This article seeks to throw some light on what is perhaps a deceptively simple question: is the central administrative bureaucracy which governs Scotland from St Andrew's House in Edinburgh in any measurable way distinctively Scottish, as opposed to the mere manifestation in Scotland of the standard U.K. civil service? Do the Scottish Departments display characteristics not found in Whitehall, and is there a ‘Scottish administrative style’ for such Departments? Answers to these questions are not only interesting in their own right but clearly have some relevance to the devolution debate. For example, if Scottish administration is recognizably ‘different’, this might be regarded as evidence that a greater degree of effective devolution than is assumed already exists; if Scottish administration shows no distinctive features at present, one might speculate what changes devolution to an Assembly would engender. We shall make some brief comments on this point at the close of this article.

Two models of local policy formation can be set forth. Although they are complementary rather than competing models, for heuristic purposes it is useful to present them as contrasting approaches. Since the first, the bargaining model, is well known, I shall concentrate in this paper on elaborating and applying a second, unitary model of policy making. The paper is divided into four main parts. After briefly identifying the limitations of the bargaining model, the first part develops the theoretical rationale for the unitary model. The second part uses the model to analyse empirically differences in the revenue sources for national, state and local governments. Part 3 does the same for expenditure policies, and, in the course of the analysis, distinguishes among three types of public policy – developmental, allocational and redistributive. In the final part, hypotheses deduced from the model are tested by means of a regression analysis of state and local expenditures.

Until the establishment of regional governments in the early 1970S, Italy was a formally centralized, unitary state. Both the legal and formal administrative frameworks within which local officials worked were such as to support the prevailing scholarly view that, in such a centralized state, the possibilities for local autonomy and initiative were very limited. The formal role of local governments after the Second World War continued to be that which had been specified by the Fascists in legislation designed to complete the centralization of the Italian state. Such a role left little room for local independence and initiative. Administratively, most decisions had to be approved by officials in Rome. Moreover, the policies and finances of local governments were closely supervised by the Prefects, who served as agents of central government.

It is not uncommon for the case-study approach to the study of policy making to be criticized for failing to produce broad generalizations or concepts applicable to systems as a whole. For example, in discussing policy making and political culture in Sweden, Anton argues that case-studies of decision making do not offer an adequate or realistic view of the process. In doing so he suggests that we should shift our focus of concern ‘from the single decision (whatever it is), to the structure of relationships between participants and the norms which serve to maintain or change those relationships through time. The focus shifts, in other words, from decisions to systems of decision-making’. Whilst not disputing the need for studying the general properties of decision making in a given political system (or indeed the need for the comparative study of policy making in different systems), this article will suggest the value of studying individual policy areas – in this case, transport – as opposed to both individual decisions and entire political systems, as a means of testing broader propositions such as those formulated by Anton.

For decades political scientists have been fascinated with the problem of establishing a formal mathematical relationship between the popular votes cast for a political party and the number of parliamentary seats it holds. The most famous attempt at solving this problem, known as the ‘cube law’, was formulated more than sixty years ago by James Parker Smith. Since then a large number of papers in support of the cube law have been published. Perhaps the most cited paper is by Kendall and Stuart. The cube law also has been stated to apply to parliamentary elections in New Zealand, the United States and Canada. But doubts about the cube law have also been expressed by Brookes, Eldersveld and just a few years ago by Tufte.

In an innovative and carefully argued article (‘The Used Vote and Electoral Outcomes: The Irish General Election of 1973’, this Journal, V (1975), 363–83) on the Irish Single Transferable Vote (STV) system in the 1969 and 1973 Dail elections, A. S. Cohan, R. D. McKinlay and Anthony Mughan state a proposition that requires further analysis. They argue that ‘the optimal number of candidates for a party is equal to the largest number of seats that the party might hope to win’ in a constituency, instead of the more common practice of nominating a larger number of candidates. The reasoning behind this proposition is that the nomination of fewer candidates means a greater concentration of the party's vote on these candidates and hence fewer vote transfers and less potential wastage of votes.

In dealing with the question of the overnomination of candidates in Irish constituencies, Lijphart and Irwin have found that overnomination does tend to injure a party's chances of electing more candidates given the vagaries of proportional representation with the single transferable vote. They conclude that ‘this favourable ten-to-one ratio appears to entail strong support for the proposition that overnomination can be dangerous and should be avoided’. My disagreement with them is not in their methods and findings, but in their interpretation of the significance of those findings.

The National Front is England's fourth largest party in terms of electoral support. The number of candidates it has fielded at local and national elections has increased at each successive election since it was founded in 1968. It has not, however, been extensively studied by academic researchers. Walker has provided a historical account of the development of the National Front. Scott has examined the activities and attitudes of one local branch. Nugent and King have compared it with earlier right-wing political organizations, particularly the British Union of Fascists. But rather less is known about the National Front's electoral support.

A small, but apparently expanding, area of research in the United States has developed around the issue of whether divisive primary elections harm the general election prospects of the winners of such primaries. The subject matter is important, being of interest to students of intra-party democracy, party organization and voting behaviour. Unfortunately the research reported in major American political science journals has been directed, with one exception, towards a trivial question, at the expense of the more complex and interesting ones. The purpose of this Note is to explain why this question is not even a poor substitute for them, and to suggest the problems towards which research in this area should be directed.

There is an error in Table 1 of my recent article, ‘Class and the Vote Before the First World War’ (this Journal, October 1978, p. 445). The ‘N’ column for middle-class Anglicans should indicate a total of twenty-four cases, with eighteen Conservative identifiers and just six respondents in the ‘Other’ partisan category. These adjusted figures correspond to the percentages which are printed correctly in the adjacent column.