I have to track the mood of this committee?  Why not have a world where, like with the price of bananas and haircuts, we observe the prices instead of trying to reverse engineer them on radio talk shows?


  1. Finally, a post from Kevin that I can understand (unlike your last one on upward nominal rigidity in asset prices, which is kind of a fascinating concept but I haven't been able to digest whether it's real, and if it is, what the implications would be.)

    Sorry, enough about my personal limitations. I think your sentiments here are exactly right. The amount of time spent analyzing what the word "patient" might mean feels so primitive.

    Perhaps if the Fed practiced NGDP level targeting, expanding or contracting the monetary base according to a prediction market, we would stop worrying so much about how hawkish or dovish a particular FOMC member is sounding on any particular day. And have enhanced economic stability too.


    Kenneth Duda
    Menlo Park, CA

    1. If NPR could only talk about subjects on which they had actual knowledge....

    2. I know. Most of their topics make me ask, "Why do we have to talk about this?"
      The answer is usually, "Because it's political."
      My response to that would be, "Then let's fix it by making this topic not political and then talk about something else."

      But, the progressives that listen to NPR want everything to be political. As I commented at Scott Sumner's blog yesterday, progressivism in power becomes sectarianism. Of course, the progressives I know would deny that they are engaged in sectarianism. They would say that we need to push everything that is contentious into the public realm, because everything we disagree about is too important for them not to impose their opinions on the rest of us.
      ...not sectarian at all.

  2. Hi Kevin,

    I commented on your comment at Bob Murphy's blog, mentioning I'd offer some academic papers on what you are doing with your housing posts. I like this one as a place to start:

    Poterba is a big name in this field, so you can find lots of stuff by him. Anyway, it's a good source of data for calibrating models, and you can see how they are calculating the "fair" return on housing.

    Keep up the good work! You have a thoughtful blog, and I'd like to convert lots of your writing into a monograph for my students.

    1. I linked to that paper in part 5. That's ok, though. Keep em coming. I'm sure there is a lot I haven't read. I started out on the first couple of posts with a model, and I've gotten so sidetracked that I'm probably at least 10 posts away from getting back to recalibrating it. Eventually, that's my plan, though.

      Thanks for the kind words.

    2. Ah, missed it on the first pass. I'm such a terrible blog reader - my eyes glaze over pretty fast just because of the medium.

  3. Meant to sign above as "Jeff" so you'd recognize the name.

  4. To be fair, the FRB sets a price "target". The error is the public's, for thinking of the votes of the board on a target level set the market price of anything. However, the target price does provide a reliable signal of board policy, particularly when the target and the market rate differ. Fortunately, these signals are often enough to induce behavioral changes that push prices towards target, meaning the fed often does not have to do much to achieve its policy goals. However, the pace of conversion may further confuse the public, by fueling the perception that the fed sets prices.

    The difference may appear semantic but it is critical. In the old Soviet Union, a committee literally set prices, and they were typically driven more by political concerns than market reality. The fed acknowledges market mechanisms in price setting, and when it sets price targets, it undertakes market operations intended to push prices toward the objective. But the board is a slave to reality - if it tries to set the price of heating oil in winter too low, there is a (costly) disconnect between target announcements and market reality, an erosion of confidence in fed target, and therefore a loss of long-run power. The board is powerful to the extent only that it is accurate, and it is accurate only to the extent that it works with the market, rather than against it. Imagine a long-run market path of short-term interest rates; call it the risk free rate of return. The fed can set market rates to this level plus some time dependent random value s.t. its long-run mean is zero.

    I don't see how NGDPLT is any significant departure from current policy. We might acknowledge (I think) that the fed cannot set a production target (let alone a production level), except through second-order effects (e.g. the board cannot order McDonald's to make more, and you to eat more, hamburgers). All that is left, then, is the price of production. The fed cannot set this either, but it can target it through the currency markets. Which it does already; we just call it inflation targeting. The difference only becomes interesting in the case of supply shocks; inflation targeting supposes that the central bank will make the shock worse in the short run, with the benefit of long-run price stability. NGDP targeting assumes that the fed will make prices worse in the short-run, with the benefit of long-run NGDP stability. In the long run both economies converge to the same equilibrium (assuming neither board is error prone). The question of which maximizes aggregate utility is not obvious to me (how does the board raise aggregate demand without lowering prices? presumably through its own consumption, which has no first-order household benefit - we would have to explore whether the additional short-run household income from the demand support is enough to offset the welfare loss from higher prices). But if we accept the starting proposition that most shocks are demand-side, the argument is moot. I don' think that's a particularly controversial position.

    1. Thanks for the input, Glenn. Good points.

      I think the level targeting part gets us much of the way there, the target adjusted according to market expectations gets us a little farther, and NGDP in place of inflation is the final piece. So, I agree that we can get some of this with the inflation target. But, I would reverse the burden of proof. What does inflation targeting get us that NGDP targeting doesn't? If we live in a world with Inflation Targeting and we are talking about demand shocks, isn't that part of the problem?

      Also, for instance, investors worry about changing profits. If corporate profits are expected to rise 8% next year, investors don't typical stress out much about whether that is 4% real or 6% real. But, if nominal profit growth expectations suddenly fall to 4%, investors will react poorly, whether that's real or inflation. It's nominal activity that creates dislocations.

  5. Glenn, you write:

    > I don't see how NGDPLT is any significant
    > departure from current policy.

    It is a significant departure in so many ways...

    1. NGDPLT would be a single target. No more of this targeting-inflation-and-unemployment-in-some-combination. Today, the market can never tell whether the Fed is committed to inflation, unemployment, or neither. For all we can tell, the Fed wants to raise rates just because it wants policy to be more "normal", because Bill Gross is "uncomfortable" with interest rates "too" low. With NGDPLT, the Fed's commitment would suddenly be clear and unambiguous.

    2. NGDPLT would be a level target. No more missing low 25+ quarters in a row, like the Fed has been missing its supposed 2% inflation target over and over. The market could count on NGDP to hit its trend path eventually, whereas under current policy, we can't count on anything; it now appears that the price level is permanently depressed compared to trend because the Fed targets change-in-price-level (inflation) rather than the price level itself.

    3. NGDPLT would mean an end to interest rate targeting. Right now, the Fed targets inflation by using a Taylor rule to set an interest rate target, and then conducts OMO's to hit that interest rate target. With NGDPLT, the Fed would target NGDP directly, using a prediction market to guide OMO's, possibly simply by pegging NGDP futures.

    4. NGDPLT would mean an end to the problem of the ZLB. We've been at the ZLB for six years, resulting in a breakdown of interest rate targeting and a muddle called "QE". NGDPLT would mean the end of QE. With no interest rate targeting, there is no need to take the ZLB into account when setting monetary policy.

    More broadly, I feel you are making two mistakes:

    1) you are confusing the real and the nominal.

    2) you are neglecting the trouble caused by the Fed's dual mandate.

    Despite your objections, there is no reason in the world the Fed couldn't target the nominal price-level of oil. I assert that if it committed to a level-path of oil prices, and let go of its other targets, it would succeed. If it set its oil price target "too low", it would suck money out of the economy until the entire price level fell (deflation) such that the supply and demand curves for oil would intersect at exactly the nominal price of the Fed's choosing. The Fed is the master of the nominal economy. Yes, this would be a total disaster, but it could do it.

    Note that nominal price-level targeting is entirely different from Soviet-style price fixing, where the government truly sets a price, making transactions at other prices illegal. In the Soviet world, setting a price too low simply causes the commodity to immediately run out, or moves activity to the black market. When the Fed targets prices, it doesn't make any transactions illegal. Markets still set prices. The Fed "merely" creates and destroys dollars as needed so as to hit its target.

    It would make no sense for the Fed to target quantity of hamburgers consumed, because that is a real quantity, not a nominal. Like you said, the Fed can't make anyone produce or consume hamburgers. But I guarantee you 100% that if the Fed abandoned its oil price level target and adopted a hamburger price level target instead, it would hit it; likewise for any single nominal aggregate. This is why I object to the Fed targeting unemployment. The Fed can't make anyone hire anyone else. The Fed should provide nominal stability, by stabilizing the level-path of NGDP, and let the market sort out the details.


    Kenneth Duda
    Menlo Park, CA