Monetary Rules: Solving the Knowledge Problem

Hayek, knowledge problem, monetary policy, monetary rules

Hayek, knowledge problem, monetary policy, monetary rulesIn its  "Free Exchange" column, the Economist recently took up the issue of monetary rules.  Provocatively titled “Rule It Out,” the column announced that “setting interest rates according to a fixed formula is a bad idea.”

Reading the column one quickly learns the author doesn’t understand what constitutes a rule, and what the argument for a rule is.  The column moves from a general consideration of monetary rules to considering specifically the Taylor Rule.  I leave it to Professor Taylor to defend his rule, which he did on his blog.  I, however, consider the general case for monetary rules.

"Free Exchange" links the case for rules to the 1977 article by Finn Kydland and Edward Prescott, “Rules Rather than Discretion: The Inconsistency of Optimal Plans.”  The title is not cited nor is the article’s central argument addressed: Discretionary economic policy cannot be optimal.  Their article undermines the case for discretion over rules that "Free Exchange" attempts to make.  "Free Exchange" lamely says that Kydland and Prescott’s argument “helps to explain the high inflation of the 1970s.”  It did far more than that.  It explains why policymakers cannot credibly commit to future policies.  That is known as the time inconsistency problem.

The argument for rules versus discretion in monetary policy goes back at least to Henry C. Simons' 1936 article, “Rules versus Authorities in Monetary Policy.”  The case for a monetary rule was re-argued by Milton Friedman in the 1960s and he anticipated the dynamics developed in Kydland and Prescott.

"Free Exchange" states that “monetary policy based on rules has one major advantage: transparency.”  Certainly a rule will likely be more transparent than policy discretion.  But it has never been the central argument for a monetary rule.  The central argument for a monetary rule is what is known as the knowledge problem in economics and social affairs.

Policymakers cannot in principle possess the knowledge required to devise an optimal (or time consistent) monetary policy.  The information required for centralized policymaking is dispersed among the millions of actors in society.  It cannot be aggregated or concentrated in one mind.  No expert or set of experts can ever know as much as the totality of individuals in society.

Rules are a response to the knowledge problem.  Uncertainty generates reliance on rules.  Rules are constructed or evolved based on accumulated experience over long periods of time.  Rules encapsulate the totality of knowledge and experience not only of all alive today, but also of those who preceded them.

Rules can be formal or informal, and could – but need not – be a formula.  (Most rules are not a formula.)  When people do not possess the information required to optimize, they rely on rules.

The knowledge problem arises in any attempt to set centralized policy or planning.  Even before Simons, F. A. Hayek articulated the knowledge problem in monetary policy and in the debate over centralized economic planning.

"Free Exchange" turns the knowledge problem upside down: “Until the day the economy is fully understood, human judgment has a crucial role to play.”  No, actually, it is just the opposite.  There would be no need for reliance on a rule if the economy were fully understood.  The less we know about the specifics of a situation, the more we must rely on rules.  A good rule incorporates the general features of a class of situations, in which the specific features vary unpredictably.  If we possess full information, why would we want to rely on a rule?

In a paper that I will present at Cato’s annual monetary conference on November 12th, I develop in more depth the case for rules in monetary policy.  I attribute the central argument to Hayek.  But I note that Friedman also adduced an argument based on the knowledge problem in support of his monetary rule.

Monetary rules and policy rules more generally are a subset of behavioral rules.  The case for a monetary rule is ultimately the same as the case for the rule of law in society.  For those would like to see that argument in print, along with the philosophical tradition undergirding it, see my recent article in the Journal of  Private Enterprise, “Hayek and the Scots on Liberty.”

  • Kenneth Duda


    I realize this article is about the idea of monetary policy rules in general, … but, I am curious if you have ideas for what specific rules would be suitable. I'd be interested to read a future post on that topic.

    (FWIW, I'm a big supporter of NGDPLT, the idea of the Fed conducting OMO to target a pre-announced desired time-path of NGDP, using a prediction market to gauge what volume of OMO's are needed to hit its target).


    Kenneth Duda
    Menlo Park, CA

    • Geral O'Driscoll

      I wanted to make the general case for rules and not to push a particular rule. I hoped to start a discussion of alternatives. An important question I didn't bring up is where should a rule come from? A model? Historical experience (e.g., arguably the Taylor Rule's origin). Evolution (e.g., the gold standard)? I think the meta-questions need to be answered first.

  • Pingback: Links for 08-26-15 | Finance()

  • Pingback: Links for 08-26-15 | Economics Blogs()

  • ThaomasH

    I echo Duda. I'd even ask what's so wrong with a PL target, if the Fed would actually carry it out?

  • joebhed

    I look forward to the publication of your thoughts on Money Rules(n) this November.
    Also appreciate your historical references here to the works of Simons' and Friedman thereon, with my qualifier that they were both advocates of a "Money quantity" rule, one that had nothing to do with "an interest rate policy Rule to (theoretically) control the money quantity."
    Economist : ""setting interest rates according to a fixed formula is a bad idea."”

    It is that error that leads to Monetarism becoming an interest rate policy discussion, rather than a money quantity policy discussion….
    as in Turner's Overt Money Finance proposal.


    • Gerald O'Driscoll

      One by one, almost all monetarists abandoned money-supply rules over time because of the breakdown in the relationship between money and prices, or money and nominal GDP. So, yes, Friedman had a rule for money growth. Today, however, with just a few exceptions, even the most stalwart monetarists have abandoned efforts to devise a money supply rule. I had dinner with one this past week, and that was his position.

      • joebhed

        Thanks, Gerry.
        Yes, I agree. To a point.
        Until Friedman published his "A Program for Monetary Stability" in the mid-90s, in which he reiterates his original plan – from his 1948(?) " A Monetary and Fiscal Framework for Economic Stability" – many claimed he had abandoned his ONLY true 'monetarist' position.
        But today, there's a whole movement, including populists (PositiveMoneyUK), money technocrats like Lord Adair Turner and Dr. Michael Kumhof, and traditional reformers like the American Monetary Institute and the Kucinich proposal,

        that have taken up the mantle of true public money issuance . a new awakening to what is indeed 'monetarism', but without the title and certainly without using money-cost to determine money-supply, when direct issuance eliminates most all problems associated with transmission and 'lag'..
        For the Money System Common.

        PS, give your dinner guest friend my comment.

  • Pingback: Hilsenrath’s Take: Dudley Sets a High Bar for September Fed Rate Move – Wall Street Journal | Creative Newspaper()

  • Pingback: Some Links()