Keynesianism in One Country

Lester Thurow

Godley and Cripps devote their first seven seven pages to acknowledging the storms that are raging around the subject of macroeconomics. Deteriorating economic performances, and monetarists, ‘converted’ governments ‘to the idea that it was impossible for them to control output and unemployment. All they could do, the new story went, was create conditions (including the elimination of inflation) in which enterprise could flourish.’ They then spend the next 284 pages building a formal algebraic macro-economic model that will sail out of the storms of intellectual confusion. Finally they reach the safety of a two-page epilogue: ‘Our claim,’ they say there, ‘is to have provided a framework for an orderly analysis of whole economic systems evolving through time.’ They continue:

This logical framework is neither ‘monetarist’ nor ‘Keynesian’; it is non-denominational both in theoretical and political terms ... The most important result is, we believe, to re-establish the quintessentially Keynesian principle of effective demand as the determinant of real output and employment. In a closed economy real demand itself is determined, with quite a short time-lag, by fiscal policy; the effects of monetary (as distinct from fiscal) policy are transitory. The rate of inflation is, largely indeterminate in terms of economic forces and, although it has a dynamic of its own, is likely to be rather unstable.

Personally I agree with the nature of the building job upon which they embark and that it is a safe port which they eventually reach, but unfortunately their journey will only be persuasive to those of us who have already reached the same destination by independent means of transportation.

No formal model can accomplish what they seek to accomplish with theirs. Models can be built with behavioural hypotheses that lead to the Keynesian result, but models can also be built with hypotheses that do not lead to the Keynesian result. Godley and Cripps may see their model as a ‘non-denominational’ model, but every monetarist or rational-expectationist is going to see it as a Keynesian economic model. Keynesian assumptions in; Keynesian results out.

The point is that the existence of a financial sector does not, as Godley and Cripps seem to think, make a non-denominational, non-Keynesian model. One could agree to money-supply rather than interest-rate targets at central banks without giving up one’s status as a card-carrying Keynesian. The heart of the difference between Keynesians and monetarists is not the role of money but beliefs as to whether the economy is self-regulating and will, if left alone, quickly return to full employment. For Keynesians, aggregate demand is not self-regulating but must be corrected with either monetary or fiscal policies. For the monetarists, no such need exists. But both sides can build formal models consistent with their positions.

While the formal Godley-Cripps model is a very good pedagogical tool for teaching students the nature of a Keynesian economy which includes monetary and financial sectors, a formal model, by its very nature, cannot be a proof that the Keynesian conclusions are true. Those who are unconvinced to start with will remain unconvinced. Those who are uncertain will remain uncertain. For Godley and Cripps to have proved that their conclusions were right they would have had to have proved that their behavioural hypotheses were the right ones and that the behavioural hypotheses of the monetarists and the rational-expectationists are the wrong ones. And this they do not attempt to do. As a result, there is a certain lack of relevance in their effort. They are fighting the good fight, but on the wrong front.

If you want to fight the battles that Godley and Cripps want to fight you have to start with a fundamental contradiction within economics, which is that what is taught in conventional micro-economics is incompatible with what is taught in macro-economics. In micro-economics every market is a price-auction market that clears through competitive bidding within a framework of supply and demand. With every individual a maximiser in his decisions to produce and consume, there are no unsatisfied bidders and the market is always in a state of equilibrium. In this situation it is impossible to have over or underemployed resources. Every factor of production which is to be employed is employed at a wage or price governed by that productivity. If an employed worker, for instance, really wants to work, he has only to lower his wage request and some employer will hire him and, if necessary, fire someone else. The labour market quickly reaches a point where everyone willing to work at or below the equilibrium wage is working, while those who stubbornly keep asking for more than they are worth remain unemployed. Given such markets, macro-economic policies, whether monetary or fiscal, which are designed to raise aggregate demand in order to eliminate unemployed resources are not only unnecessary but positively harmful. No one should intervene to improve the market’s performance, because its results are the best results it is possible to get. Interventions of any sort can never help, only hurt.

Similarly, according to this conception, inflation either can’t exist or doesn’t matter. It can’t exist because a price increase in any one area (say, an increase in the price of energy) will force reductions in demand – less income is available to be spent after paying for energy – and thus offset price and wage decreases in other areas. Inflation does not matter because the perfectly rational homo economicus, when making decisions or judging his economic success, looks only at his real income and the relative prices of various goods, and inflation affects neither of these.

If the price-auction equilibrium market adequately described the real world, every economist would agree with every other economist. If all markets (over time, across space, and those for insuring risk) existed and functioned perfectly in the way the model asserts, then a competitive economy left to itself would generate the best possible economic outcome, given the real resources and preferences of the population. No government intervention could improve upon the results. There would be no involuntary unemployment, inflation would worry no one, and growth would occur at the highest rate consistent with the preferences of the members of society. The only interventions that might be called for would be those flowing, not from economics, but from the ethical principles governing income distribution and initial endowments of resources.

Disputes arise over whether the necessary markets do or could exist and, if they exist, over the transaction costs which using them would entail. The real economic world is obviously not a literal auction market. But does it proceed ‘as if’ an auction were occurring? Put another way, does the real world come close enough to the competitive model for it to be described in accordance with the model? Sceptics point to the non-existence of certain markets – namely, those that would allow a person to trade between the present and the distant future – and to the transaction costs of implementing other markets. Many of the markets for selling risk and uncertainty either don’t exist or involve such substantial transaction costs that few individuals use them. Within the economy there are many markets where only a few firms exist and where entry costs are high.

There are also major controversies among economists as to how long it takes markets to adjust. Given that the concept of equilibrium is only relevant if markets get there rather rapidly, speedy adjustments are crucial to the price-auction view of the world. If adjustments are not rapid, one must build a micro-economic model based upon fixed prices, since the real world always lives in the fixed-price short run and never in the flexible-price long run. Conversely, there is also no disagreement as to what would happen if wages and prices were rigid downward – the basic Keynesian assumption. Demand is not self-regulating. External shocks – a rise in the price of oil, for instance – lead to inflation and unemployment. An oil price hike raises the average price level: oil prices go up and other prices fail to go down because wages fail to fall. This reduces demand for other goods and services – less disposable income is available after more has been paid for energy. Faced with lower sales, firms outside the energy areas lay off workers, unemployment rises, incomes fall, and a reduction in aggregate demand results. Stagflation commences. Keynesians would say that to reduce unemployment one must raise aggregate demand. However, Keynes himself had no answer to the question of what should be done to curb inflation in an economy of rising unemployment. The problem simply did not exist in his time. His present-day followers would advocate structural remedies – wage and price controls, incomes policies etc – to make the economy less inflation-prone.

It is important to understand that there are no theoretical controversies about what happens in a fixed or flexible-price world. The most committed monetarist or rational-expectationist would admit that if wages were rigid, downward markets would not clear. They simply deny the ‘if’ clause. The controversies are empirical: which of the two models better describes the state of the world as it actually is? And because there are smart people on both sides of the dispute, deciding which is more accurate is not easy. If it were, the issue would have been resolved long ago.

The weight of economic opinion now lies against Godley and Cripps. Why? Because almost everyone uses the flexible-price model in micro-economic analysis. And in some markets, such as the grain market, we find that prices clearly do adjust in accordance with the basic axioms of supply and demand. The weakness of Keynesian analysis (and those of us who believe in what Godley and Cripps believe in) is that there is no micro-economic theory to defend the macro-economic assumptions in other markets, such as the one for labour, where wages do not seem to adjust as they should. Any self-sustaining Keynesian macro-economics needs to be buttressed with Keynesian micro-economic foundations, but this is precisely what Keynes himself did not provide.

Godley and Cripps seem genuinely puzzled as to why other are worried about the micro-foundations of macro-economics. ‘One extensive area of controversy concerns the so-called “micro-foundations” of macro-economics. The view is widely held that no relation between aggregates (between, say, total income and total expenditures) can validly be postulated which cannot be justified precisely in terms of the behaviour of individual agents. Such a view seems perilously close to a denial that macro-economics as defined in the opening sentence of this book (“the study of how whole economic systems function”) can be a valid subject at all.’ The puzzle is not why others are worried, but why Godley and Cripps cannot see the problem.

No one objects to relationships between total income and total expenditures. The problem is that there are micro-economic models that postulate the rapid accomplishment of equilibrium, and macro-economic models that postulate the persistence of disequilibrium. Both sets of postulates cannot be right. If the micro-economic postulates are right, then the macro-economic postulates cannot be true. If the macro-economic postulates are right, then a new consistent micro-economics must be built. Monetarists and rational-expectationists solve the dilemma by junking the disequilibrium postulates of macro-economic models. If Keynesian-type macro-economic models are to prevail, their proponents are going to have to justify the junking of at least some of the equilibrium postulates of micro-economics. And this means that they have to have an alternative theory of why these markets work as they do and why they do not function in accordance with the axioms of competitive supply and demand.

To use the profession’s old way out of this dilemma – ‘market imperfections’ – simply is not acceptable. The rational-expectationists are right: no real market imperfection could last for any length of time without the protection of government; and if that is the heart of the problem, then government ought to be taken out of the economy. The solution is not to build macro-economic models which take the existence of these ‘market imperfections’ into account, but to eliminate the market imperfections so that macro-economic models and the government interventions that flow from them are no longer necessary.

The new consistent micro-economics must begin on the supply side of the economy – a part of the economy that is not represented in the Godley-Cripps model at all: ‘There is no theory of the “supply side” of the economy in this book. We are in any case sceptical about the usefulness of theory, logic or accounting to yield useful results about the aggregate productive potentials of an economy.’ This is a very strange slippery-slope argument. It is a very small step from denying the usefulness of supply-side equations (presumably they are not needed because the supply side takes care of itself in competitive markets) to believing that demand-side equations are not needed either. Since competitive markets solve the demand problem as well as the supply problem, there is no need for the help of macro-economic models and the government interventions that go with them.

If Godley and Cripps had attempted to build a supply side of the economy they might have begun to see the nature of the problem. For the problem lies precisely where Keynes himself identified it – in the labour market. Wages don’t simply fall, as they should, with rising unemployment. To explain this fact while retaining a belief in conventional supply-and-demand models requires unbelievable theoretical contortions. One such contortion is to argue that wages don’t fall because of the monopoly power of trade unions. Yet in America only 13 per cent of the work-force is unionised. Whatever one believes about the monopoly power of unions, it is difficult to explain how they stop the wages of those who are not union members from falling in the face of unemployment. From a theoretical point of view, rigid union wages should intensify the downward wage pressures in the non-union sectors of the economy.

If the Keynesian analyst is to avoid these contortions there are three words, now completely absent from the vocabulary of conventional micro-economic price theorists, which will have to be reintroduced. They are motivation, co-operation and teamwork. In the standard economic model of labour supply, workers are always 100 per cent motivated and 100 per cent co-operative (or their pay is individually adjusted to their degree of motivation and co-operation), and the concept of teamwork is completely absent: total output is simply the statistical summation of individual marginal products. But given a real world where employees can of their own accord adjust their motivation from zero to 100 per cent, where co-operation can be offered or withheld, where teamwork can be disrupted while on the job or by leaving the job, and employers know that it is very difficult and expensive to ascertain each employee’s performance at each period of time and impossible to enforce motivation, co-operation and teamwork with economic threats, there is no mystery as to why employers do not reduce wages whenever unemployment appears. At first glance, wage reductions might appear to be smart cost-cutting moves, but when actually imposed, the wage cuts so disrupt motivation, co-operation and teamwork that the short-run gains are more than lost in the long run: reductions in motivation, co-operation and teamwork follow.

Insofar as economists have a theory of human behaviour, it is the 19th-century theory that individuals act rationally to maximise their utilities. Whenever that theory is used to describe reality, it has to be finessed with unobservable variables. Revealed preferences make everyone into utility maximisers in their consumption behaviour, but the theory, because it relies on revealed preferences, has no predictive value: people simply do what people do. In a very real sense, economists do not have a theory of consumption. On the other hand, how can you have a valid theory of micro-economics without a valid theory of consumer behaviour?

Godley and Cripps try to rescue macro-economic models by building yet another macro-economic model. But no formal macro-economic model can rescue the species from extinction. That can only be done by rebuilding the environment of micro-economic models within which macro-economic models live. And even if this were done, there would still be one final problem for those of us who are Keynesians. The Godley-Cripps model portrays a closed economy, an economy that is not subject to international trade. With open economies it may prove to be the case that even if Keynesian aggregate demand manipulations are necessary, they are not possible in any one country. The world is simply too integrated. Global Keynesianism is alive. Single-country Keynesianism, as the French experience of 1981-82 shows, is dead.