Methodology
Long-run, full-employment equilibrium occupied a central role in the thinking of those economists Keynes called “classical” (by which he meant all those who believed that the economy would adjust automatically to full employment).
The prevailing long-run analysis was static and therefore abstracted from historical time. Deviations from long- run equilibrium or the processes that might bring the system back to equilibrium were rarely entertained, let alone analysed (Wicksell, and Keynes in TM, are notable exceptions). It was simply assumed that the economy would return to equilibrium if disturbed. (Keynes’s objection to this assumption is the subject of a very important article in 1934: “Poverty in plenty: is the economic system self-adjusting?”, Keynes 1934 [1973]: 485-92.)In the GT, Keynes changed all that. The long run in the GT emerges organically out of the processes that unfold in a succession of (Marshallian) “short periods”. Trend and cycle are not seen as separate but as part of the same process. Features which his contemporaries had assumed to hold in the long run, most importantly, full employment and the validity of the quantity theory, were no longer to be treated as articles of faith but as proper subjects for analysis, and they turned out to hold only in circumstances so special as to be nearly irrelevant.
Instead of a preconceived long run and transitory disturbances, Keynes adopted Marshall’s three “periods” (not “runs” - the short run refers to historical time and the long run is timeless): the market period (the “day”), when output is fixed (GT, Keynes 1936 [1973]: ch. 5); the short period, when output could vary but capital is given (ibid.: most of the book); and the long period, when even capital is variable (ibid.: mainly chs 16 and 17). These are analytical constructs, allowing the theorist to separate factors which adjust quickly from those that move more slowly.
Although artificial, the relevance of the separate periods is assured by their strong correspondence with processes in actual time. Keynes made one modification to Marshall’s short period: net new investment was a contribution to aggregate demand, but the resulting change in the capital stock was not allowed to affect the cost of producing output (aggregate supply).Uncertainty Immediately in GTs chapter 3, (Keynes 1936 [1973]) we see how an uncertain future shapes the determination of output and employment, the central concerns of the book. Output is decided by producers whose capital equipment is given: the decision takes place in the short period. Output takes time to produce, so the market for it lies in the future, which by definition is uncertain. Producers must form expectations of what the market will be like and decide their output and employment accordingly. For the first time we have an explicit link to Keynes’s earliest scholarly work, the Treatise on Probability; here is an exploration in economics of rational decision-making in the face of uncertainty. Also, for the first time in economics, we have a method which is consistent with the world economic actors actually live in.
The expectations of producers are central, but Keynes carefully reflects also on what expectations are important for other macroeconomic groups: households (GT, Keynes 1936 [1973]: chs 8 and 9), the investors in real capital (ibid.: ch. 11) and the financial speculators (ibid.: ch. 12). Economic activities with the shortest time horizons are in general the least uncertain; for instance, daily consumption is quite predictable, while investment decisions are made in the light of longer-term, and therefore more uncertain, expectations. This is the opposite of the theory of TM and of neoclassical theory today, where unexpected events (for example, windfalls) can occur in the short term, but long- run equilibrium is certain and known. (Financial speculation is an exception, in having a short time horizon but being very uncertain.)
The path-dependent system that Keynes created in the GT is an example of an open system, in this case open to future developments, which cannot be known.
Equilibrium With this new method comes a redefinition of equilibrium. It is not an endpoint or a configuration existing out of time, but a configuration contingent on the analytical constraints imposed (for example, the short period) or the policy stance assumed (monetary policy is mentioned explicitly); these are held constant for the purpose of the analysis. Equilibrium can pertain to a part of the system or to the whole: for example, “the equilibrium level of employment [is] the level at which there is no inducement to employers as a whole either to expand or to contract employment” (GT, Keynes [1936] 1973: 27), and chapter 18 outlines the equilibrium of the system as a whole (ibid.: 249). There is no suggestion of market clearing; this is important, for unemployment had continued at a high level for 15 years when the GT was published. Keynes characterized persistent unemployment as unemployment equilibrium and in the GT provided a theory which could explain how such an equilibrium (as he defined it) could occur.
For large parts of the analysis Keynes takes expectations as given. This allows him to get definite results. However, he relaxes this assumption, to show where greater realism will lead. First he varies short-term expectations, then also long-term expectations. The latter is “Keynes’s model of shifting equilibrium [which describes] an actual path of an economy over time chasing an ever changing [analytical] equilibrium - it need never catch it” (Kregel 1976: 217).
The fallacy of composition Finally we must mention the appropriate level of analysis when analysing the economy as a whole.
Though an individual whose transactions are small in relation to the market can safely neglect the fact that demand is not a one-sided transaction, it makes nonsense to neglect it when we come to aggregate demand. This is the vital difference between the theory of the economic behaviour of the aggregate and the theory of the behaviour of the individual unit...
(GT, Keynes 1936 [1973]: 85)Keynes indicated in the above passage that if one were to argue from individual behaviour to developments in the economy as a whole, the wrong answer is likely to be obtained. This occurs because the economy as a whole is not the same as the sum of its parts, and to argue from the part to the whole in such a circumstance is to commit the fallacy of composition. The most famous example of the conflict between the individual and aggregate levels in the GT concerns saving behaviour. Collectively, aggregate saving must equal aggregate investment, and investment is the variable determined by factors outside the theoretical system, so saving must conform to investment. Neither individually nor collectively do savers know this, but if there is a rise in the desire to save while investment stays the same, the shift away from consumption will decrease sales and firms will cut back output. Aggregate income will fall and savers will not be able to realize their original saving plans. Simply to aggregate the plans is to ignore the feedback which is forthcoming from other actors in the economy. This little story is known as the paradox of thrift, but it is not really a paradox, rather an example of applying the wrong level of analysis.
Another example concerns the labour “market”. Keynes’s contemporaries reasoned that if the supply of labour exceeded demand (that is, there is unemployment), the price of labour - the wage - must be “too high”, above its equilibrium level. Therefore to reduce unemployment, reduce wages. This is an example of the method practised by Marshall of looking at one market at a time: partial equilibrium analysis. Significantly, he did not apply this method to the labour market. Marshall understood that labour is only hired by producers if they expect to be able to sell the output the labour will produce. He referred to the derived demand for labour. Keynes generalized this principle to the macroeconomic level. When the demand for labour is correctly perceived as dependent on producers’ expectations of the future market for output, the policy conclusion is to stimulate demand, not to cut wages.