I’ve recently done a number of posts explaining how monetary policy impacts prices in the long run. The basic approach might be called the “hot potato model.” People have a certain demand for non-interest bearing money. When the Fed increases the supply of base money, people try to get rid of excess cash balances. Individually they can do so, but collectively they cannot. The paradox is resolved by the fact that when people try to get rid of excess cash balances, prices rise until the public wants to hold those extra cash balances.
Unfortunately, in the real world wages and prices are slow to adjust, and as a result the short run is far more complicated than the long one. In the next post I’ll look at the short run impact of money on asset prices. In this post I’ll look at the short run impact on real output. Here it will be useful to switch from a focus on the price level:
P = Ms/(Md/P)
to a focus on nominal GDP:
P*Y = Ms/k
Where k is the fraction of gross income that the public chooses to hold as base money (k=1/V.) We continue to assume no interest on reserves. For any given k, more money means higher NGDP. And a one-time change in M probably doesn’t impact k in the long run.
Because hourly wages are sticky in nominal terms, a drop in NGDP caused by tight money will lead to a fall in output and hours worked:
The aggregate demand curve should probably be called “nominal expenditure,” as it represents a given NGDP (and hence is shaped like a rectangular hyperbola.)
Tight money leads to lower NGDP, which reduces output and employment. This graph also provides one explanation for why money is non-neutral in the short run; a change in M leads to a change in output, not just prices. In the next post we’ll see that M also affects k (i.e. velocity) in the short run. The P/Y split is determined by the slope of the SRAS curve, which reflects the degree of short run wage/price stickiness.
In the long run wages and prices adjust, and hours worked/output return to the natural rate. Of course like any macro model, it simplifies certain aspects of reality. For instance, during a depression investment may be postponed and workers may lose touch with the labor market, and hence there may be a permanent loss of output. However I’d argue that the permanent effects are relatively small, as we saw in the strong bounce back after the Great Depression.
Another non-neutrality can occur if workers have “money illusion,” which means they confuse nominal and real wage changes. For this reason, the bell-shaped distribution of wage rate changes has a discontinuity at zero percent—workers are irrationally reluctant to accept nominal wage cuts. Thus very low trend rates of NGDP growth, per person, may lead to a higher natural rate of unemployment.
[Each time I make this point a few commenters try to argue that aversion to nominal wage cuts is not irrational, because of factors like nominal debt obligations. Unfortunately this argument doesn't work unless all of one's expenditures are repayment of nominal debts, which is obviously not true.]
Putting aside these special factors, most US business cycles are a pretty simple phenomenon. Because of excessively tight money, NGDP growth slows relative to what was expected when labor contracts were signed. Because hourly nominal wage growth is very slow to adjust, a sharp slowdown in NGDP growth raises the ratio of W/NGDP, which leads to fewer hours worked and less output. It may take many years for the labor market to fully adjust. (Note: if NGDP had started growing again at 5% in mid-2009, we’d be mostly out of the recession by now. The recovery was slowed by further unexpected (negative) NGDP growth shocks after 2009.)
Think of recessions in terms of the game of musical chairs. When the music stops several chairs are removed, and a few participants in the game end up sitting on the floor. Slow NGDP growth combined with sticky wages is like taking away a few chairs; several unemployed workers end up “sitting on the floor” (i.e. unemployed), as there is not enough aggregate nominal income to support full employment at the existing nominal hourly wage level.
Other variables such as interest rates also move around over the business cycle, but don’t really play a causal role in unemployment. It’s all about NGDP and hourly wage growth.
PS. Mark Sadowski sent me this graph showing the correlation between W/[NGDP/(pop)] and the unemployment rate:
If someone can update through the end of 2012, I’ll replace it with the newer version. Update: Ron M just did:
The final post in this short intro to money course will focus on money and asset prices. Then I hope to develop an online version of the intro to money course.