Monday, January 4, 2016

Scott Sumner's latest information transfer model

It looks like Scott Sumner has written down the models r : N ⇄ M0 and r : N ⇄ MB

He just needs to plot log r.

PS Here was the previous one:


  1. But Scott includes IOR, no? He writes:

    "...and the demand for reserves is negatively related to the market interest rate minus IOR"

    1. IOR is not explicit in the IT model. It could be thought of as implicit. You could say that IOR (and reserve requirements) explains the difference between the currency component of MB and MB.

      I take the difference between M0 and MB to be empirical.

      If IOR and reserve requirements (RR) explain MB - M0, then IOR and RR explain the difference between the short and long term interest rates. If an arrow means "explains" then possibly:

      IOR,RR → MB - M0 → r_long - r_short

      The ITM just uses:

      MB - M0 → r_long - r_short

      And takes MB - M0 as empirical (the same way it takes M0 to explain the price level empirically).

  2. Jason,

    I think, in this case, the title should be "Jason Smith's latest mainstream macro model," since you're model is essentially equivalent to Money-In-the-Utility-Function (MIUF) models, which have existed since the 1990s.


    Including IOR is important because, according to mainstream macroeconomics, the demand for money is decreasing in the cost of holding money relative to other assets. If money earns IOR, then the interest that agents forgo by holding money instead of some other asset with a higher rate of return is less, so they will want to hold more money.

    1. Hi John,

      However there is a major difference: the large deviations in the short term interest rate from the information equilibrium value represent an out of equilibrium market with non-ideal information transfer.

      That is to say it is MIUF-like for finite intervals between recessions, and something else (bounded by the information equilibrium solution) during recessions.

      One of the novel ideas in the ITM is that it allows markets to fail. In a sense, it is like a model where agent utility functions can go from being ordinary real functions of real variables to being undefined.

      If you have a piston filled with a gas, you can compress it and increase the temperature. However, the temperature during the non-equilibrium state between the uncompressed state and the compressed state is not defined (you can have local thermodynamic equilibrium and a local temperature).

      As far as I know, there are no macro models that say macro describes such and such, but says it is not even in principle possible to describe some system states (since they don't have macro properties).

      Another way, sometimes it's a macroeconomy (information equilibrium) and sometimes it's aggregated micro (non-ideal information transfer).

    2. John, re: IOR, that was my understanding. Thanks.


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.