A comment on "does the aggregate demand curve suffer from the fallacy of composition".

A recent article in this journal by Professors Saltz, Cantrell, and Horton questioned the validity and existence of the aggregate demand curve in macroeconomics. (1) More specifically, the authors contend that "the aggregate demand problem ... is a classic example of the fallacy of composition" and proceed to investigate this contention within each of four theoretical arguments that have typically been used to support the negative slope of the aggregate demand curve. The purpose and assertion of this comment is that Saltz, et. al., have committed several methodological errors in attempting to establish potential ambiguities contained within the traditional conceptualization and derivation of the aggregate demand curve. When these errors are corrected and the ambiguities cleared, it seems quite logical that an aggregate demand curve should exist and that there should be an inverse relationship between the general price level and aggregate quantity of goods demanded.

Perhaps the most fundamental and obvious error occurs in the analysis of the international price level effect. The authors begin by noting that the validity of this effect is tenuous because of a number of questionable assumptions that underlie it. They write, "First this argument assumes a constant foreign price level." This assumption is not a weakness but rather a methodological inevitability stemming from the ceteris paribus assumption that anchors all formal models in economic theory, both microeconomic and macroeconomic. Indeed, the authors recognize this critical assumption explicitly in their definition of aggregate demand. The implication of this assumption, of course, is that a change in any non-price level determinant of aggregate demand such as the foreign price level, only shifts a given aggregate demand curve. Changes of variables other than the price level are irrelevant to the fundamental issue of the existence and logical consistency implicit within the aggregate demand curve derivation.

With respect to the Pigou effect, Saltz et. al. contends "unexpected changes in the price level cause a redistribution of wealth or income, but not a loss of net income." Absent from the analysis however, is the acknowledgment that a price-level induced change in the distribution of income per se can affect aggregate demand. Therefore, if the distribution of aggregate income were endogenized, the higher price level would redistribute income from lower- to higher-income individuals in the authors' scenario, leading to a decrease in consumption spending and aggregate demand. This conclusion is supported by income studies, which have consistently shown that families with lower incomes have higher average propensities to consume and likely have higher marginal propensities to consume than families with higher incomes.

At least two problems are present with the authors' analysis of the Keynes effect. First, they appear to confuse a movement of the money demand curve with movements along it when they write "... we can also expect a fall in real money demand as people reduce their holdings of monetary assets in response to an increase in the opportunity cost of holding money." As the interest rate or opportunity cost of money increases from a higher price level reducing the real money supply, the resulting adjustment in the money market is not a movement of the money demand curve but rather a movement northwest along a given money demand curve, such that the interest rate rises unambiguously with a corresponding reduction in investment spending and aggregate demand.

Second, the authors again violate the ceteris paribus assumption in their analysis of the Keynes effect when they permit inflationary expectations to change endogenously and simultaneously with respect to a change in the price level. The end result is then blurred since it doesn't allow a clear delineation of how an increase in the price level alone leads to a decrease in real money balances and an increase in the interest rate. This otherwise straightforward linkage not surprisingly becomes quite vague and ambiguous when any number of price-expectations scenarios is conjectured. It is interesting to note that if a similar methodology were adopted in developing the theory of microeconomic demand, the resulting curve would also have an ambiguous slope. Again, a change in the anticipated price level must be treated as a shifter of aggregate demand and not a movement along the curve. This distinction between changes in the actual price level versus changes in the expected price level is a necessary and conventional practice in the analyses of not only aggregate demand but also both aggregate supply and the closely related Phillips curve in macroeconomics.

The essence of the argument about the possibility of ambiguous price-level effects on aggregate demand is based on the distinction between anticipated and unanticipated price expectations. This serves as a logical starting point from which to overturn a central argument presented in this paper. First, of course, is the fact that price expectations represent a non-price determinant that is a shifter of the demand curve. Therefore, any argument about price expectations is necessarily outside the investigation into whether the price level and aggregate demand are inversely related. This certainly seems to be the traditional methodology and conceptualization of microeconomic demand. If changes in both price and price expectations are embedded in movements along the microeconomic demand curve, or any aggregated relative of the market demand curve for that matter, the resulting curve would violate the ceteris paribus assumption and represent a departure from the definition of the Law of Demand. More specifically, when the price level falls, it is responded to because it is assumed to be completely unexpected. This is a well-known and standard neoclassical behavioral assumption embedded in the model. Certainly price expectations are present for all points along the aggregate demand curve but are assumed to be constant prior to any arbitrary movement along the AD schedule, after which the new expected price lag-adjusts to equal the value of the new price level. In other words, at the initial price level economic agents have the expectation that a given price will remain in place, presumably equaling the current price. Otherwise, any expectation that the price level will increase or decrease relative to its initial value would necessarily shift the AD curve at that price level. Note that this relationship would also hold for any microeconomic demand curve. That is, should there be an increase or decrease in the expected future price of a good, a shift in the market demand curve would result.

In conclusion, we believe that the analysis by Saltz, Cantrell, and Horton is flawed. Fundamentally, when the ceteris paribus assumption is honored, it seems that the international, Pigou, and Keynes effects are logically consistent and support a negatively sloped aggregate demand curve within a price level, real output plane. When these effects are presented correctly, student confusion regarding aggregate demand should be minimized.

Note

(1.) Ira Saltz, Pat Cantrell, and Joseph Horton, "Does the Aggregate Demand Curve Suffer from the Fallacy of Composition," The American Economist, Vol. 46, No. 1, pages 61-65.

Ben L. Kyer * and Gary E. Maggs **

* Professor of Economics, St. John Fisher College, Rochester, New York.

** The Benjamin Wall Ingram, III, Professor of Economics, Francis Marion University, Florence, South Carolina