The Barro-Grossman model, proposed by Robert J. Barro and Herschel I. Grossman, was first published in 1971 in what was to become, seventeen years later, the most frequently cited article in the American Economic Review. It was further developed in Barro and Grossman’s Money, Employment, and Inflation (1976), though this latter work never became a dominant graduate textbook, as it was overtaken by what has been called the rational expectations revolution. Barro himself largely abandoned the Barro-Grossman model in the late 1970s. Though work based on the model continued to appear, and in that sense the model never disappeared, it was taught in very few graduate programs and fell out of fashion in the early 1980s.
The central idea of the model is that of an equilibrium with rationing, sometimes called the economics ofdis-equilibrium. Suppose that for some reason—such as an external shock—an economy develops a significant level of unemployment. Workers will find that they cannot sell as much labor as they would normally expect to sell, and they will have to cut back on their purchases of consumption goods. Firms will find that they cannot sell the goods they wish to sell at the prevailing prices, and so will not be willing to employ more labor. The economy gets stuck in a "Catch 22" situation, in which spending cannot rise because workers cannot find jobs, even if they offer to work for a lower wage, and firms will not employ more workers, because even if they lower prices, they cannot increase their sales revenue. Both groups are rationed— that is, they face constraints on the quantities they can sell at the prevailing prices.
The key to the model is a distinction introduced by Robert Clower (1965) between effective and notional demands, the former being calculated with quantity constraints taken into account, the latter with constraints ignored (notional demands are sometimes called Walrasian). In an equilibrium with rationing, effective demand equals supply. For example, workers may wish to sell more labor and purchase more consumption goods, but they cannot do so. Thus, if price changes depend on effective demands, there is no pressure to change prices. It is therefore possible to have an equilibrium with rationing in which there is high unemployment, even if wages and prices are at the level consistent with full employment. Market forces will not bring the economy back toward full-employment equilibrium.
The Barro-Grossman model was a fixed-price model in the sense that prices were taken as a parameter when determining quantities of employment and output. They were not necessarily constant—the 1976 version of the model allowed for inflation by having prices change in response to effective excess demands—but they adjusted much more slowly than quantities. In the late 1970s, Barro interpreted this as meaning that prices were fixed by long-term contracts, and when it was shown that it would not be optimal for such contracts to specify "sticky" prices (that is, prices that do not change immediately to eliminate potential differences between supply and demand), he abandoned the approach. However, other users of such models interpreted equilibrium with rationing differently. Clower had been interested in dynamic price-adjustment mechanisms in markets that were typically not perfectly competitive. Axel Leijonhufvud (1968), who had done more than anyone to popularize the idea of disequilibrium macroeconomics, was more interested in the idea of intertemporal disequilibrium. Jacques Dreze (1975), who pioneered the incorporation of quantity constraints in general equilibrium models, came to them through problems of information.
The folklore in macroeconomics is that the Barro-Grossman model went out of fashion in the 1970s because it did not make sense to assume rigid prices when inflation was running at levels that exceeded 25 percent in some OECD countries. However, paradoxically, it was precisely because of inflation that some economists turned to such models. Edmond Malinvaud (1977) used a model of equilibrium with rationing (not the Barro-Grossman model, but a member of the same family of models) because it was the only framework he could find in which to make sense of stagflation—that is, of simultaneously rising inflation and unemployment. Taking the wage rate and the price level as parameters was the first step in discussing what might cause an economy to move between different regimes: Keynesian unemployment (unemployment accompanied by a surplus of goods), classical unemployment (unemployment and a shortage of goods), and repressed inflation (shortages of labor and goods).
Possibly the main significance of the Barro-Grossman model in the history of macroeconomics is that it was an important part of the search for microfoundations of macroeconomics. Barro and Grossman wanted to construct a macroeconomic model with rigorous microeco-nomic foundations (rigorous in the sense of being based on maximizing behavior by individual firms and households). It should be seen as parallel to the search for microfoundations of Edmund Phelps and his collaborators (1970). Because macroeconomics was at that time synonymous with Keynesian economics, Barro and Grossman created a Keynesian model. Their model comprised a representative household and a representative firm. When Barro abandoned the fixed-price approach, he retained those elements, and when household and firm were combined, there emerged the representative agent model that is characteristic of modern macroeconomics.
