This archived content originally appeared at, the predecessor site to, and does not carry the sponsorship of the Cato Institute.

Will Luther on Sound Money

Over at Atlas's Sound Money Project, Cato Adjunct Scholar (and Larry White student) Will Luther tells us what "Sound Money" means to him. A sample:

For me, the important aspects are the ability to facilitate exchange—which any money will do but some might do better than others—and the degree of macroeconomic stability enabled by that money. When comparing monies along these margins, then, we must establish a benchmark. In what follows, I’ll try to establish a benchmark for the second of those aspects: namely, the degree of macroeconomic stability enabled by a money.

Read on to discover just what Will means by "macroeconomic stability." I find his approach convincing–but I suppose I might be a tad biased.

  • George, I don't get it:

    > Likewise, P conveys important information about the relative scarcity
    > of goods transacted over time. If we become more productive, and the
    > value of transactions T increases, the price level P should decrease to
    > reflect that goods, in general, are less scarce today than in the
    > not-so-distant past.

    Why is information about relative scarcity "over time" so "important"? I get how prices are a useful signal — if apples become cheaper and bananas become more expensive today, market players can find innovative ways of using apples instead of bananas, which is great. But that doesn't work "over time" at all. If I know apples today are more expensive than apples 10 years ago, what am I supposed to do with that information? Buy more apples from the past and conserve apples in the present? That's a silly suggestion of course, but is there something serious that one could actually do with this information?

    I feel like the important thing to do with P is stabilize it, so that people can write long-term contracts (loans, supply agreements, employment contracts) in nominal terms without wanting to shoot themselves three years later, and so nominal rigidities don't prevent markets from clearing.


    Kenneth Duda
    Menlo Park, CA

    • George Selgin

      The information "value" of trend deflation is admittedly not the same as that for short-run deflation (as when there's a positive productivity shock). Still, it's there. A producer has more need for help in forecasting input prices, which are determined in other markets, than in forecasting the prices of output, and his or her own output in particular. (Most other output prices are of relatively little relevance to the producer's planning.) Stability of those input prices therefore contributes more to easing the producer's calculation and forecasting burden than stability of output prices. And remember, with productivity growth, you have to stabilize one or the other; "both" isn't an option.

    • will.luther


      Thanks for considering the ideas in my post. The information value argument is just one reason among a few that, if I recall correctly, George addresses in his book Less than Zero. Another, potentially more persuasive, reason follows from reflection on what would be required to stabilize the price level in the face of productivity growth.

      Suppose a cost-saving technology is adopted by an automobile manufacturer. Initially, the automobile manufacturer is the only person in the economy who knows of this new cost-saving technology. The automobile manufacturer offers his car at a lower price than his competitors to gain market share and, in doing so, conveys that there has been technological growth and a corresponding change in relative scarcity to everyone else. Nothing new here that isn't in Hayek's Use of Knowledge.

      But consider what happens to the price level. Suppose initially the price level was P1, the average of Pcars, Pboats, Papples, Pi… Now, since the price of cars has fallen, and no other prices have changed (except insofar as they are substitutes or complements for cars), the aggregate price level falls to P2 < P1. Stabilizing the level of prices at P1 would require raising the prices of all goods (cars, boats, apples, etc) to some new level P3 where the relative price ratios included in P3 is identical to those of P2 but the level of prices in P3 is equal to P1. In terms of conveying information about the relative scarcity of goods and services in the present period, P2 and P3 are equally good. The difference is that, to obtain P3, we have to change a lot more prices. If it is costly to adjust prices–which seems reasonable to assume–we should surely prefer P2 to P3. They both covey the relevant information but the latter comes with a higher cost.

      Take care,


  • A free banking system does not actually give us the “macroeconomic stability” which Will Luther says is so important. Nor for that matter does an “unfree” system: e.g. a system where private banks are barred from issuing their own dollar bills.

    Reason is that while the above systems issue the amount of money that people want, the amount that people want will not necessarily be the amount that brings full employment. Also those systems do not have a way of stabilising the velocity of circulation of money. Thus they don’t stabilise aggregate demand. Hence they don’t give us “macro economic stability”.

    The factors which IN THEORY stabilise AD at the full employment level are the Pigou effect and Say’s law. But clearly neither of those work too well. That’s why (as advocated by Keynes) we have the state issuing another type of money (i.e. base money) and attempting to vary and amount issued (e.g. via QE) so as to give us macroeconomic stability at the full employment level.

  • Troy Walters

    Does the Coase Theorem justify use of a Taylor rule (even if the rule turns out to be wrong im some sense)?

  • Matt Young

    let P be a ratio in a discrete variable. Define T to be P', the differential. Discrete time is an outcome, the term length along the yield curve. Then MV = P"/2, half the second differential.
    So PT= MV becomes, P*P'+ P"/2 = 0.

    For T, real transactions, value is defined as the change in P per transaction (gain or loss). MV tells you how fast the economy has to adjust before becoming disconnected, it becomes the condition of inventory flow. MV tells us the variation allowed before contango erupts in the distribution network.

    The unknown discrete variable will make summation and differences between terms stay within precision of the continuous differential version of the equation. This is Ito's calculus, look it up.

    Mathematicians have this solution, they can do a no arbitrage currency banker in a spreadsheet and support a network of bankers. The currency banker sterilizes (updates) deposit and loan rates, at the short end, when they exceed the constraint equation (by the standard uncertainty).

    Term lengths along the finite yield curve will stabilize and match inventory cycless for real distribution. It is no arbitrage, and for each level of the network, the actual interest rates and term lengths will vary proportionally by a small uncertainty and be known only after rates have been sterilized.