The closing decades of the nineteenth century saw the rise of socialism as a large-scale movement. Socialist parties were formed across Europe, and in many countries their support grew rapidly. Extension of the franchise to include the working class led to an expansion of socialist representation in European parliaments. There was great pressure for social reform, both from socialist parties themselves and from conservatives (such as Bismarck, the German Chancellor, and Disraeli, the British Prime Minister) who sought to lessen the pressure for more radical change. Government activity was extended into many new fields, new organizations emerged, and the role of the state increased. Labour unions were expanding to include unskilled as well as craft workers, and were beginning to exact improved working conditions. Though there were clear exceptions, the rise of trade unions and the rise of socialism were strongly linked.
The socialist movements that arose across Europe and in the United States took many forms. They covered a spectrum ranging from mild reformism to revolutionary Marxism. For economists, the rise of social-ism presented two types of challenge. The first was to develop principles for working out the appropriate role of the state. When, where and how should the state intervene in economic life? The second was to evaluate socialist and communist schemes for reorganizing society. Could an economy organized on socialist principles operate successfully? Though these questions clearly overlapped, they provide a useful way to think about some of the main lines of economic thought during this period.
The challenges posed by socialism came at the time when economists were increasingly taking up marginalist theories. These theories provided a framework within which problems such as the regulation of industry, the provision of welfare benefits, the establishment of government enterprises and tax policy could be tackled that was very different from that available to previous generations of economists. Smith and J. S. Mill had discussed the problem of state intervention, but their analysis had centred on long-run growth. They offered general principles by which state activities could be judged, and their observations on specific cases contained many perceptive insights, but their ability to tackle specific questions about how resources should be allocated was severely limited. Marginalism, with its mathematical apparatus of utility and profit maximization, appeared to be able to fill this gap.
Some of the early marginalists – notably the Austrians – acquired a reputation for being hostile to socialism. Some undoubtedly were. For the rest, however, although they may generally have been biased in favour of laissez-faire and individualism, this was in practice outweighed by more pragmatic considerations. Most marginalists were on the side of reform, even if their approach was sometimes paternalistic or if they were hostile to radical change. For example, although Jevons started his career a supporter of laissez-faire, by his last book, published in 1882, he had arrived at a position where he saw ‘hardly any limits to the interference of the legislator’.1 What had happened was that, during the 1870s, he found more and more contexts where state intervention was justified, to be financed mostly by continual increases in local taxation: public health, working conditions, education, transport, and many others. Marshall, the dominant economist of the following generation, saw a smaller role for state intervention than did Jevons. However, he still assigned a significant role to the state, going along with the wider movement towards support for progressive taxation (where the rich are subject to higher tax rates than the poor). Though his socialism was somewhat limited, Walras even described himself as a socialist. If there was a causal link between socialism and marginalism, therefore, it did not involve marginalism being adopted as a way of defending laissez-faire against socialist criticism. Marginalism was used to argue in favour of social reform.
In the English-speaking world, the dominant approach to problems of social welfare and reform was, for several decades, the Cambridge tradition, which originated with Henry Sidgwick (1838–1900). The fundamental part of Sidgwick's argument was a distinction between two senses in which the term ‘wealth' was used. The first was as the sum of goods produced, valued at market prices. The second was as the sum of individuals' utilities – what we would now term welfare. He offered reasons why these might be different. The clearest example is free goods: goods for which no price is paid. Such goods raise individuals' utilities – people value them – but they do not enter into the first concept of wealth at all, for their price is zero. More generally, the market value of a good to a consumer will measure the value of the last unit consumed. If the value of an additional unit falls as consumption rises, this will be less than the average value of the good to that consumer. If the ratio of price to average utility were the same for all goods, this would not matter at all. However, the ratio of price to average utility will depend on how fast marginal utility falls, and there is no reason to suppose that this will be the same for all goods. Some goods have an average utility that is high relative to their price and will be undervalued in calculations of wealth at market prices. Free goods, which have positive value but a zero price, are enough to make this point.
In developing these arguments, Sidgwick made use of Jevons's utilitarianism, according to which individuals' utilities could be measured and compared. This meant that if the marginal utility of a particular good were higher for one person than for another, total utility could be raised by redistributing goods to those who valued them most. This would leave wealth at market prices unchanged. For example, the value of an additional loaf of bread to a poor person may be higher than its value to a rich person; it may therefore be possible to increase welfare by taking a loaf from the rich person and giving it to the poor person. In other words, a community's welfare depends on how goods are distributed, not simply on the value of goods being consumed.
Having provided reasons why wealth and welfare might differ, Sidgwick argued that, for practical reasons, wealth had to be measured using market prices, except in specific cases where ‘the standards of the market fail us’.2 This provided the justification for an approach to welfare economics similar to that of the classical economists such as Smith and Mill, tackling problems of welfare by analysing first the production and then the distribution of wealth. Sidgwick also followed Mill in his analysis of the role of government. The general principle was laissez-faire, but this was subject to numerous exceptions that he explored in detail. These included cases where individuals could not obtain adequate reward for the services they provided to society (lighthouses, afforestation and scientific discovery) and also those where the gains to individuals exceeded those to society (duplicating an existing railway line). There were also cases, such as the control of disease, where cooperation was required. However, even though Sidgwick defended the classical perspective, his separation of two concepts of wealth made it possible, arguably for the first time, to conceive of welfare economics as something distinct from economics in general.
Marshall followed in the same tradition, but made Sidgwick's analysis more precise. He defined wealth very clearly as a sum of money values – national income, or national dividend as he called it – that was distinct from utility or welfare. His main contribution, however, was to develop a way to measure utility in terms of money. This was the theory of consumers' surplus. An individual consumer's surplus is the difference between what the consumer is willing to pay for a commodity and what he or she actually pays. Marshall showed that such surpluses could be added together and used to measure changes in social welfare only under certain circumstances: in particular, the value of an additional unit of income had to be the same for all individuals. In general, this would not be true. His response was to confine his use of consumers' surplus to situations where it could plausibly be argued that it was approximately true.
On the whole, however, it happens that by far the greatest number of events with which economics deals, affect in about equal proportions all the different classes of society; so that if the money measures of the happiness caused by two events are equal, there is not in general any very great difference between the amounts of happiness in the two cases.3
Marshall confined his use of consumers' surplus to goods that accounted for only a small proportion of consumers' spending. This meant that a change in price would have a negligible effect on real income and would have only a minor effect on the value to the consumer of an additional unit of income.
The practical, utilitarian approach to welfare economics reached its culmination in the work of Pigou (who gave the subject its name), in his two books Wealth and Welfare (1912) and The Economics of Welfare (1920). Pigou's welfare economics was utilitarian, in that he regarded the elements of welfare as ‘states of consciousness' that could be compared with each other.4 Like Sidgwick and Marshall, he focused on national income and the way in which it was distributed. National income was linked to what he called ‘economic welfare' – ‘that part of welfare that can be brought, directly or indirectly, into relation with the measuring rod of money’.5 In other words, he recognized that there were aspects of welfare about which economists could say little.
Pigou's main innovation was to replace Marshall's concept of consumers' surplus with an analysis of marginal private and social products. If the marginal private product of an activity (the benefits obtained by the person undertaking the activity) were different from its marginal social product (the benefits to society), welfare was unlikely to be maximized. There were, Pigou argued, many situations where private and social products would be different from each other. One was where one person owned an asset (for example, land or a building) that was managed by a tenant. If the benefits accrued to the landlord, the tenant might have no incentive to improve or even to maintain the asset. The marginal social product of improving land, therefore, would be higher than the marginal private product to the tenant. Another situation was where one person's activities directly affected someone else's welfare. The obvious examples of this are pollution and traffic congestion. Monopoly would also cause private and social products to differ: instead of simply looking at the value of the additional output, the monopolist will also take account of the effect of increased sales on the price of goods that are already being produced. Economic policy therefore involved eliminating differences between marginal private and social products. Using this approach, Pigou offered a detailed programme for economic policy, virtually providing a blueprint for the welfare state.
Unlike the Cambridge economists, Walras and Pareto at Lausanne did not assume that the welfare of different individuals could be measured and added together. Instead, Walras started from the notion of justice in exchange – ‘commutative justice’. He argued that this type of justice required that every trader faced the same price for a given product and that prices did not change. He then showed that, given justice in exchange, free competition would produce maximum welfare. The significance of this result was that it offered a way in which questions of welfare could be analysed without either adding up or comparing the well-being of different individuals. This approach was developed by Pareto, who defined a social optimum as a situation in which any change would be agreeable to some individuals and disagreeable to others – in other words, a position where it was impossible to make anyone better off without making someone else worse off.
Though Walras, like his English counterparts, proposed detailed policies of social reform, centred on getting rid of monopolies, he also considered the question of socialism at a more abstract level. Central to this was a proposal for land nationalization that would, he contended, provide a way to reconcile individualism and socialism. Pareto took the discussion of socialism a stage further, paying attention to the question of how a socialist state might be organized. He observed that, even if the state owned the entire stock of capital and prohibited all buying and selling, prices and rates of interest would have to remain, at least as accounting entities:
The use of prices is the simplest and easiest means for solving the equations of equilibrium; if one insisted on not using them, he would probably end up by using them under another name, and there would then be only a change of language, and not of things.6
Without prices and interest rates, ‘the ministry of production would proceed blindly and would not know how to plan production’.7 Individuals' desires and the obstacles to satisfying them would be the same under a collectivist organization of society as under capitalism, with the result that both societies would have to solve similar problems. The main difference between socialism and capitalism was the principles by which the distribution of income was determined. Under capitalism, incomes were linked to ownership of means of production (and hence by the way in which society has evolved), whereas under socialism they were determined according to ethical and social considerations.
Pareto's argument was in turn taken a stage further by one of his students, Enrico Barone (1859–1924). Barone pointed out that the same conditions had to be fulfilled in a collectivist economy seeking to maximize the welfare of its members as in a perfectly competitive equilibrium. The ministry of production in a socialist state could start with the prices and wages inherited from the previous regime. It could then raise or lower them, in a process of trial and error, until two conditions were fulfilled: prices were equal to costs of production and costs of production were minimized. These arguments led him to claim that such a ministry would face an immense task, though not an impossible one.
The period of ‘war communism' in Soviet Russia in 1918–21 was a brief attempt to dispense completely with markets and prices – the basis of capitalist economies – replacing them with centralized planning. This resulted in chaos, and was followed in the early 1920s by the New Economic Policy, which reintroduced markets for many goods, though maintaining extensive state control over the economy. The time was thus right for a more detailed examination of socialism. Had the Soviet experiment in instituting a non-market economy collapsed because of the intense pressures created by wartime, or because it was theoretically flawed? Several economists took up the challenge of showing that it was the latter.
One such economist was Gustav Cassel (1866–1945), who used the example of a socialist economy to make certain points that applied to any exchange economy. The socialist economy had the advantage that it was the simplest possible economy, with the result that it offered a benchmark against which more complex economies could be assessed. Comparison with socialism would reveal which institutions were essential and which could be dispensed with. This led him to elaborate on Barone's point that, even if a socialist state tried to dispense with prices and wages, these would inevitably re-emerge, for they reflected fundamental economic realities. However, he went further than Barone in arguing that, in the absence of private property and a fully developed system of exchange, a socialist state would be unable to direct production in the best way. The necessary prices would not be available.
The economist who provided the most radical critique of socialism was Mises, in his article ‘Economic calculation in the socialist commonwealth' (1920). This provoked what has come to be known as the socialist-calculation debate, in which many of the period's leading economists participated. In his article, Mises argued, in uncompromising terms, that socialism was impossible – it could never work. His reasoning was that, in any economy, rational calculation required the existence of freely established money prices for both consumers' and producers' goods. Without such prices it would be impossible for anyone to work out how resources should best be used. This was, Mises emphasized, not a purely technical problem, as some socialists seemed to assume. The main difficulty arose not with consumer goods (one might not need prices to say, for example, that 1,000 litres of wine was more valuable than 500 litres of oil) but with producers' goods. A railway, for example, is valuable because it reduces costs for other industries, enabling them to produce more of the goods that consumers require. Without money prices, it would be impossible to calculate whether or not it should be built.
In a static economy, where nothing changed, rational calculation might be possible. A socialist state could continue the pattern of production that prevailed under a previous competitive system. However, the world is not static. Tastes and technology are forever changing, with the result that new ways of producing goods have continually to be worked out. In a socialist state, there would be no one with the responsibility and initiative to change the way in which activities were organized in response to these changes. Managers of capitalist enterprises, Mises argued, have an interest in the businesses they administer that is quite different from anything that could be found in public concerns. ‘Commercial-mindedness' will not exist when people are moved from business into public organizations. Even if human nature could be changed so that people all exerted themselves as much as if they were subject to the pressure of free competition, there would still be a problem. In the absence of prices, people would not know what it meant to economize – to balance the costs and benefits of alternative activities.
The main response to Mises came from a group of economists who have come to be known as ‘market socialists’, including Fred M. Taylor (1855–1932), H. D. Dickinson (1899–1969) and Oskar Lange (1904–65). The reason for this label is that they argued that it was possible to design an economy that was socialist in the sense that the state owned the means of production but in which there were markets for consumer goods and labour. Households would thus be free to sell labour and to buy consumption goods in response to market wages and prices. Production would be organized by plant managers, who would be given the task of producing at minimum average cost and setting prices equal to marginal cost (the cost of an additional unit of output). Behind these plant managers would be industry managers, who would make investment decisions, including when to open new plants and close old ones. A central planning board would monitor the whole process, setting the prices on which the decisions of industry managers would be based.
The reasoning behind these rules was that, if they were followed, it would be possible for a socialist economy to mimic the behaviour of a perfectly competitive one. If the market-socialist system were correctly administered, both systems would give the same outcome. There might be practical problems with socialism (no one disputed this), but it was argued that socialism was theoretically possible.
The most forceful response to this came from Hayek, in a series of articles the first of which appeared in 1935. Hayek argued that the market socialists had not shown that rational calculation was possible under socialism. They had just shown that if one had complete knowledge of all the relevant data (including knowledge of consumers' tastes and of all the technical possibilities for producing goods) it would be possible to solve a set of equations to determine what goods should be produced. However, this did not solve the problem of how efficiency could be achieved under socialism it showed that it had not been understood. In a real-world economy, full information on technical conditions of production does not exist. What does exist is engineers with techniques of thought that enable them to discover new solutions when confronted with new problems. In other words, the knowledge required by socialist planners does not exist it needs to be created. This means that the initiative in adopting new methods, developing new products and so on has to come not from planners, but from managers who are aware of new developments and are able to respond to them. The problem with socialism is not merely a computational problem: it is one of generating the information required for the system to operate. The market socialists, by taking technical conditions as given, simply assumed the problem away.
Hayek also raised further problems with the market-socialist arguments. In equilibrium, prices can be calculated by solving a set of simultaneous equations. But the economy never is in equilibrium. It is not clear how the planners should operate out of equilibrium. It might not even be appropriate to start with existing prices, for there was no reason to believe that the transition to socialism would not produce large changes in equilibrium prices. Such problems would be compounded by the problem of new goods: planners would have no idea about which new goods should be produced and in what quantities. Comparisons with state enterprises in a capitalist economy would provide no guidance, for it would no longer be possible to make comparisons with the private sector.
This critique of the so-called ‘competitive solution' to the problem of socialist planning was developed by Hayek into a theory of competition that differed radically from the one that had, by the 1930s, come to dominate the profession. Where the theory of perfect competition focused on an equilibrium in which no firm was able to affect the prices it faced, Hayek focused on rivalry. The essence of competition was that businesses competed with each other, discovering new technologies and new ways in which production could be organized. The importance of the market was that it provided a means whereby decentralized decision-making by individual firms could be coordinated. Prices conveyed information that would not otherwise be available to decision-makers. Competition was not only a means of moving the economy towards equilibrium, but also a procedure for discovering new ways of doing things.
The socialist-calculation debate overlapped with another controversy that arose in the 1930s. This was about the foundations – of welfare economics. Lionel Robbins argued that there was no scientific basis on which interpersonal comparisons of welfare could be made. Though people made such judgements all the time, they should not form part of the science of economics. This undermined the foundations of the Cambridge tradition in welfare economics. There was therefore a need to rebuild the subject. The outcome was what came to be called the ‘new welfare economics’, developed by Lange and a group of Robbins's younger colleagues at LSE, notably Hicks, Abba Lerner (1903–82) and Nicholas Kaldor (1908–86). There was a close link with the calculation debates, for it was impossible to ask whether a socialist economy could operate efficiently without examining what an efficient allocation of resources might look like. Lange, one of the architects of market social-ism, was also a major contributor to the new welfare economics.
The main contribution of the new welfare economics was the development of the concept that came to be known, using the term coined by Ian Little (1918–) in 1950, as ‘Pareto optimality' or ‘Pareto efficiency' (the two terms are used interchangeably). This is the situation, described by Pareto, where it is impossible to make one person better off without making someone else worse off. The contribution made by Hicks, Lange, Lerner and their contemporaries in the 1930s was to work out the conditions that had to be met if this condition were to be satisfied. There was, however, a problem with the criterion of Pareto optimality and the associated concept of a Pareto improvement (a change that would make at least one person better off without making anyone worse off): they failed to provide any guidance on real-world policy changes, which virtually always benefited some people and harmed others. A stronger criterion was required.
Hicks and Kaldor, again taking up an idea found in Pareto, tried to strengthen the Pareto criterion by introducing the idea of a ‘compensation test’. A change would be beneficial if the gainers could compensate the losers and still remain better off. If this criterion were met, the result would be a potential Pareto improvement. It would not be an actual Pareto improvement, of course, unless compensation was actually paid; however, the concept of a compensation test was thought to provide a way in which the question of whether resources were being used efficiently could be separated from questions of income distribution. But the idea turned out to be flawed. Tibor Scitovsky (1910–) showed in 1940 that it was easy to find examples where the compensation test would be satisfied in both directions: it gave contradictory results.
In the course of these discussions, economists approached the problem of social welfare in many different ways, often deriving different versions of the conditions for a social optimum. One of the main problems was that, though economists wrote of ‘optimality' and ‘ideal output’, it was never made clear exactly what it was that was optimized in a social optimum. An answer was provided by Abram Bergson (1914–), who proposed the idea of a social-welfare function. This was a relationship between social welfare and all the variables on which social welfare might depend. In itself, this was entirely devoid of content: it simply stated that social welfare depended on whatever variables it depended on. However, it provided a framework within which different approaches to the problem could be analysed. It was possible to use the social-welfare function to analyse the implications of different value judgements or ethical criteria. For example, individualism implies that the only variables entering the social-welfare function are variables affecting individuals' levels of well-being. The Pareto criterion implies that if an individual's welfare increases (without anyone else's changing) social welfare must increase. Using such arguments, it was possible to clarify the meaning of the concept of Pareto optimality and to resolve the paradoxes surrounding compensation tests.
During the 1950s economists worked extensively on welfare economics. Kenneth Arrow's Social Choice and Individual Values (1951) completely reoriented welfare economics by proposing a social-welfare function that was very different from Bergson's. Arrow thought of a social-welfare function (or social-choice function) as being similar to a voting mechanism. Every voter has a preference for a particular political party, and a voting mechanism is a rule that translates such individual preferences into a social choice (a government is elected). Possible mechanisms include simple majority voting as well as much more complicated procedures. Arrow viewed the problem of social choice in exactly the same way – as the problem of getting from individuals' views about how society should be organized to a social decision.
The way in which Arrow managed to say anything about such an abstract problem was by specifying a list of conditions that any voting procedure, or social-choice function, should satisfy. These included conditions such as ‘If everyone prefers A to B, then A should be chosen' (this is known as the Pareto principle); ‘No individual should be a dictator’; and so on. He then proved that, although every condition looked extremely reasonable, there was no social-welfare function that would satisfy all of them. This was his so-called ‘impossibility' theorem. It stimulated the emergence of an entirely new field of economics – social-choice theory – which had strong links with the analysis of voting rules by political scientists.
At around the same time, Arrow, together with Gérard Debreu, formulated what have become known as the two ‘fundamental theorems of welfare economics’. These results formalized what had been discovered in the 1930s with the new welfare economics. The first theorem is that every competitive equilibrium is Pareto efficient. In other words, that in a competitive equilibrium it is impossible to make anyone better off without making someone else worse off. The second theorem approaches the problem the other way round. It is that any Pareto-efficient allocation of resources can be made into a competitive equilibrium, provided that income is distributed in an appropriate manner.
The Arrow–Debreu theorems mark the culmination of a particular approach to welfare economics, as their inventors' existence proof did for the theory of general competitive equilibrium. They establish all that can be said about the merits of perfect competition as a way in which to allocate resources. Their limitations, however, are that Pareto optimality is an extremely weak optimality criterion and that they tell us nothing about what happens when some of the criteria for optimality are not satisfied. For example, if there are monopolies in several other industries, will it be socially beneficial to remove a tax that distorts incentives in a particular industry? In 1956 Richard Lipsey (1928–) and Kelvin Lancaster (1924–), worked out their theory of the ‘second best’, which showed that this would generally not be the case. If there were distortions in other parts of the economy (such as monopolies or taxes) then removing a distortion was as likely to make the overall situation worse as to improve it.
The result of these developments was that by the end of the 1950s the outlook for welfare economics looked very bleak. The new welfare economics had failed to provide any welfare criterion stronger than Pareto optimality. Arrow's impossibility theorem had shown that there was no acceptable way to get from individual preferences to a social preference. Lipsey and Lancaster had undermined the idea that piecemeal reforms could be shown to be beneficial. Arrow and Debreu had established the precise relationship between perfect competition and Pareto efficiency, but nothing could, in general, be said about whether actual policy changes would raise or lower social welfare.
The displacement of the ‘old' welfare economics of Sidgwick, Marshall and Pigou by the ‘new' welfare economics did not mean that the old problems were neglected. In 1951 Samuelson worked out the theory of pure public goods. Public goods are goods (like the services of a lighthouse, a healthy environment, or a public fireworks display) that, if they are provided at all, are provided for everyone. People cannot be excluded from benefiting from them, and one person can benefit from them without reducing the benefits available to anyone else. (The qualification ‘pure' is used to acknowledge that these conditions describe an ideal – problems of congestion, for example, mean that after some point many goods cease to exhibit these characteristics.) The significance of public goods is that, as Samuelson showed, the amount supplied will typically be less than the amount that is socially desirable. Everyone benefits, but no one has an incentive to pay. Similar problems arise with externalities (of which pollution is the main example), where one person's action causes harm (or possibly benefit) to a third party.
Public goods and externalities are both examples of market failure where competitive markets fail to allocate resources in a Pareto-efficient way. If allocations are not Pareto-efficient, it means that there can be unanimous agreement that a better allocation of resources is possible (at least one person can be made better off without anyone being harmed). These concepts have been widely used to justify government intervention. The government has the responsibility to provide goods that the market will not supply in sufficient quantities, and to use its power to tax in order to correct defects in the market mechanism. In the 1960s such beliefs fitted in well with the belief that the government also had to intervene at the macroeconomic level to ensure full employment. Since then, however, this rationale for government intervention has been challenged.
The first challenge arose with what has come to be called the ‘Coase theorem’, proposed by Ronald Coase (1910–) in 1960. Coase made the point that most discussions of externalities, like Pigou's, failed to take account of the legal framework within which economic activities were undertaken. The failure of markets to allocate resources efficiently should, Coase argued, be attributed not to a failure of competition but to the absence of clearly defined property rights. If property rights were clearly defined, markets could develop that would ensure efficient use of resources. For example, if the rights over the use of a river were clearly established, a factory owner wishing to pollute the river and fishermen with an interest in clean water could negotiate over the amount of pollution that would be allowed. If the factory owner held rights over the river, fishermen could pay him or her to limit pollution; if fishermen held the rights, the factory owner could buy the right to pollute. The result of this perspective was that Coase saw a much greater scope for the market and a more limited role for the state than did Pigou.
The second challenge to the conventional view of the role of government, also around 1960, came with the development of theories of how voters, governments and bureaucracies behaved. These theories, developed by economists such as James Buchanan (1919– ), Gordon Tullock (1922– ), Mancur Olson (1932–98) and Anthony Downs (1930– ), abandoned the notion that governments are disinterested organizations that act in the public interest (see p. 312). They replaced it with a view of governments as made up of individuals who are seeking to achieve their own ends. Politicians offer policies that will maximize support in elections. Managers run their organizations in ways that increase their own status and income. Taxes and government spending came to be seen as the outcome of political processes in which competing interests were expressed. The result was that the concept of government failure came to be placed alongside that of market failure.
Since at least the eighteenth century economists have been concerned with the question of whether the market mechanism is an effective way to organize economic activity. In this sense they have always been concerned with welfare economics. The major theme in Adam Smith's Wealth of Nations was that a system of natural liberty, or free competition as it came to be called, would promote economic growth and hence increase welfare. Producers would be led, as if by an invisible hand, to serve the public good even though they were concerned only with furthering their own interests. This was a theorem about competition and economic welfare.
During the period covered by this chapter, the way in which welfare economics was conceived changed dramatically. Theories of marginal utility provided a new way to analyse markets. Economists began to focus on whether the resources available at any moment were allocated efficiently. Concern with the growth of resources faded into the background. At the same time, economists began to think of competition in a different way. Instead of Smith's vision of natural liberty, in which competition meant actively competing with other people, competition came to mean a situation in which market power – the ability to influence prices – was absent. This change was clearly illustrated in the socialist-calculation debate, in which the market socialists – for many years perceived as the clear victors – defended socialism on the grounds that it was possible to design a socialist system in which resources would be allocated efficiently. They failed to recognize that Mises and Hayek, like Smith and the classical economists, had a different vision of what competition involved and of how the efficiency of an economic system should be judged.
Economists were, as so often in the history of the subject, also trying to make economics more ‘scientific’. In the 1930s many of them interpreted this to mean that value judgements should be eliminated from the core of the discipline. In this they were possibly influenced by the arguments made in philosophy by the Viennese logical positivists (and brilliantly conveyed to the English-speaking world by A. J. Ayer). The ‘old' welfare economics of Sidgwick, Marshall and Pigou was strongly criticized and replaced by the ‘new' welfare economics based on the principle of Pareto-optimality. It turned out, however, that few clear results could be obtained. The Pareto criterion was too weak a foundation on which to base welfare economics. However, the Arrow-Debreu theorems about the efficiency of a competitive equilibrium made it possible to claim that Smith's problem of the invisible hand had now been rigorously proved. What was less often noted, however, was that the interpretation of the invisible-hand theorem had changed dramatically. It was no longer (as it was for Smith) a proposition about the dynamic effects of competitive rivalry in the real world; instead it had become a theorem about optimal resource allocation in an abstract world where market power was absent.
From around the 1970s the situation began to change. The work of Buchanan, Tullock and others has already been mentioned. In addition, social-choice theory developed as an abstract discipline that sat somewhere between economics, ethics and political science, strongly influenced by Arrow's impossibility theorem. Social-choice theorists such as John Harsanyi (1920–) and Amartya Sen (1933–) explored issues such as whether it might be possible to measure individuals' utility, the nature of individual rights, and the ethical criteria on which social decisions might be based. More widely, economists began to use models not to provide a value-free science but to explore the consequences of different possible value judgements.
There were also important changes in the way in which markets were conceived. In the mid-1970s ‘Austrian' economics experienced a revival. It was actively promoted as an alternative to conventional economics, based on radically different conceptions of knowledge and the market process. However, although support for this approach grew, and Hayek once again became a widely known figure within the profession, it remained very much a minority tradition. Within the mainstream of the subject, economists began to construct models in which there was uncertainty about the future and information was scarce. Joseph Stiglitz (1943–) showed that, once information was introduced, markets could not be completely efficient. If someone tried to use information he or she possessed (say by trading on the stock exchange), the very act of trading would reveal information to others, reducing its value. Differences in the information available to different agents were shown to produce results that were far removed from the perfectly competitive ideal. For example, if banks were unable properly to monitor the performance of businesses to whom they had made loans, it might be rational for them to maintain a low rate of interest and ration borrowers. There were also attempts to construct more dynamic models of competition in which firms actively competed against other firms, trying to be the first to patent a new technology.
The economics of socialism versus capitalism received a sharp stimulus from the break-up of the Soviet Empire around 1990. It is too soon to see this in proper perspective. To many economists it seemed to offer the final vindication of the claim made by Mises and Hayek that socialism could not work, although it was only one type of socialism, implemented in very peculiar circumstances, that had failed. However, it is tempting to argue that the developments described in this chapter proved of little help in designing a rational transition from socialism to capitalism. One might claim that the most important lesson the reformers needed was to be found in Adam Smith, who emphasized the importance to any capitalist system of a secure framework of law, morality and property rights. The socialist-calculation debate, along with most welfare economics, missed this point entirely.