Chapter 5 TRADITIONAL KEYNESIAN THEORIES OF FLUCTUATIONS

5.1 Introduction

This chapter and the next develop models of fluctuations based on the assumption that there are barriers to the instantaneous adjustment of nominal prices and wages. As we will see, sluggish nominal adjustment causes changes in the aggregate demand for goods at a given level of prices to affect the amoimt that firms produce. As a result, it causes purely monetary disturbances (which affect only demand) to change employment and output. In addition, many real shocks, including changes in government purchases, investment demand, and technology, affect aggregate demand at a given price level; thus sluggish price adjustment creates a channel other than the intertemporal-substitution mechanism of real-business-cycle models through which these shocks affect employment and output.

This chapter takes nominal stickiness as given. It has two main goals. The first is to investigate aggregate demand. We will examine the determinants of aggregate demand at a given price level and the effects of changes in the price level. The second is to consider alternative assumptions about the form of nominal rigidity. We will investigate different assumptions implications for firms willingness to change output in response to changes in aggregate demand and for the behavior of real wages, markups, and inflation. Chapter 6 then turns to the questions of why nominal prices and wages might not adjust immediately to disturbances.

The Keynesian Approach to Modeling

.As will quickly become clear, Keynesian models differ from real-business-cycle models not just in substance, but also in style. Real-business-cycle models typically begin with microeconomic assumptions about households preferences, firms production functions, the structure of markets, and the evolution of quantities over time. Thus the models are fully specified

dynamic general equilibrium models. Keynesian models, in contrast, often begin by directly specifying relationships among aggregate variables. The relationships are often static, and the models Implications for the behavior of some variables (such as the capital stock) are often omitted from the analysis. Even versions of the models that build up the behavior of some variables from microeconomic foundations often specify the behavior of others directly.

The idea behind this shortcut aggregate approach to modeling is threefold. First, it is simple. The solutions to even relatively basic real-business-cycle models are complicated, and usually require numerical methods. Basic Keynesian models, in contrast, can be analyzed graphically.

Second, many features of the economy are hkely to be robust to the details of the microeconomic environment. For example, the opportunity cost of holding non-interest-bearing money is the nominal interest rate. Thus, regardless of the precise reasons that people hold money, the quantity of money demanded is likely to be decreasing in the nominal interest rate. Building a model of money demand from microeconomic foundations would probably add little insight on this issue.

And third, by insisting on microeconomic foundations we could in fact miss important effects. In the case of money demand, for example, beginning with microeconomic foundations could lead us to a specific functional form for the money demand function; yet that functional form would probably not be robust to reasonable changes in the microeconomic assumptions. To give a more significant example, traditional Keynesian models give current income a particularly important role in consumption demand. If we build up consumer behavior from intertemporal optimization with freedom to borrow against future income, we find that current income is no more important than discounted future income in determining households consumption. But, as we saw in the discussion of Ricardian equivalence in Section 2.9, there are microeconomic models that imply a greater role for current income. Yet these models are often complicated, and thus difficult to embed in models of the entire economy. Thus we may get greater insight by specifying directly that consumption depends particularly on current income.

Of course, there are also disadvantages to the Keynesian approach to modeling. Without microeconomic foundations, welfare analysis is not possible. More importantly, specifying aggregate relationships directly may cause us to overlook important effects. For example, stating directly that consumption depends on current disposable income neglects the possibility that temporary and permanent income movements may have different effects; similarly, it neglects the possibility of Ricardian equivalence. When consumption behavior is derived from microeconomic foundations, in contrast, these possibilities are immediately apparent.

Finally, aggregate relationships may change when the structure of the economy or the nature of policy changes. Thus, working with aggregate re-

Textbook treatments of the Keynesian model of aggregate demand include Abel and Bernanke (1992, Chapters 12-13); Dombusch and Fischer (1994, Chapters 3-7); Gordon (1993, Chapters 3-6); Hall and Taylor (1991, Chapters 6-7); and Mankiw (1994, Chapters 8-10, 13). The presentation in Sections 5.2 and 5.3 is most similar to Mankiws.

lationships rather than microeconomic assumptions may lead us astray in assessing the likely consequences of changes in policy. This is the basis of the Lucas critique of traditional macroeconomic models, which we will discuss in Chapter 6.

Overview

The remainder of the chapter consists of five sections. Sections 5.2 and 5.3 develop the aggregate demand side of the standard Keynesian model. These sections take as given that nominal prices and wages are not completely flexible, and that firms change their output in response to changes in demand. Section 5.2 assumes a closed economy, and Section 5.3 considers the open-economy case.

Sections 5.4 and 5.5 consider aggregate supply. Section 5.4 shows how different combinations of wage rigidity, price rigidity, and non-Walrasian features of the labor and goods markets yield different implications about the effect of shifts in aggregate demand on output, unemployment, the real wage, and the markup. Section 5.5 discusses short-run and long-run output-inflation tradeoffs.

Finally, Section 5.6 discusses some empirical evidence about the real effects of monetary changes.

5.2 Review of the Textbook Keynesian Model of Aggregate Demand

The textbook Keynesian model is traditionally summarized by two curves in output-price or output-inflation space, an aggregate demand (AD) curve and an aggregate supply {AS) curve. The AD curve slopes down and the AS curve slopes up. Ihese curves are shown in Figure 5.1.

The fact that the aggregate supply curve is upward-sloping rather than vertical is the critical feature of the model. If the AS curve is vertical, changes on the demand side of the economy affect only prices. But if it is merely upward-sloping, changes in aggregate demand affect both prices and output.

The AD curve summarizes the demand side of the economy. It is derived from two familiar curves in output-interest rate space, the IS and LM curves. These are shown in Figure 5.2. The curves are drawn for a given price level; as we will see shortly, considering different values of the price level allows