Within six months of taking office in February 1974 Labour had funded three worker co-operatives: Meriden, Kirkby Manufacturing and the Scottish Daily News (SDN). All three ventures were the subject of considerable economic and political controversy. The Conservatives attacked them as ‘wasteful projects’ which would only encourage ‘creeping militancy’; the Daily Telegraph referred to them as a ‘diet of sub-Trotskyism’. Michael Heseltine, Conservative spokesman for Industry, attacked Labour's support for the three co-operatives, suggesting that they encouraged ‘others to follow these illegal precedents, commit national resources to wasteful projects, and create a growing sense of injustice among the overwhelming majority of hardworking, law-abiding citizens who totally fail to understand why creeping militancy should attract government support at their expense’. There was criticism too from less obviously partisan sources: the Sixth Report of the Committee on Public Expenditure, for example, concluded that worker co-operatives were non-viable and thus a waste of public money.

When public programme demands exceed resources, tradeoffs between different public policies occur; that is, one policy gains while another loses in the allocation of these resources. Tradeoffs are most easily measured in terms of budgetary expenditures. If the budget is expanding, the consequences of tradeoffs are less severe than when the budget contracts or remains stable, and budgetary decisions involve zero-sum games. However, the concept of tradeoff is still relevant in an expanding budget since decisions are made about how the additional resources are to be allocated, and while all policy areas may gain, some obviously will receive more resources than others.

The relationship between ministers and the senior civil servants with whom they are in closest contact has long fascinated students of British government. For all the attention lavished on it, however, the relationship remains obstinately elusive and unsusceptible to clear categorization and analysis. The problem is partly one of finding a suitable frame of reference. Despite their well-established limitations, the Weberian model of an instrumental bureaucracy and the closely-related ‘politics-administration dichotomy’ still loom surprisingly large in academic analyses of bureaucratic behaviour. But the attempt to specify roles appropriate to civil servants, on the one hand, or ministers, on the other, runs the risk either of proving inadequate in face of the empirical evidence or of leading to the conclusion that one of the groups – usually the civil servants – is usurping the other's role or roles. Elements of the latter can be seen in the concern evinced in recent years about the power of civil servants vis-à-vis ministers in Britain. (Discussion of minister-civil servant relationships has, indeed, been almost entirely confined to aspects of power – particularly that of the minister's ability to get his way on policy.)

Central both to formal theories of coalition formation and many, actual, collective decision-making processes is the distribution of some reward, or payoff as it is commonly called, among those participants able to control a decision. The payoff constitutes an object of value for the players, or the stakes of the game, the disposition of which becomes a primary focus of bargaining among them. It is usual to assume that participants in coalition-forming contexts are rational actors, and this is often interpreted to mean that they pursue strategies intended to maximize their individual shares of the available payoff. Theories based upon this general structure have furnished a variety of important propositions, focusing mostly upon the composition of winning coalitions.

One of the richest data sources for the study of public opinion is the Survey Research Center's panel study conducted in the late 1950s. Because the SRC interviewed its national panel of Americans three times over a four-year period, the SRC panel data allows the analysis of changes in survey responses over time. The most remarkable discovery from the SRC panel was that panelists changed their reported opinions on policy issues with considerable frequency when asked the same policy questions in different years. Moreover, the amount of observed change in the individual responses varied little with the time interval between responses. That is, the correlations between responses to the same issue item in 1956 and 1958 or in 1958 and 1960 (two years apart) were almost as low as the correlations to the same issue item in 1956 and 1960 (four years apart).

The sharp drop in the Conservative and Labour parties' combined share of the poll in the two 1974 parliamentary elections is perhaps the most important development in British electoral politics since the rise of the class alignment of partisanship. From an average of 91·8 per cent between 1950 and 1970, the share of the vote won by these two parties dropped sharply to an average of 75·3 per cent in 1974, a figure substantially below their previous low point of 87·5 per cent in 1964. The Liberal party's share of the vote increased almost threefold over the same two periods, jumping from an average of 6·8 per cent between 1950 and 1970 to an average of 18·8 per cent in 1974.

In a recent exchange in this Journal, K.I. Macdonald took issue with Steven Lukes's claim that power is an essentially contested concept. Macdonald claims that in arguing that power is essentially contested Lukes locates disputes about power in the concept itself rather than in the divergent uses of that concept. Macdonald's claim is that there are at least two types of dispute about concepts: disputes which arise from within the concept and disputes which arise in the differing values ‘from which the concept depends’. In the case of power, Macdonald argues, the dispute lies between contending perspectives and values and not within the concept of power itself. Whilst not denying that there are essentially contested concepts, Macdonald is clear that the concept of power is not essentially contested but reflects a contest between differing value positions. Macdonald argues that whenever a concept is ‘truly essentially contested the proper ground for the contest is the essence of the concept. Our disagreements inhere in that concept’.

Following the exchange between Shackleton and Marsh on the work of Olson and its applicability to the case of the Confederation of British Industry, I wish to criticize some of Shackleton's points and develop some of those made by Marsh. I intend to ignore some of Shackleton's criticisms of the methodology Marsh and I employed in our book (for example, our use of semi-structured interviews, a respectable technique in elite interviewing) and instead to concentrate on substantive questions of more general interest.