Keynesian theory and the aggregate-supply/aggregate-demand framework: A defense

INTRODUCTION

In a paper published earlier in this journal Barro argues that the widely used aggregate–supply/aggregate–demand (AS/AD) framework is “unsatisfactory and should be abandoned as a teaching tool” [1994, 1]. This recommendation is based on the claim that the models usually represented by the AS/AD framework fall into two categories: those that are inconsistent or those that can be represented more conveniently in a different way. Barro is not alone in having misgivings about the AS /AD framework: others, including Fields and Hart [1990], Colander [1995] and Bhaduri, Laski and Riese [1995],1 have also recently argued that the framework is internally inconsistent and fundamentally flawed, although they have not called for its abandonment as a pedagogic device. As we understand it, the AS/AD framework is a general diagrammatic approach.

The defining characteristics of the framework is the use of a price-output diagram to illustrate two reduced-form relations, the (misleadingly labelled) AS and AD curves. The AS curve represents a reduced form of equations describing the labor market and pricing behavior while the AD curve represents a reduced form of equations depicting the nature of asset markets and the demand for goods.

Since this definition of the AS/AD framework says nothing about the nature of labor, goods, and assets markets, it cannot be called a model of the economy. For present purposes, this observation has two implications. First, many different models can be depicted using the AS/AD framework. Second, while specific AS/AD models may have inconsistencies, such inconsistencies are not necessarily attributable to the framework itself.

In this paper we argue that Barro’s judgment regarding the AS/AD framework is incorrect, that his analysis is riddled with some basic and fundamental errors, and that the AS/AD framework, pace Barro, can represent a variety of macroeconomic models in a consistent way. As such, it may be useful in teaching basic macroeconomic theory, enabling students to appreciate the differences among alternative macroeconomic approaches in terms of a common simple framework. We hope to convince others not to join Barro in abandoning what can be a useful teaching device.2

Our main disagreement with Barro does not concern pedagogical issues, however, for Barro’s paper is not primarily about how one should teach a given material. Underlying the discussion of teaching tools loom substantive issues of economic theory. In Barro’s view there are currently only “two types of internally-consistent models that allow for cyclical interactions between monetary and real variables” [1994, 4]: the fix-price models associated with the work of Barro and Grossman [1976] and others, and the new classical models of market clearing.

This position is blatantly wrong. In fact, Barro (somewhat inconsistently) acknowledges the existence of other types of models elsewhere in his paper when he refers to “what used to be called the complete Keynesian model” [ibid., 1]. Barro does not criticize this model on grounds of internal inconsistency but suggests that the model “was rejected long ago for good reasons” [ibid., 1]. These reasons, we are told are partly theoretical (essentially that it does not capture some of the main Keynesian ideas) and partly empirical (it is inconsistent with the observed cyclical pattern for the real wage) [ibid, 4]. Barro’s theoretical criticisms are in our opinion unjustified, and the empirical objection – which has a long history going back to Dunlop [1938] and Tarshis [1939] and which applies to monetarist and new classical models as well can be met by a straightforward reformulation of the theory to take account of imperfect competition and non-diminishing returns to labor.

The other aforementioned authors who have criticized the AS/AD framework have put forward their own suggestions about how the framework can be modified to free it of its internal inconsistencies and empirical shortcomings. Although we do not concern ourselves with the details of their arguments, we do comment on their work to the extent that it is directly related to the issues relevant for understanding the problems with Barro’s criticisms and our own interpretation of the AS/AD framework and models. We should say at the outset that we disagree with many of the claims made by these authors and our interpretation of consistent AS/AD models are significantly different from theirs.