Wednesday, April 5, 2017

Anatomy of a tweet

Lars Christensen's tweet about the slight fall in 10-year interest rates is a thing of beauty, if by beauty one means the smug security of a circular argument:

Anatomy of a tweet

I thought his qualitative circular hand-waving theory that lacks the capacity for even an order of magnitude estimate for observable variables would be an excellent counterpoint to my update of the quantitative 10-year interest rate model forecast (where the performance has been tracked publicly since 2015):

Per the infographic above, I guess this represents signalling of my own ...


  1. Interesting post. I hesitate to take on someone with vastly superior math and other data analytic skills, but I'd like to defend market monetarism. I think one of the limitations of many of the qualitative descriptions, implied or explicit, is that the expectations components can be difficult to describe formally in a comprehensive way. Where market monetarists see the silhouette of expectations of money supply growth, you see hand waving.

    This, of course, does not mean the market monetarists are wrong, but that more formal modeling and testing against the data is necessary. I do not share the views of many critics that the model has been demonstrated to be wrong or that it isn't truly falsifiable.

    The safest statement I can make is that I don't see how the possibility that expectations of timid central bank action can explain the relative impotence of monetary policy at the zero lower bound has been eliminated.

    Further, perhaps due to my own cognitive limitations, I cannot get past the logic that the expectations for a permanently expanded money supply would lead inevitably to higher nominal output than would occur otherwise, ceteris paribus, both when inflation is below 10% and when interest rates are near or below zero.

    The more dollars there are expected to be versus output they can purchase in the future, the lower the value of each dollar versus that output. This would seem to be the closest thing to a natural law in a social science.

    1. I have some issues with any theory that involves expectations that are related to the actual future (my more formal argument is here).

      But I think there's a more generic argument that essentially reduces market monetarism to a "no true Scotsman" argument.

      Imagine some central bank (say, Scotland) targets some higher inflation rate that it's currently experiencing (say, 2%). The market monetarist then says (according to the theory) the Bank of Scotland will reach 2% inflation.

      Now let's say that doesn't happen. The market monetarist answer is that the BoS didn't actually want to achieve 2% inflation (it wasn't a true Scotsman). All kinds of post hoc rationalizations are made (2% was just an upper bound, other BoS statements sounded hawkish). These should have come out with the original theorizing. But then if you said all these things with the original theorizing, it boils down to this:

      The BoS will achieve 2% inflation unless it doesn't.

      This is to say that forecasts made with market monetarism are conditional on the forecast being correct.

      Now I understand all the supply and demand arguments with money, and there are information equilibrium models you can build that are basically the quantity theory of money (e.g. P:N⇄M which says log N ~ k log M and log P ~ (k-1) log M).

      However the quantity theory of money appears to be an effective description of an economy when inflation ~ 10% or more.

      So if we're talking about inflation of 2% or (heavens, no!) 4%, we're not really in a "quantity theory of money" regime and other effects are much stronger such that the relationship between money and output or inflation is of secondary importance.

    2. In response to your formal argument, I could make many comments. I think though it's most relevant to say that over finite time frames, some possible current distributions are more probable than others, and this is increasingly true, the shorter the time frame. Correct me if I'm missing something.

      On your "no true Scotsman" argument, I think it's unfair to characterize market monetarist claims about central bank timidity this way. Many central banks in the developed world today have many things in common that lead to tighter than optimal monetary policy, especially near or below zero target rates.

      For one, they almost all use inflation-targeting regimes that depend upon signaling via interest rate targets, without level targeting. To market monetarists, the interest rate is not the price of money. Output is.

      Then, you have decisions by committees, which usually include hawkish members with some uncertainty as to the membership composition, often even over just a few years.

      And I would point out that many market monetarists don't only claim monetary policy is usually too timid near ZLB, but also is typically too timid in general. Many market monetarists believe that most recessions, that is, those caused by nominal shocks, were entirely avoidable.

      So, market monetarists don't just offer post-hoc stories about central bank timidity. There is a robust model there that is not entirely discernible from Keynesian and New Keynesian perspectives.


Comments are welcome. Please see the Moderation and comment policy.

Also, try to avoid the use of dollar signs as they interfere with my setup of mathjax. I left it set up that way because I think this is funny for an economics blog. You can use € or £ instead.