CD P

FIGURE 6.4 The impact of a fall in aggregate output on the representative firms profits as a function of its price

the firms profits as a function of its price. The fall in aggregate output affects this function in two ways. First, since the demand curve for the firms good falls, the profit function shifts down. As described above, the firm cannot undo this change. Second, the firms profit-maximizing price is less than before. This the firm can do something about. If the firm does not pay the menu cost, its price remains the same, and so it is not charging the new profit-maximizing price. If the firm pays the menu cost, on the other hand, it can go to the peak of the profit function.

The firms incentive to adjust its price is thus given by the distance AB in the diagram. This distance depends on two factors: the difference between the old and new profit-maximizing prices, and the curvature of the profit function. We consider each in turn.

Since other firms prices are unchanged, a change in the firms nominal price is also a change in its real price. In addition, the shift in aggregate demand with other prices held fixed changes aggregate output. Thus the difference between the firms new and old profit-maximizing prices (distance CD in the figure) is determined by how the profit-maximizing real price depends on aggregate output-that is, by the degree of real rigidity. Greater real rigidity, holding the curvature of the profit function fixed, reduces the firms incentive to adjust its price in response to an aggregate

This corresponds to the assumption that the profit-maximizing relative price is increasing in aggregate output; that is, it corresponds to the assumption that > 0 in the pricing equation, (6.45). As described in Section 6.6, this condition is needed for the equilibrium with flexible prices to be stable.

rates.

*This analogy is originally due to Friedman (1953, p. 173), m the context of exchange

demand movement if other firms do not change their prices. The intuition is the same as the intuition for why greater real rigidity increases the real effects of nominal shocks in the Fischer and Taylor models: greater real rigidity means that firms do not want their prices to depart greatly from others prices.

The curvature of the profit function determines the cost of a given departure of price from the profit-maximizing level. The less sensitive the profit function is to departures from the optimum, the smaller the incentive for price adjustment (for a given ), and so the larger the range of shocks for which nonadjustment of prices is an equilibrium. Thus, in general terms what is needed for small costs of price adjustment to generate substantial nominal rigidity is some combination of real rigidity and of insensitivity of the profit function.

Seen in terms of real rigidity and insensitivity of the profit function, it is easy to see why the incentive for price adjustment in our baseUne calculation is so large: there is immense "real flexibility" rather than real rigidity. Since the profit-maximizing real price is [ )/( ) - 1)]Y, its elasticity with respect to output is 1/p. If the elasticity of labor supply, p, is small, the elasticity of (P, / )* with respect to Y is therefore large. A value of p of 0.1, for example, implies an elasticity of (P, / )* with respect to Y of 10.

A well-known analogy may help to make clear how the combination of menu costs with either real rigidity or insensitivity of the profit function (or both) can lead to considerable nominal stickiness: monetary disturbances may have real effects for the same reasons that the switch to daylight saving time does.2 The resetting of clocks is a purely nominal change-it simply alters the labels assigned to different times of day. But the change is associated with changes in real schedules-that is, the times of various activities relative to the sun. And in contrast to the case of monetary disturbances, there can be no doubt that the switch to daylight saving time is the cause of the changes in real schedules.

If there were literally no cost to changing nominal schedules and communicating this information to others, daylight saving time would just cause everyone to do this and would have no effect on real schedules. Thus for it to change real schedules, there must be some cost to changing nominal schedules. These costs are analogous to the menu costs of changing prices; and like the menu costs, they do not appear to be large. The reason that these small costs cause the switch to have real effects appears to be that individuals and businesses are generally much more concerned about their schedules relative to one anothers than about their schedules relative to the sun. Thus, given that others do not change their scheduled hours, each individual does not wish to incur the cost of changing his or hers. This is analogous to the effects of real rigidity in the price-setting case. Finally, the

Specific Sources of Real Rigidity

To see what types of factors can give rise to real rigidity and insensitivity of the profit function, retum to the marginal revenue-marginal cost framework of Figure 6.3. On the cost side, the smaller the faU in marginal cost is as a result of the fall in aggregate output, the smaller the firms incentive to cut its price and increase its output, and thus the more likely nominal rigidity is to be an equilibrium. This can occur in two ways. First, a smaller downward shift of the profit function in response to a fall in aggregate output implies a smaller decline in the firms profit-maximizing price-that is, it corresponds to greater real rigidity. Second, a flatter marginal cost curve imphes both greater insensitivity of the profit function and greater real rigidity.

On the revenue side, the larger the fall in marginal revenue is when aggregate output falls, the smaller the gap between marginal revenue and marginal cost at the representative firms initial price, and so the smaller the incentive for price adjustment. Specifically, a larger leftward shift of the marginal revenue curve corresponds to increased real rigidity, and so reduces the incentive for price adjustment. In addition, a steeper marginal revenue curve (for a given leftward shift) also increases the degree of real rigidity, and so again acts to reduce the incentive for adjustment.

Since there are many potential determinants of the cyclical behavior of marginal cost and marginal revenue, the hypothesis that small frictions in price adjustment result in considerable nominal rigidity is not tied to any specific view of the structure of the economy. On the cost side, factors that may make costs much less procyclical than in our baseline case include: thick-market externalities that make purchasing inputs and selling final products easier in times of high economic activity (for example. Diamond, 1982); other external economies of scale or agglomeration economies that make costs lower when other firms are producing more (for example. Hall, 1991; CabaUero and Lyons, 1992; and Cooper and Haltiwanger, 1993); capital-market imperfections that make the cost of finance higher in recessions, when firms cash flow and credit worthiness are lower (for example, Bernanke and Gertler, 1989, and Kiyotaki and Moore, 1995); and input-output linkages among firms that cause firms to face constant costs for their inputs when prices are sticky (Basu, 1993). On the revenue side, some potentially important factors are: thick-market effects that make it easier for firms to disseminate information and for consumers to acquire it when aggregate output is high, and thus make demand more elastic (Warner and

Recall that for simplicity the marginal cost curve was not shown as shifting down in Figure 6.3 (see n. 24). There is no reason to expect it to stay fixed in general, however.

less concerned that individuals are about precisely what their schedules are, the less willing they are to incur the cost of changing them; this is analogous to the insensitivity of the profit function in the price-setting case.