Nominal GDP targeting: a simple rule to improve Fed performance.

• Author: Sumner, Scott B.

The history of central banking is a story of one failure after another. This record does not mean that our actual monetary regimes have been the worst of all possible regimes--far from it. But it does mean that we can improve policy by learning from experience. Every proposed reform is a response to a previous failure, an implicit display of lessons learned.

A big part of this story has been the search for a robust monetary system that could produce good outcomes under a wide variety of conditions, without having to rely on a central bank run by a benevolent and omniscient philosopher king. It is a search for a monetary rule that can provide the appropriate amount of liquidity to the economy, under widely differing conditions. In this article, I argue that the optimal monetary rule is a nominal GDP (NGDP) target, or something closely related. To understand the advantages of this approach, it helps to see how the theory and practice of central banking have changed over time--that is, to see what went wrong with some previous monetary regimes, and how past reformers responded to those failures.

The Gold Standard

It is not hard to see why gold and silver were used as money for much of human history. They are scarce, easy to make into coins, and hold their value over time. Even today one finds many advocates of returning to the gold standard, especially among libertarians. At the same time most academic economists, both Keynesian and monetarist, have insisted we can do better by reforming existing fiat standards.

It is easy to understand this debate if we start with the identity that the (real) value of money is the inverse of the price level. Of course, in nominal terms a dollar is always worth a dollar, but in real terms the value or purchasing power of a dollar falls in half each time the cost of living doubles. During the period since we left the gold standard in 1933 the price level has gone up nearly 18-fold; a dollar today has less purchasing power than six cents back in 1933. That sort of currency depreciation is almost impossible under a gold standard regime; indeed the cost of living in 1933 wasn't much different from what it was in the late 1700s. This long-run stability of the price level is the most powerful argument in favor of the gold standard.

The argument against gold is also based on changes in the value of money, albeit in this case short-term changes. Since the price level is inversely related to the value of money, changes in the supply or demand for gold caused the price level to fluctuate in the short run when gold was used as money. Although the long-run trend in prices under a gold standard is roughly flat, the historical gold standard was marred by periods of inflation and deflation. (1)

Most people agree on that basic set of facts, but then things get more contentious. Critics of the gold standard like Ben Bernanke point to periods of deflation such as 1893-96, 1920-21, and 1929-33, which were associated with falling output and rising unemployment. This is partly because wages are sticky in the short run (see Bernanke and Carey 1996; Christiano, Eichenbaum, and Evans 2005). Supporters point out that the U.S. economy grew robustly during the last third of the 19th century, despite frequent deflation and a flawed banking system that was susceptible to periodic crises. They note wages and prices adjusted swiftly to the 1921 deflation, allowing a quick recovery. Countries with more stable banking systems, such as Canada, did even better. The big bone of contention is whether the Great Depression should be blamed on the gold standard or meddlesome government policies (see Cole and Ohanian 2004). My own research suggests the answer is "both" (see Silver and Sumner 1995).

I do see some weaknesses in the arguments put forth by advocates of the gold standard. It is true that some of the worst outcomes were accompanied by unfortunate government intervention, particularly during the 1930s (see Cassel 1936 and Hawtrey 1947). However it is worth pointing out that governments also intervened during the classical gold standard in the period before World War I.

Advocates of gold often base their arguments for gold on the assumption that it's dangerous to give the government control over money. They claim it is much easier and more tempting for governments to debase fiat money, as compared to gold coins. That's true, but it doesn't mean that a gold standard prevents meddlesome governments from creating instability in the short run, as in the 1930s. For instance, during the interwar years major countries such as the United States and France often failed to adjust their money supplies to reflect changes in the monetary gold stock.

Here is how I see the debate today. Advocates of gold correctly claim that a gold standard will tend to preserve the value of money over long periods of time, and will sharply reduce the ability of governments to extract wealth from savers. Critics are right that a real-world gold standard is likely to deliver unacceptably large short-term fluctuations in the price level. I think they are also correct in assuming that wages are much stickier than they were during the gold standard's heyday, and that the sort of deflation that led to just a brief surge in unemployment during 1921 (when wages quickly adjusted downwards) might now lead to unacceptably high and persistent unemployment rates. (2) A classical gold standard could probably do considerably better than the sort of regime we had between the world wars. However, if we could count on the authorities to accept the discipline of such a standard, why not make them adhere to a monetary rule to stabilize inflation or the growth of NGDP?

Obviously this debate could go on to look at all sorts of political models of policymaking. Instead, I will focus on purely technical issues and sketch out what I think are the pros and cons of various fiat money regimes, and leave for others the public choice issues of whether such regimes are politically feasible. However, I will return to politics at die end, when I argue that NGDP targeting would help avoid many extremely counterproductive government interventions in nonmonetary aspects of the economy. There are good reasons why many economists with libertarian leanings, including Friedrich Hayek, have embraced some version of this policy target (see Selgin 1995 and White 2008). (3)

Money Supply Targeting and the Taylor Rule

In the United States, gold was phased out in two steps: (1) domestically we left the gold standard in 1933, and (2) internationally the last links were broken in the late 1960s and early 1970s. What followed was a period of very high inflation, which led to renewed interest in finding some sort of anchor for tire price level. Between 1979 and 1982, Paul Volcker was seen as leading a "monetarist experiment" trying to control inflation by reining in tire money stock.

Contrary to the belief of many economists, the Fed never really adopted the sort of rigorous money supply rule that had been advocated by Milton Friedman (1968) and other monetarists. Even during the early 1980s there was significant variation in the money supply growth rate. The problem is that monetary velocity--that is, the ratio of nominal GDP to the money stock--also seemed volatile, especially in the wake of the so-called monetarist experiment. That is not to say that Volcker's experiment was a complete failure; he did break the back of double-digit inflation, and by doing so proved that monetary policy rather than fiscal policy (which was expansionary under President Reagan) was the key determinant of inflation.

Like central bankers everywhere, Fed policymakers greatly prefer to target interest rates, not the money supply. So once inflation was brought down to relatively low levels, they went back to targeting the federal funds rate. But memories of the Great Inflation of 1966-81 led many economists to look for a policy rule that would prevent a recurrence of high inflation. John Taylor proposed a rule for adjusting the fed funds target in such a way as to keep inflation near 2 percent and output as close to potential as possible, reflecting the Fed's dual mandate. The key insight was that as inflation rose above target, nominal interest rates had to be raised by more than one for one with inflation, assuring that even real interest rates were higher than before.

It is hard to overstate the importance of the Taylor Rule. In America, Paul Volcker and Alan Greenspan were feted as heroes who had adeptly steered the economy into the Great Moderation, the period of relative stability between 1983 and 2007. In fact, there was no miracle. All of the foreign central banks that operated under somethig like the Taylor Rule also achieved success in bringing inflation down to low and stable levels. It may be politically difficult to bring down the rate of inflation, especially when contracts have been negotiated on the assumption that high inflation would continue. But once this is done, it turns out to be very easy to prevent a recurrence of high inflation. Just promise to raise nominal interest rates by more than any increase in the inflation rate, until you are back on target.

Obviously something went wrong after 2007 (or maybe even before). (4) If the Great Moderation had continued, there would be little reason to abandon the Taylor Rule. But before we consider alternatives, let's discuss what did not go wrong with that rule; high inflation did not return. Over the past five years the CPI (even including food and energy prices) has risen at the slowest rates since the mid-1950s, barely over 1 percent per annum. (5) Instead, the problem since 2007 has been a severe recession and accompanying financial distress.

Robert Hetzel (2009, 2012) makes a distinction between the "market disorder view" and the "monetary disorder view." Although the market disorder view is the conventional wisdom, the fact...