Thursday, 23 March 2017


Taylor, Sumner and Selgin at the Cato Monetary Conference

January 7, 2016

In this post I want to briefly comment on three points raised by J. B. Taylor, George Selgin, and Scott Sumner. Though these points have been raised before in the literature, they are certainly worth reviewing.

J. B. Taylor delivered his lecture on the challenges of monetary policy in an international context. The first challenge, of course, is that the strategy, or policy decision, of a major central bank affects the decision making of other major central banks. This could result in unintended loose policies at the international level as central banks around the world react to an expansionary policy by a major central bank like the Federal Reserve. Say, for instance, that after 2001 the Fed would have decided to reduce the federal funds rate target and expanded the monetary supply. In response, a major trade partner like China might have decided to peg its exchange rate to the U.S. dollar in order to avoid the effects on its trade with the U.S. To do this, China would have to mimic the Fed’s policy. The international effects of the Fed’s policies are certainly significant. It is worth noting that two largest crises in Latin America happened after the two largest deviations by the Fed from Taylor’s rule (here). But to be conscious of these issues does not mean that the solution is easy. Leith and Wren-Lewin (2009) show that when assuming open economies, the Taylor rule may be indeterminate or produce spill over to other economies.

Selgin draws a distinction between monetary policy rules and pseudo-rules. A pseudo-rule is a rule that fails to accomplish its objective because there is no reason for the monetary authority to actually observe the rule and hit a given target. The reason why a monetary authority may fail to observe a given rule is because no penalty is awarded if the given target (i.e. an inflation target) is not achieved.

This principle should make sense.

If the CEO of a company fails to achieve the objectives set by his firm’s owners and investors, he can expect to lose his job. But this is not always the norm in central banking. The “CEO” of a central bank may suffer no penalty for failing to follow a rule, which is tantamount to his mandate. Certainly there are reasons beyond the monetary authority that might result in failing to achieve the given target. Real shocks, international political and economic events, other policy makers’ behavior, are all examples of unexpected effects that a central banker is unable to predict. Similarly, there are reasons beyond the CEO of a firm’s strategy that might explain why he might fail to achieve a target. Just as any fail should not be excusable, not any fail in achieving a target is necessarily the monetary authority‘s fault. This doesn’t mean that the monetary authority should go unchecked, but rather that the rules imposed and targets chosen must be savvy and immune to discretion.

Finally, Sumner says that the Federal Reserve communications are not clear enough. What does it mean, exactly, when the Fed says that monetary policy has been tight or loose? To summarize his argument, it can be misleading to reason just from a change in interest rates. A fall in interest rates is associated with a loose monetary policy because it is associated with a shift of the credit supply. But it can be the case that credit demand has actually fallen. Sumner argues this is likely the case after 2008. He goes on to point to a similar scenario that was labeled as tight during the Great Depression that is now labeled as loose. A way to think of a loose monetary policy is when money supply is expanded beyond money demand producing a monetary disequilibrium. This excess of money supply can be channeled through the credit market (as is usually the case in the U.S.) or through government spending (think of Argentina and Venezuela in recent years.) A tight monetary policy can be understood as the opposite scenario. Sumner suggests looking at NGDP, a proxy of total nominal spending. In particular (as I understand it,) a deviation from a stable NGDP trend means that nominal spending is exceeding (loose) or falling short (tight) with respect to expected nominal income.