Thursday, September 17, 2009

Friends Don't Let Friends Mix Say's Law with Money

Nick Rowe recently noted that John Cochrane, in his rebuttal to Paul Krugman, effectively invokes Say's law with this sentence:
Paul’s Keynesian economics requires that people make plans to consume more, invest more, and pay more taxes with the same income.
Cochrane's point is that it is impossible to have total planned expenditures exceed total income. This line of thinking or Say's law, taken to its logical conclusion, implies it is impossible to have a general, economy-wide glut since (1) total income must equal total planned expenditures and (2) total income comes from the production of goods and services upon which total expenditures are spent. (i.e. total planned expenditures = total income = total value of production). Obviously, history has not been kind to this proposition. As Nick notes, this is because Say's law assumes a barter economy and ignores the complications found in a money economy. Here to explain these complications is Leland Yeager (Source: The Fluttering Veil: Essays on Monetary Disequilibrium, p.4-6):
Say's law, or a crude version of it, rules out general overproduction: an excess supply of some things in relation to the demand for them necessarily constitutes an excess demand for some other things in relation to their supply...
The catch is this: while an excess supply of some things necessarily mean an excess demand for others, those other things may, unhappily, be money. If so, depression in some industries no longer entails boom in others...

[T]the quantity of money people desire to hold does not always just equal the quantity they possess. Equality of the two is an equilibrium condition, not an identity. Only in... monetary equilibrium are they equal. Only then are the total value of goods and labor supplied and demanded equal, so that a deficient demand for some kinds entails and excess demand for others.

Say's law overlooks monetary disequilibrium. If people on the whole are trying to add more money to their total cash balances than is being added to the total money stock (or are trying to maintain their cash balances when the money stock is shrinking), they are trying to sell more goods and labor than are being bought. If people on the whole are unwilling to add as much money to their total cash balances as is being added to the total money stock (or are trying to reduce their cash balances when the money stock is not shrinking), they are trying to buy more goods and labor than are being offered.

The most striking characteristic of depression is not overproduction of some things and underproduction of others, but rather, a general "buyers' market," in which sellers have special trouble finding people willing to pay more for goods and labor. Even a slight depression shows itself in the price and output statistics of a wide range of consumer-goods and investment-goods industries. Clearly some very general imbalance must exist, involving the one thing--money--traded on all markets. In inflation, an opposite kind of monetary imbalance is even more obvious.

Also see Brad DeLong's comments on this issue.


  1. Everyone is raving about Krugman's essay on whats wrong with macro. But a better review is provided by David Laidler available at
    He argues that modern macro is simply incapable of thinking about the current crisis by its use of an Arrow-Debreu framework. Although there are now attempts to build financial features into the DSGE framework, like Bernanke and Gertler and others.

  2. In that same essay, Yeager discusses how an excess demand for short term bonds (like T-bills) can shift over to an excess demand for money as nominal yields approach the zero bound. He notes that this may have implications for the implementation for monetary policy.

    In other words, an excess demand for money that is a secondary effect of an excess demand for short term bonds because they have hit (or are near) the zero bound on yields, is unlikely to be corrected by open market purchases of those very bonds. Creating new money and purchasing the bonds that are in excess demand is going to exacerbate that problem by reducing the outstanding stock of those bonds, thus increasing the demand for money more or less in proportion to the increase in the quantity of money. (A monetary disequilibrium account of the liquidity trap, more or less.)

  3. Are you quite sure that really was Cochrane's point? I read DeLong's post too and I find it very hard to believe Cochrane would disagree with either of you that plans to spend are not fixed to real income.

    However, knowing that planned expenditures vary from total income does not necessarily give us license to push for further decoupling. Failure to save at a prudent level creates a more vulnerable economy, one of the many things ignored in Krugman's babysitting fable.

    It seems narrow to me to keep harping on the short-term fact that in a recession increased saving reduces spending. Am I wrong in thinking that too many Keynesians spend too much of their attention on how to prevent unpleasant events, which could be rendered less unpleasant if people had more (dreaded) savings?

  4. Bill: I was fortunate to have a professor, Dr. George Selgin, who made Yeager required reading. I hadn't opened Yeager's book in sometime and only did so after Nick Rowe referred to him in his post. I need to go back and reread the rest of his book.

    Josh: You may be right on Cochrane's view, although it is hard to see how given that sentence. Like Nick Rowe, though, I found the rest of Cochrane's paper a good read. As you know, my view is that this collapse in nominal spending would have been better dealt with by a more aggressive monetary policy rather than more government deficit spending.

  5. Says Josh:
    Am I wrong in thinking that too many Keynesians spend too much of their attention on how to prevent unpleasant events, which could be rendered less unpleasant if people had more (dreaded) savings?

    If the US had a "safety net" such as they have in Europe and Australia, maybe recession prevention wouldnt be such a high priority for the Fed?
    Perhaps this lack of protection for the working man is causing the US economy to be more crisis prone than other countries.

  6. please excuse this long comment to follow. I’m trying to understand this issue and so far am hitting a wall.

    It seems like there are a few potential problems I’ve read about in the way this particular window of time is analyzed by the Keynesians, where people’s demand for higher cash balances (right?) leads them to cut back on spending in general and “purchase money with their output”.

    So let me see if I understand this causal chain.

    #1. Something happens or "Animal Spirits" wane/
    #2. "People" in the aggregate decide to save more, or demand more cash balances
    #3. In so doing, they are effectively producing in exchange for money and not any immediate good beyond the money in the short run.
    #4. Aggregate demand falls, leading firms, again in the aggregate, to cut back on employment.
    #5. Deflationary spiral to depression.

    Here's the problems with this "Saw's law is dead, we need more M or G to fill the gap" as I understand them from my amateur reading.

    Problem #1. "Something Happens" is likely something SPECIFIC. Perhaps it's a combination of negative savings driven in part by a false sense of wealth as a result of an inflationary asset boom caused by the last effort to fill the "gap" with M or G or both. Perhaps it's because of a financial crisis brought on by the inevitable raising of interest rates to curb the prior inflation caused by the last effort to fill the gap with M, resulting insolvency in production projects, defaults on loans for these incomplete projects, etc.

    Problem #2. The continued use of aggregate reference to “people’s preferences” as if they are usefully averaged into one global “propensity to save” or “money demand” or “liquidity preference”. That seems like an oversimplification and over-aggregation to the point of breaking useful thinking about what’s going on in the economy. Surely there are relative degrees of these changes which have actual reasons that are specific. As such there are all kinds of relative adjustments to prices and incentives going on. Relative price changes matter.

    Problem #3. This "hoarding" analysis appears to assume that the money goes into a mattress instead of into a bank which can lend it at increasingly lower interest rates to businesses in pursuit of more productive and perhaps more "roundabout" processes. This investment spending would or at least could "fill the gap". But then, if #1 is not "animal spirits" but instead a case of bank insolvency, this inability of savings "leaks" to become investment spending has a cause that we can understand. That cause, again, is not "animal spirits".

    Problem #4. If there is a nominal price deflation as people's higher propensity to save drives down aggregate demand, isn't it possible that the increasing REAL value of their savings potentially provide what they are potentially trying to do: build up a larger buffer of purchasing power to weather uncertain times?

    I suppose that could be offset by declines in their nominal wage, but I thought that “wages are sticky downward” according the very same school of thought. And, again, the relative prices and wages and the actual real world industrial changes seem pretty important for any understanding here.

    So, it seems like this Keynesian/Malthusian approach to "under-consumption" at best can explain a deflationary spiral after a boom turns to bust but seems to offer no explanation whatsoever for the boom or the change point itself aside from "animal spirits" which in the context of the sequence of events appears to be a completely silly, superficial explanation.

    Or maybe I'm totally confused because I'm trying to teach myself this stuff. Any clarity or insights would be appreciated.

  7. Thanks. My guess (I'm not in his head) is that Cochrane feels it is undesirable that the individual should be impelled to spend at or beyond his total expected income - not that it doesn't happen.

    My concern about Farmboy's point (and Bill's too, maybe) is that "excess demand" for money is hard to determine, and enhancing government protection neither eliminates demand for money (who will completely entrust the next generation's, the next year's, the next week's income and needs to any government?) nor is necessarily more efficient at compensating for emergent situations which justify having saved money. And it comes at the cost of reducing choice... I'm not an idealogue, but whenever I read someone advocating policy to impede savings in a nation where the savings rate is not 30% but close to 0%, I get the shivers.