*Change in recessions Is computed relative to the variables average growth over the full postwar period, 1947-1992.

flnflation was zero at both the peak and the trough of the 1948-49 recession. 11 only beginning m 1959. Source: Citibase.

in the unemployment rate are generally smaller than the movements in output. The relationship between movements in output and the unemployment rate is known as Okuns law. As originally formulated by Okun (1962), the "law" stated that a shortfaU in GDP of 3% relative to normal growth produces a 1 percentage point rise in the unemployment rate; a more accurate description of the current relationship is 2 to 1.

The remaining lines of Table 4.3 summarize the behavior of various price and financial variables. Inflation generally (but not always) falls. The real wage, at least as measured in aggregate data, tends to fall slightly in recessions. And nominal interest rates and the real money stock both generally dechne.

4.2 Theories of Fluctuations

It is natural to begin by asking whether aggregate fluctuations can be understood using a Walrasian model-that is, a competitive model without any externalities, asymmetric information, missing markets, or other imperfections. If they can, then the analysis of fluctuations may not require any fundamental departure from conventional microeconomic analysis.

Other ways of summarizing the cyclical behavior of inflation and the price level give different results. Because of this, the cyclical behavior of inflation and the price level, and the implications of that behavior, are controversial. See Ky diand and Frescott (1990); Cooley and Ohanian (1991); Backus and Kehoe (1992); Ball and Mankiw (1994); and Rotemberg (1994).

TABLE 4.3 Behavior of some important macroeconomic variables in recessions

Average change Number of recessions Variable in recessions in which variable falls

King, Plosser, and Rebelo (1988, Section 3) and King and Rebelo (1986) investigate the consequences of using an endogenous growth model rather than the Ramsey model as the starting point for analyzing fluctuations.

The seminal papers include Kydland and Prescott (1982); Long and Plosser (1983); Prescott (1986); and Black (1982).

"See Aiyagari, Christiano, and Eichenbaum (1992); Baxter and King (1993); and Chris-tiano and Eichenbaum (1992a).

As emphasized in Chapter 2, the Ramsey model is the natural Walrasian baseline model of the aggregate economy: the model excludes not only market imperfections, but also all issues raised by heterogeneity among households. This chapter is therefore devoted to extending a variant of the Ramsey model to incorporate aggregate fluctuations. This requires modifying the model in two ways. First, there must be a source of disturbances: without shocks, the Ramsey model converges to a balanced growth path and then grows smoothly. The initial extensions of the Ramsey model to include fluctuations emphasized shocks to the economys technology-that is, changes tn the production function fromperiod to period. More recently, work in this area has also emphasized changes in government purchases. Both types of shocks represent real-as opposed to monetary, or nominal-disturbances: technology shocks change the amount that is produced from a given quantity of inputs, and government-purchases shocks change the quantity of goods available to the private economy for a given level of production. For this reason, the models are known as real-business-cyde (or RBQ models.

The second change that is needed to the Ramsey model is to allow for variations in employment. In all the models we have seen, labor supply is exogenous and either constant or growing smoothly. Real-business-cycle theory focuses on the question of whether a Walrasian model provides a good description of the main features of observed fluctuations. Models in this literature therefore allow for changes in employment by making households utility depend not just on their consumption but also on the amount they work; employment is then determined by the intersection of labor supply and labor demand.

As discussed in the final section of this chapter, there is considerable debate about whether fluctuations can in fact be understood using Walrasian models. In particular, a large number of macroeconomists believe that the technology shocks and the propagation mechanisms of real-business-cycle models are of little relevance to actual fluctuations, and that nominal disturbances and a failure of nominal prices and wages to adjust fully to those disturbances are central to fluctuations.

Chapters 5 and 6 are therefore devoted to Keynesian theories of fluctuations. To keep the analysis tractable, and to do justice to how macroeconomics is actually done. Chapters 5 and 6 do not pursue the strategy of adding incomplete nominal adjustment to a Ramsey-style model. Instead, to focus on the consequences and causes of incomplete nominal adjustment, they investigate price stickiness in models that are dramatically simplified on the real side. Chapter 5 takes nominal stickiness as given and investigates

its effects. Chapter 6 tackles the questions of why nominal prices might not respond fully to disturbances.

One conclusion of Chapter 6 is that significant nominal stickiness is much more likely to arise if there are departures from a Walrasian model in addition to some type of direct impediment to instantaneous nominal adjustment: imperfections in the goods, credit, and labor markets may greatly magnify the consequences of barriers to nominal flexibihty. Thus modem Keynesian theories differ i±om real-business-cycle models not only by including barriers to complete nominal adjustment, but also tn their analysis of how the economy would operate in the absence of those barriers.

This division of theories of fluctuations into ones focusing on real shocks impinging on a Walrasian economy and ones focusing on nominal disturbances affecting an economy with significant imperfections omits the possibihty of real non-Walrasian theories. That is, il may be that nominal shocks and nominal stickiness are not important to fluctuations, but that there are other departures from the Walrasian baseline that are central to fluctuations. There are a host of possible non-Walrasian features of the economy-such as imperfect competition, externalities, asymmetric information, departures from rationahty, and failures of markets to clear-and thus a host of possible real non-Walrasian theories of fluctuations. Thus we will not attempt to discuss them comprehensively. Instead, we will consider them briefly at the end of Chapter 6.

43 A Baseline Real-Business-Cycle Model

We now turn to a specific real-business-cycle model. The assumptions and functional forms are similar to those used in most such models (see, for example, Prescott, 1986; Christiano and Eichenbaum, 1992a; Baxter and King, 1993; and Campbell, 1994). The model is a discrete-time variation of the Ramsey model of Chapter 2. Because our goal is to describe the quantitative behavior of the economy, we will assume specific functional forms for the production and utility functions.

The economy consists of a large number of identical, price-taking firms and a large number of identical, price-taking households. As in the Ramsey model, households are infinitely-lived. The inputs to production are again capital (K), labor (i), and "technology" (A). The production function is Cobb-Douglas; thus output in period t is

Ff = KfiAtLt)", 0 < a < 1. (4.1)

Output is divided among consumption (C), investment (/), and government purchases (G). Fraction 6 of capital depreciates each period. Thus the capital stock in period t -i- 1 is