W P

F-\Y) F-HY) L

FIGURE 5.13 A competitive labor market when prices are sticky and wages are flexible

labor supply curve (to Point E in the diagram). If labor supply is relatively umesponsive to the real wage, the real wage varies greatly when aggregate demand changes.

Finally, this model implies a countercyclical markup (ratio of price to marginal cost) in response to demand fluctuations. A rise in demand, for example, leads to a rise in costs, both because the wage rises and because the marginal product of labor declines as output rises. Prices, however, stay fixed, and so the ratio of price to marginal cost falls.

The cyclical behavior of the markup has received much less attention than the cyclical behavior of the real wage. It plays an important role in many of the models of this chapter and the next one, however. Because of Its importance to theories of fluctuations, it has begun to be the subject of intensive study. The evidence to date seems inconsistent with the view that the markup is strongly procyclical; whether it is significantly countercyclical or approximately acyclical, however, is an open question.

The reason that incomplete nominal adjustment causes changes in aggregate demand to affect output is quite different in this case than in the

See, for example, Bils (1987); Rotemberg and Woodford (1991); and Chevalier and Scharfstem (1994). Kalecki (1938) was an early advocate of the importance of the behavior of the markup for fluctuations.

Case 3: Sticky Prices, Flexible Wages, and Real Labor Market Imperfections

Since output fluctuations appear to be associated with unemployment fluctuations, it is natural to ask whether movements in aggregate demand can lead to changes in unemployment when it is nominal prices that adjust sluggishly. To see how this can occur, suppose that nominal wages are still flexible, but that there is some non-Walrasian feature of the labor market that causes the real wage to remain above the level that equates demand and supply. Chapter 10 investigates characteristics of the labor market that can cause this to occur and how the real wage may vary with the level of aggregate economic activity in such situations. For now, let us simply assume that firms have some "real-wage function." Thus we write

= w(I), wC) > 0. (5.29)

For concreteness, one can think of firms paying more than market-clearing wages for efficiency-wage reasons (see Sections 10.2-10.4). As before, prices are fixed at P, and output and employment are related by the production function, Y = F(I).

These assumptions, like the previous ones, imply a flat aggregate supply curve up to the point where marginal cost equals P; thus again changes in aggregate demand have real effects. This cases implications for the labor market are different than the previous ones, however. This is shown in Figure 5.14. Employment and the real wage are now determined by the intersection of the effective labor demand curve and the real-wage function. In contrast to the previous case, there is unemployment; the amount is given by distance EA in the diagram. Fluctuations in labor demand lead to movements along the real-wage function rather than along the labor supply curve. Thus the elasticity of labor supply no longer determines how the real wage responds to aggregate demand movements. And if the real-wage function is flatter than the labor supply curve, unemployment rises when demand falls.

previous one. A fall in aggregate demand, for example, lowers the amount that firms are able to sell at the prevailing price level; thus they reduce their production. In the previous model, in contrast, a fall in aggregate demand, by raising the real wage, reduces the amount that firms want to sell.

This model of aggregate supply is important for three reasons. First, it is the natural starting point for models in which nominal stickiness involves prices rather than wages. Second, it shows that there is no necessary connection between nominal rigidity and unemployment. And third, it is easy to use; because of this, models like it often appear in the theoretical literature.

w{L) is

f-hy ) f-\y) i

FIGURE 5.14 A non-Walrasian labor market when prices are sticky and nominal wages are flexible.

Case 4: Sticky Wages, Flexible Prices, and Imperfect Competition

Just as Case 3 extends Case 2 by Introducing real imperfections in the labor market, the hnal case extends Case 1 by introducing real imperfections in the goods market. Specifically, assume (as in Case 1) that the nominal wage is rigid at W and that nominal prices are flexible, and continue to assume that output and employment are related by the production function. Now, however, assume that the goods market is imperfectly competitive. With imperfect competition, price is a markup over marginal cost. Paralleling our assumptions about the real wage in Case 3, we do not model the determinants of the markup, but simply assume that there is a "markup function." With these assumptions, price is given by

F(LY

(5.30)

WIF(L) is marginal cost and p is the markup.

Equation (5.30) implies that the real wage, W/P, is given by F(L)lix(L). Without any restriction on p{L), one cannot say how WIP varies with L. If