Background

A large amount of economic theory and analysis is based on the idea of economic equilibrium - the state where economic forces, such as supply and demand, are balanced, and variables, such as price and quantity, are stable. Equilibrium theory sits at the heart of most modern economics.

Classical, 1750-1830

Adam Smith discussed the concept of equilibrium in Part I, chapter 7 of The Wealth of Nations.¹⁶ He talks about a "natural price" for everything in the economy, which he defined as "the central price, to which the prices of all commodities are continually gravitating". This analogy of economic forces being like physical forces, pulling variables towards their equilibrium values, has been an influential way of thinking. The concept of equilibrium was extended to the entire economy with Say's Law. Say argued that "supply creates its own demand", meaning there can never be a general surplus in the entire economy, as the income generated from all goods produced is equal to the total spending on all goods produced - in effect arguing that aggregate supply and demand in economy must always be in equilibrium. Say's Law was widely believed in the Classical period (except by Malthus, who had an extended argument with Ricardo over it) , but has since been criticised by many economists, including Marx and, most influentially, Keynes.

Neoclassical partial equilibrium, 1830-1890

Neoclassical economics formalised the notion of equilibrium. This process was started by Auguste Cournot, one of the early figures in neoclassical economics, who was one of the first economists to think of supply and demand as mathematical functions. He analysed the equilibrium that competitive markets would tend towards based on these functions, and was the first person to draw intersecting supply and demand curves to illustrate this.¹ He extended this to study the differences between markets with different numbers of firms, and his model of oligopoly is still in use today. Cournot's work on supply and demand was extended and popularised by Alfred Marshall, who clarified the different roles of demand and supply in setting equilibrium prices. Marshall emphasised the role of time, describing three time periods where different factors determine the equilibrium. In the short term, it is impossible to change supply, so the equilibrium price depends mainly on demand; in the medium term, supply can be changed, but only in some ways, so variable costs (such as wages or material costs) have most effect on the equilibrium price; and in the long term, it is major fixed costs (such as rent or energy costs) that determine the price.⁷

Walrasian general equilibrium, 1870-1900

Cournot and Marshall only analysed conditions of partial equilibrium - looking at the equilibrium for one market, and ignoring all other markets. However, all markets are inter-connected, and changes in price or quantity in one market affect the demand and supply in other markets. Léon Walras worked on a theory of general equilibrium, where the entire economy is in a state of equilibrium. He constructed a set of simultaneous equations to give the prices and quantities for every good, including markets for factors of production and money. He then proved that there is one set of solutions to these equations, meaning there is a possible state where every market is in equilibrium. He also described a theoretical mechanism by which prices could reach this equilibrium, in a state of perfect competition: an auction where everyone submits their demand and supply for every good, and this is used to find the price. Although this mechanism is highly artificial, and relies on assumptions of unrestricted markets and perfect competition, it illustrates how a price mechanism can balance supply and demand for most markets, and illuminates the factors affecting changes to prices and quantities.^{1, 8}

Modern general equilibrium theory, 1950-1970

General equilibrium theory was developed greatly in the mid 20^th century by Kenneth Arrow, Gérard Debreu and Lionel McKenzie. One major advance was the theory of market completion. The economists observed that, when Walras talks about equilibrium in the market for every good, it is not enough to think about just different types of goods. A good that is delivered at a different time or in a different place ought to be thought of as being a different good, with its own market. A general equilibrium model assumes that all these markets exist, which is very unlikely to be true in the real world. However, this theory suggests that creating markets for different times and places, such as with futures contracts, will help to bring the economy closer to its general equilibrium and increase efficiency. The other key advance was the Welfare Theorems, which mathematically described some characteristics of an economy in the equilibrium described by Walras. The First Welfare Theorem states that any economy in general equilibrium is Pareto efficient; that is, there is no way to make anyone better off without making someone else worse off. This implies that intervention in the economy will not make it more efficient unless that intervention reduces market failure, bringing the economy closer to the idealised conditions of free and complete markets, perfect competition, and lack of externalities. However, an efficient economy is not necessarily in an ideal state: if one person owned all the resources in the economy, it would be impossible to make anyone better off without making the owner worse off, so this would be a Pareto efficient outcome. There are an infinite number of efficient states of the economy, some much fairer than others. The Second Welfare Theorem states that competitive markets can lead to a Walrasian equilibrium in any one of these efficient states; all that is necessary to achieve a different outcome is redistributing the resources each person starts the process with. Taken together, these theorems imply that free markets will produce an efficient outcome, co-ordinating all of the decisions about demand and supply in the economy, but redistribution can make sure that the outcome that is produced is fair. These theorems are based on strong and unrealistic assumptions, but their basic insights are useful and the general equilibrium theory based on them works well as a simplified model of the economy.^{8, 11}

Dynamic Stochastic General Equilibrium, 1980-

An extension of general equilibrium modelling that has become increasingly common in recent decades is Dynamic Stochastic General Equilibrium (DSGE). Dynamic means that they take account of changes over time, rather than looking at a static equilibrium like older GE models, and stochastic means that they take account of random fluctuations in the economy. The basic foundations for the model were laid out by Finn Kydland and Edward Prescott in 1982. As this model is based on the actions of individual agents and markets, it is based on microeconomic theory. This means that it satisfies the Lucas critique (Lucas criticised purely macroeconomic models for not taking account of the expectations of individuals in the models), and can be used to forecast the effects of policy changes. They have received criticism, like most general equilibrium models, for having unrealistic assumptions, and also for poor forecasting performance; however, there are few alternative models that satisfy the Lucas critique, so these models are still likely to continue to be used. There have also been many attempts in recent years to extend DSGE models to improve their realism.

The future

DSGE models seem to be gaining ground as a form of macroeconomic modelling. They are now used as the Bank of England's primary model, as well as by the European Central Bank and the Federal Reserve. However, many economists have criticised reliance on DSGE models, based on their poor empirical performance. There are a number of suggested alternatives. One is to pay more attention to an alternative style of modelling pioneered by Paul Samuelson, overlapping generations modelling, which takes into account the fact that agents age and die; this type of model violates the First Welfare Theorem, which states that equilibriums must be efficient. A second approach is so-called "agent-based modelling", in which multiple agents interact based on rules about their behaviour. Agent-based models do not assume that the economy is in or will reach equilibrium. However, they require a lot of computational power, and so far have produced few useful results. There are various other attempts to model economies out of their equilibrium, but none of these have gained much ground within the economics profession. However, one of these models may prove to be particularly useful and then be accepted; it is difficult to say whether equilibrium or non-equilibrium models will be more important for the future of economics.