In 2001 William Niskanen proposed a test of the ‘demand rule’ – a target path for nominal spending (NGDP):

It is important to recognize that a demand rule is consistent with any desired price-level path, including a stable price level. My primary point is that a demand rule is potentially superior to a price rule, whatever the desired price-level path, because of the different response to changes in supply conditions. A central bank following a demand rule would not respond to either positive or negative supply shocks; such shocks would lead to a one-time change in the combination of price and output changes in that year, but would not lead to a long-term change in the inflation rate. A central bank following a price rule, in contrast, would increase the monetary base in response to a positive supply shock and would tighten the base in response to a negative shock, thereby increasing the variance of output. Similarly, a demand rule is potentially superior to a money-supply rule because it accommodates unexpected changes in the demand for money. The general case for a demand rule, thus, is that it minimizes the variance in output in response to unexpected changes in either supply or demand.

He presents a version of this chart:

And argues:

Focus first on the sample period through the first quarter of 1998. It is important to recognize that this demand path was maintained over a period of two quite different shocks. Productivity growth was unusually high in 1992, 1996, and 1997 and unusually low in 1993 through 1995, but the demand path was maintained over this large change in productivity growth rates. And the demand path was also maintained over the Asian financial crisis of 1997. Figure 1 also illustrates the effects of the major tightening of monetary policy in late 1994, a tightening that led to a temporary weakness in demand in 1995 but maintained the demand path. In retrospect, this was an extraordinarily successful period of monetary policy, marked by a very steady growth of demand and a continued decline in both unemployment and inflation. Focus next on the very different record since the sample period. From the first quarter of 1998 through the second quarter of 2000, demand increased at a 7.3 percent annual rate to a point that was nearly 3 percent above the demand path established during the sample period. In 1998, the Federal Reserve appears to have responded to the near collapse of Long-Term Capital Management, the Russian default, and the Brazilian financial crisis with monetary stimulus that led to significantly higher demand growth, a rate of growth that the interest rate increases in 1999 did not constrain until mid-2000. This led to an increase in the consumer price inflation rate by about 2 percentage points relative to the 1998 rate, and there was no realistic prospect that this higher rate could be reduced without reducing demand growth.

As the future showed, that action by the Fed caused instability with demand falling far below trend through mid-2003. By early 2006, just as Greenspan passed the baton to Bernanke, spending was almost exactly back on trend.

But Bernanke completely ignored the “demand rule” and concentrated on inflation. The result was a monetary policy that was so bad it managed to recreate the conditions that prevailed during the “Great Depression” when nominal spending was allowed to crash. This time around the crash was much less deep because Bernanke, remembering all his “Great Depression” research, did not allow the “blow-up” of the financial system.

Advertisements