Sumner's Monetary Economics

Scott Sumner has an article in National Affairs defending a target for nominal GDP targeting.

I think he gives a good account of his views, but I didn't find his arguments for 5 percent nominal GDP growth and 2 percent inflation persuasive.

He argues that 2% inflation will allow central banks to better use conventional monetary policy (interest rate targeting) because the zero nominal bound on the target interest rate is a -2% real bound.

He also argues that it is easier to refuse to give workers cost of living raises than to impose pay cuts. This allows employers to cut real wages by 2% per year without explicitly telling the workers that they are getting real pay cuts. (His blog is called The Money Illusion.)

In my view, the best reason to stick to a 5 percent nominal GDP growth path is because that is what we had for 20 years, and shifting to a lower one would not be worth the disruption. The reason I favor such a change now is that we have already suffered massive disruption and when we shift to a new regime, we should get it right.

What about conventional monetary policy? Accommodating the preference of central bankers to stabilize short term interest rates has turned out to be a mistake. If watching short term interest rates provides a good estimate of whether supply and demand conditions for money, (the quantity of money in relation to the demand to hold money,) is consistent with keeping nominal GDP on the target growth path, then perhaps central bankers should take this into account.

But putting all of the monetary policy eggs in that one basket, and insisting that "setting" the current short term interest rate and, more importantly, generating expectations for short term interest rates in the future, is the whole of monetary policy is a proven disaster. Central bankers need to learn that when short term interest rates don't work, they have to go to with some alternative approach. Perhaps a targets for nominal GDP would allow central bankers to shift to looking at a longer term interest rate. Perhaps looking directly at base money or some alternative measure of the quantity of money would help. But in the end, the goal must be to purchase as many assets as it takes to get nominal GDP back to its targeted growth path.

As for the wages, a 3 percent growth path for nominal GDP involves nominal wages being on a 3 percent growth path. In some particular labor market, more rapid increases in labor supply or slower increases in labor demand can slow the growth in equilibrium real and nominal wages. The actual growth path can shift down 3 percent over a year without any absolute decrease in nominal wages. Shifting to a 5 percent growth path of nominal GDP does increase the maximum rate of reduction in growth path given stable nominal wages, but only by an additional 2 percentage points. Of course, that extra 2 percent is a decrease in real wages.

If slower growth in labor demand in some segment of the labor market requires an actual decrease in real wages, and real wages don't decline, then employment in that sector decreases and given the reduced labor input, real output in aggregate shifts to a lower growth path. This implies that with a given target for nominal GDP, the price level moves up to a higher growth path. Real wages do shift down at least slightly. While presumably this reduction in real wages slightly improves condition in the market where real wages need to fall, the larger effect is to enhance profitability in expanding sectors of the economy.

Usually, the appropriate response to a shift in the demand for labor is for workers to abandon contracting industries where real wages are growing less than trend and gain employment in growing industries where real wages grow faster than trend. If real wages must fall to maintain employment in some industry, this almost surely a sign of a needed shift. It isn't entirely clear that making it easier for firms to lower real wages in those circumstances and postpone the adjustment is desirable. Nor is it clear how effective this will be.

It is certainly correct that if there is an excess demand for money, clearing it by having prices and wages shift to a lower growth path is disruptive. Similarly, if there is some economy-wide productivity shock that reduces real wages across the board, having the price level rise to a higher growth path is the least disruptive course. But when what is needed is a shift in the allocation of labor between industries, perhaps layoffs in the shrinking industries matched by more rapid hires in the growing industries is the least bad option.

Rules of thumb that have proven effective in recruiting and retaining a productive workforce in competition with other firms do not apply when there is an excess demand for money or even when there is a adverse supply shock impacting real wages throughout most of the economy. The benefit of these rules of thumb is exactly in situations where there are shifts in demand between firms and industries. Lowering real wages in those situations is apparently problematic for the firms.

And, of course, with nominal income growing, even without inflation, there is room for new firms to enter, hiring workers at lower nominal and real wages. And those firms with growing demand can start their new workers at lower nominal wages. These are not small factors considering the huge turnover in employment in a healthy, growing economy. The very serious problems that develop when there is a need to lower nominal real wages more or less across the board, do not necessarily apply when there is some small segment of the labor market requires lower real wages to maintain current levels of employment.

Finally, the more official and open the commitment to a trend inflation rate, the less we can pretend that the absence of a cost of living pay increase is anything other than a pay cut. Worse, the entire notion that "cost of living increases" are something to be expected is entirely counterproductive when there is an adverse supply shock. Cost of living increases to correct of that sort of inflation greatly reduce the benefit of nominal GDP targeting. In my view, it is better to not get into the habit of cost of living increases than to require firms to explain that they can give the usual 2 percent cost of living increase, but the additional inflation that has reduced your real income can't be compensated because that wasn't part of the normal increase in the cost of living planned by the Federal Reserve.