Fikret Adaman and Pat Devine
The political revolutions of 1989–91 in Central and Eastern Europe and the former Soviet Union have created a new historical conjuncture in which the very future of the socialist project has been called into question.  Socialists of all varieties have been affected by a profound loss of confidence and lowering of sights, including those who had long rejected even the term ‘actually existing socialism’ as an acceptable description of the Soviet and East European experience. The underlying reason for this loss of confidence is that the historical conjuncture coincides with a deep-seated crisis of socialist theory, above all of the theory of a socialist economy.
The classical Marxist model of a socialist economy may be characterized in terms of a dichotomy between capitalism and socialism in the two dimensions of ownership and coordinating mechanism. Under socialism, capitalist private ownership of the means of production is replaced by social ownership; the operation of market forces is replaced by socialist economic planning. Social ownership enables exploitation to be abolished; socialist economic planning enables the anarchy of production to be replaced by conscious social control of the economy. The two dimensions are connected in that market forces operate through the interaction of numerous separate decisions over the use of discrete parcels of society’s productive resources, which are therefore de facto privately owned whatever the de jure position, whereas economic planning involves a single set of coordinated decisions over the use of society’s productive resources as a whole, which necessarily precludes private ownership and fragmented decision making. 
This classical Marxist model, interpreted as the centralized planning of all production decisions, has been discredited by the Soviet experience. At the same time, the revival in the 1980s of the socialist economic calculation debate that had first taken place in the 1920s and 1930s provided theoretical support for the proposition that central planning, indeed any society-wide economic planning, is necessarily inefficient and strictly speaking impossible. Any socialist economic model today must take account of the Soviet experience and, even more importantly, address the theoretical issues that have been identified in the revived calculation debate.  These theoretical issues in relation to the efficiency and/or possibility of socialism are part of the more general theoretical climate, associated with the rise of the neo-liberal Right, which challenges the desirability of any form of collective action and celebrates the superiority of ‘spontaneous order’ over conscious social action. 
The Socialist Economic Calculation Debate
The debate on the possibility of rational economic calculation in a socialist economic system, defined as a system in which the means of production are publicly owned, was effectively initiated by Barone in 1908, gathered momentum in the German literature during the 1920s, continued in the English literature during the 1930s, and then died out in the 1940s, after the summing-up article by Bergson in 1948.  The two principal sides in the debate were economists from the Austrian school, who denied the possibility of rational calculation under socialism, and socialist economists arguing within the neoclassical paradigm, who defended such a possibility. Underlying the debate were not only different views on the desirability of capitalism and socialism, but also different epistemological and methodological approaches. These differences of approach were, however, largely implicit at the time and it has been argued that the distinctive Austrian position only crystallized as a result of the debate itself and subsequent reflection on it.  In the neoclassical paradigm, knowledge is conceptualized as being objective, with production possibilities, costs and revenues constituting knowable objective constraints given which economic agents make maximizing decisions. In the Austrian school, by contrast, knowledge is conceptualized as being subjective. Instead of being taken as given for the purpose of decision making, subjective—or ‘tacit’—knowledge must be discovered through action, through competitive entrepreneurial activity in the market. The emphasis on the centrality of the ‘discovery’ process is the hallmark of the modern Austrian school and the basis of its reworking of the debate on the possibility of rational calculation under socialism which gathered pace during the 1980s. It is the foundation of the modern Austrians’ claim that the impossibility of a rational socialist economic system has been established not just historically but, more significantly, also theoretically.
In the debate, the role of the market in making rational economic calculation possible has been discussed under three headings—calculation, motivation and discovery. In 1920 Mises argued that economic calculation is only possible in a free-market system, based on private property, which establishes the exchange value of all goods and services and thus provides economic agents with the necessary information, in the form of prices, for them to decide how to act.  It then emerged that Barone had already refuted Mises’s argument by demonstrating that a socialist economy could achieve the same level of efficiency as its capitalist counterpart by solving the set of simultaneous equations, based on production and utility functions, that describes the interdependent behaviour of the economic actors in the system.  Hayek’s 1935 contribution in response to this has historically been interpreted as arguing that, although Barone’s analytic solution is conceivable in theory, it is impossible in practice, given both the amount of information that would have to be collected centrally and the scale of the computation required to solve the simultaneous equations.  Lange responded to Hayek’s argument by producing his seminal model of market socialism, drawing on earlier decentralized models. 
In this model the central-planning bureau announces an initial set of prices for producer goods; the managers of the state-owned firms take these prices as given (parametric) and follow the rule that they should seek to minimize costs and set price equal to marginal cost; firms hire labour supplied by workers in pursuit of utility maximization, sell goods to consumers who purchase them according to their utility-maximizing demands, and buy and sell producer goods from and to each other. On the basis of information supplied by the firms on increases or decreases in inventories of producer goods, reflecting excess supply or excess demand, the planning bureau announces a new set of producer goods’ prices and the process continues through a series of iterations until the supply and demand for all goods is equalized.  Lange’s model incorporates real markets for labour and consumer goods and ‘pseudo-markets’ for producer goods. Since the state owns the means of production, there is no unearned income and ‘profits’ are distributed by the state in accordance with democratically determined criteria. The model combines the allocative efficiency of idealized, neoclassical, perfectly competitive capitalism with a socialist income distribution.
Although most of the discussion in the 1920s and 1930s was of the calculation problem, there was also some discussion of motivation. Hayek responded to Lange’s model by stressing the relationship between ownership, incentives and economic efficiency.  However, his arguments did not make the impact he might have expected and by the early 1940s the ‘standard account’ of the debate was that the neoclassical socialist school had succeeded in refuting the Austrian challenge and establishing that a rational socialist economy was indeed possible.
The Principal–Agent Problem
This is how things remained until the 1980s, when the modern Austrian interpretation of the debate began to challenge the standard account. In the meantime, however, work within the neoclassical paradigm on the so-called ‘principal–agent’ problem was taking place, which, although not explicitly related to models of socialism, has formed the basis of the treatment of the motivation/incentive problem in modern work on market socialism. An ‘agent’ is a person representing the interests of her/his ‘principal’ in return for a reward, where agents might be managers and principals might be shareholders or the state, although strictly speaking the state is itself an agent for its principal, ‘society’. In a context of asymmetric information, with principals not in possession of all the information available to agents, a monitoring problem arises. Principals are unable to tell whether their agents are performing on their behalf as efficiently as is possible; agents may take advantage of this uncertainty and pursue their own interests, rather than those of their principals.
The principal–agent problem faced by any neoclassical decentralized socialist model is readily apparent. Since the centre cannot monitor perfectly what the socialist enterprise does, incentive problems may be expected to arise. The problem for the centre, then, is to find an optimal contractual scheme between itself and the enterprise, or to create a parametric environment which the enterprise takes as given, which would induce the enterprise to behave in such a way that, while maximizing its own interests, it also contributes to the achievement of maximum social efficiency. It should be noted that this principal–agent relationship is symmetrical for both socialist and corporate-capitalist economies. The problem the socialist principal—the centre—has to deal with is structurally the same as that faced by the capitalist principal—the shareholder— namely, to induce their agents to act properly in their interests. There is now a significant literature on the design of ‘incentive-compatible’ mechanisms which achieve this objective to a greater or lesser extent. 
However, although the problem is structurally the same, there still exists the question as to whether the centre would have the same incentive as shareholders to design and implement incentive-compatible mechanisms. As noted above, the real socialist principals are strictly speaking the people constituting the society in question. It has been argued that individual members of society lack the knowledge that would enable them to monitor the centre. Furthermore, they have no incentive to acquire such knowledge since the effort involved would be out of all proportion to the difference that any individual’s action could make and, therefore, to the return that they would receive. Hence people would ‘free ride’, with everyone relying on everyone else, and consequently there would be no effective monitoring.
Once again, however, economists within the neoclassical school have been investigating precisely this problem in the context of corporate capitalism. Although in a market-socialist context, it might be true that individual voters would tend to free ride, rather than actively monitor the centre, the results of capital-market theory suggest that shareholders in a corporate-capitalist enterprise do the same. Shareholders owning insignificant fractions of the total shares of a firm have little incentive to monitor its performance. As Stiglitz puts it, ‘there is always some cost associated both with obtaining information to determine whether a manager is a good manager and with evaluating alternative management teams, in other words, to voting intelligently, and there is a negligible benefit’. 
Thus, citizens and shareholders will both lack the incentive to monitor their agents. In a capitalist economy, one possible solution to this problem is that intermediate financial institutions play a positive role by monitoring managers on behalf of shareholders. Financial institutions may monitor managerial performance and intervene, either at shareholders’ meetings or in a take-over bid, though the empirical evidence suggests that the market for corporate control is far from efficient.  However, to the extent that monitoring by financial institutions is possible in a capitalist economy, it is equally possible to create comparable intermediate institutions in a market-socialist economy. 
So within the neoclassical paradigm as applied to a corporate-capitalist economy, there is no convincing theoretical support for the argument that private property has intrinsic efficiency claims in dealing with the principal– agent problem that are absent in the case of public property. In relation to both calculation and motivation, neoclassical market socialism is theoretically at least as efficient as neoclassical corporate capitalism.
However, the original challenge to the possibility of rational socialism came from Mises and Hayek, two prominent members of the Austrian, not the neoclassical, school. Although in the historical debate they focused on the problems of calculation and motivation, the defining characteristic of the Austrian school is its subjective concept of knowledge and the consequent centrality of the process of discovery through entrepreneurial activity. While it has been claimed by Lavoie that the discovery aspect of the market process was implicit in the 1920s and 1930s contributions of Mises and Hayek,17 the full articulation and expression of the process of discovery, and the challenge it presents to the possibility of socialism, have been developed by the modern Austrian school’s revival of the debate in the 1980s. According to the Austrian school as it has crystallized today, the economic problem is not, as the neoclassical school maintains, simply a question of how to allocate limited resources with alternative uses to limitless wants; rather, the problem is how subjective, or tacit, knowledge, necessarily fragmented and dispersed, can be socially mobilized. Hayek expresses the Austrian argument on the impossibility of an objective analysis of given data as follows:
"The peculiar character of the problem of a rational economic order is determined precisely by the fact that the knowledge of the circumstances of which we must make use never exists in concentrated or integrated form but solely as the dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess." 
The Austrian answer is that the mobilization and coordination of this incomplete and contradictory knowledge occur through the actions of entrepreneurs, competing against one another in the market process, discovering and learning what is and is not possible.  Thus in the Austrian setting, the means–ends nexus of economic life is not predetermined but is subject to the creative action of market participants. Hence, the economic problem is envisaged, not as how to optimize the use of available resources, but as how to optimize the use of knowledge.
Although the tacit nature of knowledge and the discovery-and-learning process are central to all members of the Austrian school, their views do encompass different emphases. As McNulty has pointed out, whereas the Misesian–Kirznerian entrepreneur detects and exploits opportunities within the existing range of economic activities which others have failed to notice, and thus plays an equilibrating role, the Schumpeterian entrepreneur plays a disequilibriating role by interrupting the existing set of activities and creating new ones through innovation. In the market process both types of entrepreneur complement each other—the one creating change, the other responding to it. 
The calculation (or computation) and motivation (or incentive) functions of the market mechanism are common to both the neoclassical and the Austrian schools. The discovery-and-learning function, however, is distinctive to the Austrian school. The cognitive discovery process of markets is seen, like the process of human discourse in language, as an intrinsically social process:
"Like verbal conversation, the dialogue of the market depends on the specific give-and-take of interaction, a creative process of interplay in which the knowledge that emerges exceeds that of any participants...Competition is not seen as a psychological attitude but as a creative learning process among minds...There is thus a bidirectional communicative process that produces a kind of social intelligence that depends on, but goes beyond, the individual intelligences of the system’s participants." 
It is this discovery-and-learning function of the market process, the crux of the Austrian school, which distances its position sharply from the neoclassical perspective. It underlies the Austrian school’s assertion that neoclassical market socialism, like neoclassical central planning, is either impossible or irrational, since the efficient discovery and mobilization of necessarily dispersed subjective knowledge, it is claimed, require a market process based on private ownership of the means of production. Lavoie argues that entrepreneurs, motivated by private incentives, direct their efforts towards the achievement of potential profits and, in doing so, make use of their dispersed knowledge to discover new opportunities: ‘The role of profits is not primarily to motivate people to do the right thing but to find out, through the process of interplay itself, what the right thing to do might be.’  Private property not only induces effort but also, and more importantly, is a necessary condition for discovery.
Now, while the Austrian epistemological objection to the neoclassical paradigm, and hence ipso facto to neoclassical socialism, may be convincing, this does not yet establish the stronger claim that all forms of socialism per se are inefficient. For that, Kornai’s argument for an ‘affinity’ between private property and the market process must be accepted.  The argument, that discovery and learning can only occur through the operation of the market process and that this process requires private property would, if true, be a convincing refutation of the possibility of any form of market socialism. However, even accepting this argument, the Austrian assertion of the superiority of capitalism would only be relevant to the case where private entrepreneurs use their own capital or borrow capital for their own use. The case of modern corporate capitalism, where the question of the relationship between capital owners as shareholders and entrepreneurial activity arises, has been largely unexplored by the Austrians.  This has left space for the construction of models of what may be called ‘Austrian market socialism’, which incorporate entrepreneurship and a capital market and thus seek to address the problem of discovery.
The Relevance of the Debate for Socialism
Any model of a socialist economy must deal with the problems of calculation, motivation and discovery. The Soviet model of centralized administrative- command planning eventually failed under each heading. Within the neoclassical paradigm, it has been established that the problems of calculation and motivation can be dealt with at least as efficiently on the basis of state-owned as of privately-owned enterprises. Within the Austrian paradigm, as we have seen, the focus is on motivation and, above all, discovery. In our view, the Austrian critique of the neoclassical paradigm, and with it of neoclassical market socialism—and a fortiori neoclassical central planning—is overwhelming. The market process in neoclassical capitalism is essentially a metaphor to enable the analysis of equilibrium states and their static welfare properties. The neoclassical paradigm contributes nothing to an understanding of how the capitalist-market process in particular and market processes in general operate.
The Austrian school, by contrast, offers real insights into the operation of actual capitalist-market forces in conditions of imperfect knowledge, uncertainty, continuous change, endemic disequilibrium and an endless process of discovery. Any model of a socialist economy must be able to address the Austrian challenge, reinforced by the weight of historical experience, that socialism—models or actual societies purporting to be socialist—cannot orchestrate the efficient social mobilization of dispersed, subjective, personally held and experienced, incomplete and contradictory knowledge. However, although the Austrian insight is compellingly insistent, it is itself incomplete. The Austrian school’s ideological commitment to private property and market forces leads it to underestimate the force of the third, albeit minor, contribution to the historical socialist economic calculation debate, that of Dobb.
Dobb insisted that economic planning, in the sense of the ex ante coordination of major interdependent investments, is a defining characteristic of a socialist economy and is incompatible with private ownership.  He criticized the decentralized models of the neoclassical market socialists on the grounds that they sought merely to mimic neoclassical capitalism, not to develop a qualitatively new system. However, his contribution was largely ignored by the other participants in the debate.
Dobb’s argument was that the neoclassical socialist economists, having defined socialism solely in terms of the legal-ownership status of the means of production, were preoccupied with the issue of allocating a given quantity of scarce resources between various uses. They focused on the marginal conditions for stationary equilibrium and gave little or no attention to problems of coordination and long-term growth paths. Dobb refused to accept that the framework for the debate should be that established by the problem of scarcity and the need to arrive at a relative valuation of alternatives in order to optimize the use of scarce resources. He criticized the decentralized Lange-type models for their reliance on the market mechanism, arguing that ‘those who dream of marrying collectivism to economic anarchy must, at any rate, not pretend that the progeny of this strange match will inherit only the virtues of ill-mated parents’.  For Dobb, the ability to plan economic activity was the main difference between a capitalist market economy and a socialist system. As he put it, ‘the essential contrast is between an economy where the multifarious decisions which rule production are taken each in ignorance of all the rest and an economy where such decisions are coordinated and unified’. 
Instability and Waste in the Market Economy
Dobb’s claim was that economic planning would enable the two major drawbacks of the market mechanism—instability and externalities—to be overcome. He maintained that disequilibrium in a market economy is only corrected after the event, with the consequence that economic fluctuations are endemic and resources are wasted, whereas economic planning would enable decisions, especially those concerned with major interdependent investments, to be coordinated in advance of any commitment of resources. As Dobb saw it, this ex ante coordination would bring about two sets of advantages. First, in an unplanned economy, market equilibrium ‘is only reached through the mechanism of fluctuations, which are themselves conditioned by the uncertainties inherent in production for a market when each autonomous decision is necessarily “blind” in part with respect to related decisions’;  planning enables the uncertainties arising from the atomized nature of individual decision-making units to be overcome. Second, planning enables externalities to be taken into account explicitly, and interrelated decisions, such as those involving infrastructure, to be coordinated before they are implemented. Finally, Dobb argued that, with planning, things that appear as ‘data’ in the static problem can be converted into ‘variables’ in a dynamic framework. Thus, it would become possible for conscious social decisions to be made, not only about the overall rate of investment, but also about the distribution of investment between capital-goods and consumer-goods industries, the choice of techniques, the regional distribution of investment, and so on.
We have seen that, during the historical debate, both the Austrians and Dobb criticized the neoclassical market-socialist school, though from different perspectives. From Dobb’s point of view, the atomistic organizational form of a market economy necessarily generates imperfections of knowledge. However, he argued that these imperfections could be remedied by a planning agency and did not address the modern Austrian insight in relation to the discovery-and-learning process through which tacit knowledge is socially mobilized. According to the Austrian school, by contrast, this process of discovery-and-learning necessarily involves entrepreneurs acting in the market. Austrians accept that the atomistic organizational form of the market mechanism generates imperfections of knowledge, but they claim, in Kirzner’s words, that ‘to describe the competitive process as wasteful because it corrects mistakes only after they occur seems similar to ascribing the ailment to the medicine which heals it, or even blaming the diagnostic procedure for the disease it identifies.’ 
The relevance of the socialist economic-calculation debate for socialism today arises from the theoretical issues it raises. One way of characterizing these is in terms of the two types of imperfection of knowledge that have to be taken into account when thinking about economic systems: those generated by the tacit nature of dispersed knowledge, and those generated by the atomistic organizational form of the market mechanism. On this basis, three alternative approaches vis-à-vis these two kinds of imperfection can be identified in the debate. The neoclassical school assumes away both sources of imperfection.  The Austrian school admits the existence of both types but argues that the discovery and mobilization of tacit knowledge through the market process require autonomous action on the part of economic agents, which precludes any attempt at ex ante coordination in order to deal with the second type of imperfection. Finally, Dobb advocates the ex ante coordination of economic activity to deal with the imperfections of knowledge intrinsic to the market process, but fails to recognize the tacit nature of information.
The analysis so far suggests that models of a socialist economy should be evaluated in terms of (i) their ability to deal with the issues of calculation, motivation and discovery, and (ii) the way in which they deal with the imperfections of knowledge arising from the tacit nature of knowledge and the atomized decision making of the market mechanism. If we start from the assumption that socialism requires some form of non-capitalist ownership, we can identify three contemporary contenders: models of market socialism involving real, as opposed to pseudo-markets; models based on direct calculation, with no real markets; and a model of participatory planning, in which market forces are replaced by a combination of real market exchange and the ex ante coordination of major investment through a process of participatory decision making.
Contemporary Contenders: Market Socialism
Contemporary models of market socialism are based on state-owned enterprises, or enterprises that are owned by various non-state but non-private institutions, or worker cooperatives. These enterprises operate in real product markets and sometimes also real capital markets. They seek to combine the efficiency allegedly inherent in markets with the income equality associated with socialist values. Our evaluation of these models will focus on efficiency in the use of existing capital stock, efficiency in the allocation of investment, and the extent to which the models incorporate wider social considerations. Under each of these headings we shall distinguish between models that are essentially within the neoclassical paradigm with respect to their epistemological and methodological assumptions and those that fall within the Austrian paradigm. As a rule of thumb, models that implicitly or explicitly recognize the need for discovery through entrepreneurial activity may be characterized as Austrian market socialism and those that do not as neoclassical market socialism. 
Efficiency in the Use of Existing Capital Stock
The issues that arise here concern calculation, motivation and the associated principal–agent problem. Nove’s model of ‘feasible socialism’, perhaps the most celebrated model of market socialism after Lange’s, envisages a slowly changing economy in which enterprises meet market demand from their existing capacity and undertake incremental investment, financed by retained profits or bank credits, to adapt to slowly changing demand and technology.  Competitive market prices provide the basis for rational calculation and competitive markets provide the motivation for efficiency. The model falls essentially within the neoclassical paradigm, since its focus is on marginal adjustment with no role for entrepreneurship or discovery. Major, non-marginal change and investment, together with the regulation of enterprise behaviour, is assumed to be undertaken by the state, but there is no discussion of how this is to be done or of where the knowledge on the basis of which these decisions are to be made comes from. Thus, the principal–agent problem is not discussed and neither is the Austrian theoretical challenge. In addition, the lessons of historical experience in relation to actual attempts at versions of market socialism are, for the most part, ignored. 
Most of the work on market socialism since Nove, referred to by Bardhan and Roemer as ‘fifth-generation models’,  has been concerned with analyzing the efficiency properties of an economy consisting of nonprivately- owned, profit-maximizing enterprises within a neoclassical framework. A distinctive feature of this generation of models is an acceptance of the principal–agent problem as a, perhaps the, central problem that market socialists have to address, and the elaboration of various ways of allocating property rights such that self-interested and selfish people will act in a way that results in an efficient use of resources.  It should be noted that, within the neoclassical paradigm, the concept of efficiency used is that of ‘Pareto efficiency’, a technical definition referring to a situation in which, on certain highly restrictive assumptions, one person’s utility cannot be increased without another’s being diminished.
The most crystallized model of neoclassical market socialism today is that of Bardhan and Roemer, in which the problem of how the principal can monitor the agent is explicitly addressed.  They propose two mechanisms, which both consciously mimic the main institutional arrangements for disciplining managers to be found in capitalism, ‘bank-centric’ monitoring and a pseudo-capital market. The bank-centric mechanism, which has much in common with the Japanese keiretsu organization, consists of a corporate group with a main public bank as its nucleus. Member firms are joint stock companies, a minority of whose shares are owned by the firm’s workers, with the rest belonging to the group’s nucleus bank, other firms in the group and third parties outside the group, such as other firms, pension funds or local government. The nucleus bank is owned by the state, with a majority interest, and other financial institutions. Thus, firms would not be owned directly by the state, but indirectly through the nucleus bank and the other firms within the group.
Monitoring would be undertaken primarily by the nucleus bank, but also by the other members of the group. In general, the group would be technologically interrelated, so that, among other things, the nucleus bank ‘can specialize in some relatively narrow and well-defined technological area, for the purpose of monitoring and scrutinizing its loans and equity involvements in the associated companies’.  Monitoring would be effective, since both the nucleus bank and the other firms in the group would ‘have a larger stake in, and more “inside” information about, a company than the ordinary shareholders in a stock market-centric system [and would] be capable of detecting and acting on early signs of trouble more easily than a diffuse body of stockholders [and would] be more prone to take a longer view in the matter of risk-taking and innovations’.  They would thus be able to monitor effectively, but would they also have an incentive to do so? The argument is that they would, because all the firms in the group would be interested in the return on their shareholdings in one another, and the nucleus bank would ‘desire to retain its reputation of credibility as a delegated monitor in a system of reciprocal delegated monitoring with a small number of other main banks’ and would ‘not want to lose the intangible asset it has accumulated specific to its relationship with the affiliated firm’. 
The second monitoring mechanism proposed, that of a pseudo-capital market, is the so-called ‘clamshell’ system. All citizens receive an initial equal endowment of vouchers, which can only be used to buy shares in firms, denominated not in currency but in coupons—or clamshells. Shares cannot be bought or sold for currency, but can be traded at prices denominated in coupons. Share ownership carries with it an entitlement to a share of the firm’s profits. The coupon prices of shares will therefore oscillate, as share prices do on capitalist stock exchanges. The shares of firms whose management is performing badly, with consequential low profits, will be sold and their coupon price will fall, and vice versa. Thus, this pseudo-capital market would ‘provide the same signals that a capitalist stock market does’. 
Monitoring on the basis of these signals would be undertaken by the banking system. Firms are financed in this system by loans from publicly owned banks. The clamshell stock market is not a source of equity finance but only a market for property rights in the profits of firms. If the coupon price of a firm’s shares started to fall, the banks would have an incentive to investigate the performance of the firm’s management in order to ensure that the firm remained sufficiently profitable to repay its loans to the banks. Clearly the clamshell system could be combined with the inside monitoring of the bank-centric keiretsu system, with the signals of the coupon prices of the clamshell shares influencing the intensity of the inside monitoring by the nucleus bank. Thus ‘one could design the system to permit varying degrees of influence that the stock market could have on capital allocation and firm monitoring, depending on one’s view concerning the efficiency of the stock market.’ 
Bardhan and Roemer recognize that in both systems the question of the relationship between the state and the banks remains an issue. They regard the central question here as being whether the banks would have enough independence from the state to be able to act on the basis of economic rather than political criteria. They argue, however, that their model would largely succeed in removing firms from the state’s orbit without granting individual citizens unbridled private property rights in them. The shareholding of the nucleus bank and the interlocking shareholdings of the member firms of a corporate group would provide a buffer against direct political influence in a firm’s activities. The shares in the nucleus bank held by other banks and financial institutions, albeit a minority interest, would make it more difficult for the state to put pressure on the bank. Even so, Bardhan and Roemer recommend a constitutional settlement that narrowly limits the circumstances in which the state could use its majority holding to intervene in the operations of the banks.
Running parallel to these models based on non-state, non-private forms of ownership are a group of models based on worker cooperatives or labour-managed firms. Early neoclassical analysis of the labour-managed firm identified several inefficiencies arising from the assumption that workers seek to maximize income per worker while capitalists maximize profits.  Contemporary models in this tradition have been concerned to show how these inefficiencies might be overcome. Sertel has demonstrated that the inefficiencies do not arise if a market in membership rights is introduced, in which case the behaviour of a labour-managed firm and its capitalist equivalent is identical.  Fleurbaey proposes a model based on workplace democracy in which indirect external financing— by household savings channelled through the banking system— enables most of the traditional inefficiencies to be overcome.  Weisskopf develops a model in which worker self-managed enterprises obtain assets by, inter alia, selling non-voting tradable shares to independent self-managed mutual funds. He then analyses the efficiency properties of the model and concludes that: "it is possible to meet various lines of criticism in such a way as to diminish considerably the force of the argument that my model of market socialism will prove inefficient. I do not mean to suggest that it will be just as efficient (by conventional efficiency criteria) as a capitalist model. Against any remaining efficiency disadvantages . . .however, must be set not only its contributions to . . . socialist goals . . . but also certain countervailing efficiency advantages." 
Weisskopf’s conclusion encapsulates the essence of the neoclassical approach to market socialism—the analysis of different allocations of property rights in terms of the extent to which self-interested opportunistic behaviour based on them gives rise to contractual agreements that are Pareto efficient. Neoclassical market socialists have been more or less successful in dealing, on their own terms, with the issues of calculation and motivation in relation to the Pareto efficient use of existing capital stock, but they have not even begun to address the issue of discovery. Nor is it at all evident, even in principle, that they could ever do so.
However, in what could perhaps be regarded as an early model of Austrian market socialism Liska seeks to deal with the issue of discovery, in the context of the use of existing capacity, by replacing the ‘socialist manager’ by the ‘socialist entrepreneur’. Liska’s model consists of leasing state-owned production and property units to private entrepreneurs through a process of competitive bidding.  The state owns the existing means of production and makes decisions on investment, but the highest bidders decide how the existing means of production are to be used, on the basis of their initial expectations and what they discover through using them. When the lease expires, the bidding process is repeated and the operating entrepreneur must either match the highest bid or leave the business to any overbidder. The emphasis is therefore on the ‘human assets’ of entrepreneurs, which are in the end manifested as a residual surplus that belongs to them. Hence, the model can be interpreted as an attempt to ensure that state-owned production and property units are used in an entrepreneurial way and their most efficient use is discovered.
Efficiency in the Allocation of Investment
However, although Liska’s model addresses the issue of discovery in the context of the use of existing capacity discovery really only arises as a fundamental issue in relation to investment decisions, in a context of change and uncertainty. This is why neoclassical models of market socialism, dominated by the analysis of the efficiency properties of equilibrium states in conditions of certainty or certainty equivalence (risk), are unable to respond to the modern Austrian challenge, concerned with the problems of motivation and discovery rather than of calculation and motivation.  Most neoclassical models of market socialism simply assume that the pattern of investment generated by perfectly competitive market forces is efficient, and explore ways in which this pattern can be achieved in an economy consisting of non-privately-owned enterprises.
Yet there are some market-socialist models that do try to deal explicitly with the issues of entrepreneurship and discovery in an uncertain dynamic context, and hence can be thought of as models of Austrian market socialism. Brus and Laski’s model of what they call ‘market socialism proper’ seeks to combine ‘state ownership in one form or another’ with ‘full independence of firms and true entrepreneurship’, through the functioning of a real capital market along with real product and labour markets.  However, the conditions that have to be met to achieve this, effectively replicating capitalism, are such as to render the retention of formal state ownership artificial and redundant. Thus, Brus and Laski conclude that ‘the pure logic of the fully-fledged market mechanism seems to indicate the non-state (private) enterprise as the more natural constituent of the enterprise sector’. 
Estrin envisages an economy in which self-managed worker cooperatives borrow capital from competing profit-maximizing holding companies whose shares are owned by the government and the self-managed firms themselves. He argues that the existing labour forces of the self-managed firms should bear the risks of production and receive any residual surpluses, while the holding companies should perform the entrepreneurial function of innovation, research and development, and market research. The holding companies would manage social capital by creating and closing firms—in the latter case even against the opposition of the workers involved. 
Schweickart proposes a model of ‘economic democracy’ based on worker-managed production units, producer- and consumer-goods markets, and socially controlled investment. Enterprises compete against each other and apply to regional community banks for finance to invest in new opportunities as they perceive them. Schweickart claims that ‘Economic Democracy provides for, indeed requires, “socialist entrepreneurs”, individuals or collectives willing to innovate, take risks, in hopes of providing new goods and services, or old ones in new ways.’ 
What these models of Austrian market socialism have in common is the attempt to mimic the institutions of capitalism as closely as possible on the basis of different forms of non-capitalist ownership. The closer they get to this, and thus incorporate a process for the discovery and mobilization of tacit knowledge, the further they move from being able to deal with the imperfections of knowledge arising from the atomized decision making inherent in the operation of market forces.
Wider Social Considerations
Market-socialist models emphasize the efficiency of market forces, albeit interpreted in different ways, and claim that their superiority in relation to capitalism lies primarily in the more equal income distribution that non-private ownership makes possible. However, some market socialists also recognize that the operation of market forces is characterized by problems of coordination, externalities and risk aversion. They have therefore sought to build into their models institutions and procedures that enable wider social considerations to be taken into account than can be captured by the unregulated operation of competitive market forces. 
A recent neoclassical model by Ortuno-Ortin, Roemer, and Silvestre explores different ways of implementing a politically determined pattern of investment through tax and interest rate incentives that differ across sectors.  They develop a general equilibrium model in which the centre implements the predetermined pattern of investment either through direct commands or through ‘indirect planning’ by shaping the parametric market signals to which the publicly owned, profit-maximizing firms respond.  However, the crucial factor to note is that in this model ‘the public authority has complete information about the technology of firms’, so that the Austrian critique of the epistemological assumptions underlying the neoclassical paradigm applies with full force. 
Austrian market-socialist attempts to take account of wider social considerations than those captured by the operation of competitive market forces have been made, inter alia, by Estrin and Winter and by Schweickart. Estrin and Winter, in the context of Estrin’s model of selfmanaged firms and profit-maximizing holding companies, advocate indicative planning in order to reduce the imperfections of knowledge associated with the ex post coordination of atomized decisions through the operation of market forces. Schweickart proposes that the overall rate of investment should be decided democratically through the political process, with regional community banks then being allocated a share of the available investment funds on the basis of agreed social priorities. Worker-managed enterprises would compete for these investment funds and the banks would allocate them primarily on the basis of expected profitability, but would also take into account other politically determined criteria. 
A related model has been proposed by Elson, seeking ‘the democratization of the market within a framework of strategic planning’.  She argues that the market process does not permit collective reflection before individual units take decisions, thus fostering the pursuit of individual objectives at the expense of long-term cooperation, and that the market should be transformed, or ‘socialized’, by new institutions of democratic regulation and control. In her model, firms are publicly owned, with representatives of consumers and the local community on their boards, but are internally self-managed; a ‘Regulator of Public Enterprises’ performs the functions of the capital market; and prices are shaped through a dialogue between firms and ‘Wage and Price Commissions’. This setup, according to Elson, would change the antagonistic nature of social relations between buyers and sellers and make the process of price formation public rather than private. Furthermore, the workability of the system would be enhanced by the creation of ‘buyer–seller networks’ to facilitate information exchange and enforce information disclosure.
Unresolved Contradictions in Market Socialism
Most modern work on the economic theory of socialism consists of the elaboration of alternative versions of market socialism. However, the coherence of the market-socialist project has been called into question, both by those who believe that capitalism is intrinsically more efficient and by those who believe that socialism necessarily involves economic planning. The aim of market socialists is to combine equality with efficiency, while taking into account the wider social considerations referred to above. Yet there are unresolved, indeed unresolvable, tensions within the market-socialist project.
In order to achieve a greater degree of equality than is thought possible under capitalism, private ownership is replaced in market-socialist models by various forms of non-private ownership. In order to avoid the inefficiencies associated with administrative-command central planning, or political intervention by the state, enterprises need to be more or less autonomous. The decisions of the non-privately-owned enterprises making up the economy are therefore coordinated, to different extents in different market-socialist models, by market forces. However, for the efficiency properties allegedly associated with market forces to be fully realized, enterprises have to be more rather than less autonomous. Thus, the logic of market socialism is inexorably in the direction of Austrian market socialism, of Brus and Laski’s ‘market socialism proper’. However, the incentive mechanism of market forces, competitive success or failure, generates inequality. This, plus the wider social considerations that market socialists tend to regard as important, creates pressures to limit the autonomy of enterprises and thus the scope for the operation of market forces. Yet limitations on enterprise autonomy recreate the very inefficiencies that the market-socialist project sets out to overcome. 
In addition to these tensions, or contradictions, within the market-socialist project, there are other considerations that cause some socialists to reject the very concept of market socialism. First, there is Dobb’s insistence that the atomized decision making by autonomous enterprises through which market forces operate necessarily creates imperfections of knowledge that result in the ‘anarchy of production’. This, it is argued, can be overcome by the planned ex ante coordination of major interdependent decisions, but not by the indicative planning advocated by both Estrin and Winter and, although not referred to as such, by Elson.  Since, in the end, enterprises engaging in indicative planning remain autonomous and act atomistically, indicative planning itself cannot overcome the anarchy of production inherent in the operation of market forces. Second, many socialists reject the market-socialist behavioural assumption that people necessarily act in a selfish, narrowly self-interested, opportunistic way, arguing instead that behaviour is shaped by social institutions and that socialist economic arrangements should encourage cooperative behaviour, rather than promote adversarial competitive social relations. Finally, the belief that market forces create alienation and a sense of helplessness, of being subject to forces beyond one’s control, rather than promote participation in a process of individual and collective self-determination, remains strong, and for good reasons.
Yet, despite this critique of market socialism, in the apparent absence of what is seen as any convincing alternative, most socialists, albeit reluctantly, have turned to market socialism as the only way of sustaining any overall socialist project. There are, however, two other contemporary approaches to the economic theory of socialism which we will now discuss: direct allocation and participatory planning.
Contemporary Contenders: Direct Calculation
Although Lange in his celebrated 1938 model used pseudo-markets for capital goods as a decentralized computer, in 1967 he published a contribution to Dobb’s Festschrift in which he argued that the modern electronic computer could be a substitute for the market.  For Lange this was his affirmation that his seminal market-socialist model had been superseded by subsequent theoretical and technical developments. The two most prominent modern models of direct calculation are those by Cockshott and Cottrill and by Albert and Hahnel.
Cockshott and Cottrill push to the limit the possibilities offered by computer technology for the classical vision of the direct allocation of society’s productive resources in accordance with a democratically determined set of objectives and priorities. They confront head on Nove’s apparently conclusive argument that the calculation involved in solving an input–output system for a complex modern economy is impossible.  According to their calculations, in an economy consisting of one million distinct products, a computer capable of performing roughly 200 million operations per second, using the standard method of solving simultaneous equations, would take 16,000 years to compute the labour values of each product. By contrast, using the approach of successive approximation it would provide a solution correct to four significant decimal digits within a few minutes—and modern computers operate even faster. 
Cockshott and Cottrill then develop a decentralized algorithm for calculating and revising the (direct and indirect) labour content of each product. They advocate a quasi-market in which the difference between the calculated labour content of each product and the labour tokens people are prepared to exchange for the product provides the central planners with information on the basis of which to adjust the production instructions that they issue to producers. However, the information that the central planners would need in order to do this includes, ‘lists of the products being produced...regular updates on the technology used in each production process . . . the available stocks of each type of raw material and every model of machine’; furthermore, ‘on the basis of a central evaluation of the different production technologies, the planning system would choose the intensity with which each technology was to be used.’  Thus, while Cockshott and Cottrill’s model effectively disposes of the argument that direct calculation is practically impossible, on which the plausibility of Nove’s advocacy of market socialism as the only ‘feasible’ socialism primarily rests, it does so on the basis of the neoclassical epistemological assumption and nowhere addresses the issues of discovery and entrepreneurship.
Albert and Hahnel argue for their model of ‘participatory economics’ by asking:
"why cannot workers in different enterprises and industries, and consumers in different neighbourhoods and regions, coordinate their joint endeavours themselves—consciously, democratically, and efficiently? . . .What is impossible about a social, iterative, planning procedure in which workers and consumers propose and revise their own activities in light of accurate information about what is efficient and what is fair?" 
Their approach is based on a decentralized, iterative planning procedure, through which workers’ councils and consumers’ councils achieve equilibrium. Workers’ councils make decisions through democratic voting procedures and are organized on the basis of ‘balanced job complexes’, in which people rotate over a reasonable time period through a sequence of tasks, in order to combine different types of work in the proportions in which they need to be performed in society as a whole. They are asked to maximize their members’ utility functions, taking into account the objective conditions of their activities and subject to the constraint that their contribution to output should exceed the social costs they incur in producing it. Similarly, consumers’ councils, in which people indicate their consumption requests for personal and public goods, are asked to draw up their consumption-bundle proposals in order to maximize their members’ utility functions, subject to a budget constraint.
The process of allocation consists of consumers’ and workers’ councils proposing and revising their own activities, prior to initiating those activities, via a price signal mechanism in which prices are adjusted in an automated way according to excess supplies and demands. Thus, the model replicates Lange’s iterative procedure—in its ex ante rather than ex post version—with two new dimensions: it claims to guarantee a participatory decision-making process by having workers’ and consumers’ councils propose and revise their own activities; and it aims to have non-hierarchical production relations which promote equity and participation through the institution of ‘balanced job complexes’.
Albert and Hahnel focus on the use of existing capacity, in ways which take account of both efficiency and wider social considerations, but they do not discuss investment.  Cockshott and Cottrill’s model formally embraces investment, but only on the neoclassical assumption of a complete knowledge of all production functions. Neither model explicitly addresses the issues of discovery and entrepreneurship. Furthermore, on a political note, both reject representative political democracy in favour of various forms of direct-polling procedures and referenda, with no provisions for face-to-face social interaction and negotiation.
Thus, models of direct calculation are able to deal with the technical problem of calculation, but not with the actual problem of discovery that confronts real economies. They operate within the neoclassical epistemological paradigm in which knowledge is assumed to be objectively given and readily codified and transmitted. While they are in principle able to deal with the imperfections of knowledge arising from atomized decision making, they are unable to deal with the imperfections arising within the Austrian epistemological paradigm, in which knowledge is tacit and has to be discovered through a process of social mobilization.
Contemporary Contenders: Participatory Planning
Models of market socialism and of direct calculation, in so far as they are internally coherent, are constructed within what is essentially the neoclassical paradigm. Austrian models of market socialism, as we have seen, are either incoherent or effectively replicate capitalist private property. Furthermore, the Austrian school rejects in principle the possibility of dealing with the imperfections of knowledge that arise from atomized decision making.
The present authors have advocated a system of participatory planning, the third contemporary contender in discussion of the economic theory of socialism, arguing that what is needed is a paradigm shift in the way in which economic interactions are conceptualized in order to overcome the two sources of imperfection of knowledge identified in the evaluation of the socialist economic calculation debate. The system of participatory planning advocated is one in which the values of individuals and collectives interact and shape one another through a process of cooperation and negotiation. Such a process, it is claimed, would enable tacit knowledge to be articulated and economic life to be consciously controlled and coordinated in a context that dispenses with coercion, whether by the state or by market forces. Thus, the dilemma to which Nove refers, when he asserts that there are only two dimensions: ‘There are horizontal links (market), there are vertical links (hierarchy). What other dimension is there?’,  can be transcended.
The premises of the paradigm shift underlying the model of participatory planning need careful attention. The Austrian claim is that only entrepreneurs making decisions about the use of private capital have both access to tacit knowledge and the incentive needed for action leading to its discovery. The participatory planning approach claims that people in general have access to tacit knowledge and that they can discover and articulate this knowledge provided that they are equipped with the capacity to analyze and evaluate the consequences of their decisions. Thus, generalized participation in decision making would enable the social mobilization of the tacit knowledge of people in general, rather than confining the process of discovery to the small subset of people who have access to private capital in a capitalist economy.
Market Exchange and Market Forces
Although participatory forms of economic organization have a long history, the modelling of a socialist economy based on participatory planning is relatively new.  One of the present authors, Devine, has developed a model which envisages the basis of economic organization as a process of ‘negotiated coordination’ among representatives of those affected by the decisions involved, informed by participatory discussion among the multiplicity of affected interests.  In Devine’s model, the distinction between ‘market exchange’ and ‘market forces’ is of crucial importance:
"Market exchange involves transactions between buyers and sellers, where what is being exchanged consists of either stocks (inventories) or goods and services produced by enterprises using their existing capacity. Market forces refer to the process whereby changes are brought about in the underlying allocation of resources, the relative size of different industries, the geographical distribution of economic activity, through the interaction of decisions on investment and disinvestment that are taken independently of one another, with coordination occurring ex post." 
The process of negotiated coordination is one in which information generated by market exchange on the profitability of different activities is not rejected but is used in a cooperative and coordinated way, in conjunction with other quantitative and qualitative information, to decide on the pattern of investment. Thus, market exchange coordinates the use of existing productive capacity, but changes in the structure of productive capacity are negotiated and coordinated ex ante by those affected by them. Hence:
"The model combines planning with decentralization, without relying on market forces. The overall allocation of resources is planned at the level of society as a whole. Major investment is coordinated centrally but implemented on a decentralized basis. Production units know what to produce with their existing capacity in order to meet society’s collectively and individually determined demand. Changes in the capacity of production units are decided in negotiated coordination bodies by those affected by them. The social interest is defined in each case by the general and specific interests involved. Production units are under public scrutiny which encourages them to operate efficiently. Since people participate at all levels in taking the economic decisions that affect them, they are likely to be committed to the effective implementation of those decisions." 
Devine defines social ownership as ownership by those affected by the use of the assets in question. Thus, in his model, enterprises are owned by their workers, customers, suppliers, the communities and regions in which they are located, the more general interests represented by the regional, national or global planning commission, depending on the geographical reach of their operations, and issue-based groups, such as those concerned with the environment or equal opportunities.  Prices would be set equal to long-run average cost, based on labour costs, the cost of bought-in inputs and a centrally determined capital charge. Enterprises would compete for customers, thus engaging in market exchange and generating information about the effectiveness with which they are using their assets:
"There would be no attempt to coordinate ex ante transactions between producers and users. The degree of capacity working and the state of order books and inventories in an industry as a whole would indicate whether the aggregate capacity of the industry needed expanding or contracting. Thus, the need for changes in an enterprise’s capacity would arise as a result of its performance relative to that of other enterprises in the industry and/or as a result of an imbalance between industry supply and demand." 
Changes in capacity would be decided by negotiated coordination bodies consisting of representatives of the social owners of the industry: the enterprises in the industry; major supplying industries and customers; the relevant regional, national or global planning commission; and other groups with a legitimate interest in the set of decisions at issue. Negotiated coordination bodies
"would have available to them three sorts of quantitative information: first, accounting data on the performance of each enterprise, generated by the use of their existing capacity (i.e., by market exchange); second, estimates of expected changes in demand or costs in relation to existing activities; third, estimates of expected demand and costs in relation to potential product or process innovations. They would also have available to them two sorts of qualitative information, supplied by the representatives of the different interests participating in the negotiation process: first, judgements about the reasons underlying any differential performance by enterprises; second, the views of those affected about the economic and social situation prevailing in the communities and regions in which investment or disinvestment might occur, the priorities for regional distribution agreed through the democratic political process, and the concerns of other interests represented." 
Devine’s model of participatory planning incorporates market exchange but not market forces. Since the use of existing capacity is decided by enterprises engaging in market exchange on a decentralized basis, the danger of administrative overload is minimized.  Interdependent investment decisions are taken by representatives of those affected by them. Thus decisions with global scope, such as those concerned with international redistribution, the global pattern of economic activity, or activities with global ecological consequences, would be negotiated at the global level. In general, however, the principle of subsidiarity would be applied to ensure that decisions were negotiated at the most local level consistent with the participation of representatives of all the major affected interests. Since the process of negotiation would be continuous, those involved would discover their tacit knowledge, would find out what was possible and what was not possible, and would therefore be able to evaluate their past decisions and learn from them. Hence
"while interdependent decisions would be coordinated as far as possible ex ante, through a process of negotiation that enabled discovery and learning before resources were committed, implementation would result in further discovery and learning ex post that would enable subsequent correction in the next round of decision making. However, these integrated processes of ex ante and ex post decision making would be based on cooperative negotiation rather than coercion or competition." 
In summary, Devine’s model places the utmost importance on negotiation and interaction, emphasizing the transformative aspect of participation. It seeks to promote cooperation and recognition of interdependent common interest, while acknowledging that people have distinct interests which they need to be able to articulate and argue for in a context which also encourages recognition and respect for the interests of others.
The twentieth century’s historical and theoretical experience of attempts to construct or model socialist economic systems has been rich but prima facie discouraging. We have tried to draw some conclusions from this experience in order to inform contemporary discussion of the economic theory of socialism. Models of a socialist economy need to be able to deal with the problems of calculation, motivation and discovery; and they need to address the imperfections of knowledge associated with the tacit nature of knowledge and the atomized decision making of market forces. The three contemporary contenders are market socialism, direct calculation and participatory planning.
Market-socialist models are of two sorts: neoclassical and Austrian. Neoclassical models are primarily static and have been concerned with the issues of incentives and monitoring. They are able to deal with the problems of calculation and motivation, but not with the problems of discovery and entrepreneurship. The few models that discuss investment assume certainty or a certainty equivalent. Neoclassical models do not recognize the tacit nature of knowledge and are unable to deal with real uncertainty. Like the neoclassical paradigm as a whole, they have nothing to tell us about the ways in which capitalist market economies actually work or socialist market economies might work.
Austrian models of market socialism do seek to address the problems of discovery and entrepreneurship, as well as those of calculation and incentives. However, they are either incoherent or mimic capitalism so closely that they are based on de facto private property. Furthermore, while they may be able to deal with the problem of mobilizing tacit knowledge, they are by definition incapable of dealing with the imperfections of knowledge associated with atomized decision making. Thus, they cannot incorporate economic planning, in the sense of the ex ante coordination of major interdependent investment.
Models of direct calculation are able to address the formal problem of calculation, with modern computers having disposed of the technocratic information collecting and processing arguments. However, these models are quite unable to deal with the issues of tacit knowledge, discovery and entrepreneurship. In order to deal with investment they have to assume certain knowledge or its equivalent and thus their analytic framework is essentially the same as that of the neoclassical paradigm.
If this evaluation is correct, participatory planning offers the most promising direction for future work on the economic theory of socialism. It is able to deal with the problem of calculation through quantitative information generated by real market exchange and qualitative information supplied by those closest to, and most affected by, decisions. Since major investment is not undertaken on the basis of atomized decisions it can be coordinated ex ante. At the same time, the tacit nature of knowledge is recognized. A social process of discovery and mobilization through generalized participation and negotiated coordination lies at the heart of participatory planning. Thus, the imperfections of knowledge arising from both tacit knowledge and atomized decision making can be dealt with. Participatory planning is the only one of the three contemporary contenders able to do this.
Why, then, is market socialism so predominant in contemporary discussions of the economics of socialism? One explanation is the experience of Soviet-style central planning. Another is the fascination with the neoclassical paradigm, and the intellectual capital invested in it, that characterizes most economists, even socialist economists. However, there is a third reason, and it is ultimately perhaps the most telling. Models of direct calculation and of participatory planning are based on the traditional socialist assumption that people have the capacity to recognize that their own interests and those of others are interdependent and to be concerned not only for their own but also for the general good. Market socialists have lost faith in that assumption. Indeed, many would regard it as dangerous, interpreting it as a belief in the perfectibility of human beings, with the corollary that it might be used to justify social engineering. Thus, market socialists design their models in order to deal with the incentive and monitoring problems that arise in the context of opportunistic self-interested behaviour by ‘the selfish manager, planner, and public bank director’. 
Our advocacy of participatory planning, by contrast, is based on an insistence that socialism must have a transformatory dynamic. Marx, after all, may have been right when he expected socialism to be created on the basis of the highest level of development reached by capitalism. We can today interpret this to mean not just the level of labour productivity, but also the level of involvement in self-organization, in the institutions of civil society, in participatory democratic processes. That is why evolution towards participatory politics and economics must involve the abolition of the social division of labour, the transformation of society from one which is dominated by a de facto ruling class, whether defined on the basis of property ownership or institutional position, towards a society which is self-governed, with people in the course of their lives undertaking their fair share of the business of running things, whether at the micro level, the macro level, or both.  Not surprisingly, the economic theory of socialism is inseparably linked with the political theory of socialism.