The US Congress has often sought to limit presidential influence over certain public policies by designing agencies that are insulated from presidential control. Whether or not insulated agencies persist over time has important consequences for presidential management. If those agencies that persist over time are also those that are the most immune from presidential direction, this has potentially fatal consequences for the president's ability to manage the executive branch. Modern presidents will preside over a less and less manageable bureaucracy over time. This article explains why agencies insulated from presidential control are more durable than other agencies and shows that they have a significantly higher expected duration than other agencies. The conclusion is that modern American presidents preside over a bureaucracy that is increasingly insulated from their control.

Public bodies adopt procedures for the competitive bidding for funds in the belief that they improve public welfare, while critics regard such practices as a waste of resources and open to political manipulation. We test the operation of a competitive bidding regime through Tobit models of data drawn from successful and unsuccessful bids in four years of the Single Regeneration Budget programme in England. We derive hypotheses from a model of competitive bidding, the official evaluation of the programme and the pork-barrel literature. Our data and statistical models show that successive rounds did not greatly improve the quality of the bids, did not systematically reward needy communities and diverted resources to ministers' parliamentary seats in some regions.

The virtual disappearance of moderate and cross-pressured members from the US Congress is analysed in this article. There were substantial numbers of these partisan non-conformists in both parties and in both chambers until the early 1980s when the middle began to shrink. This trend continued and accelerated in the 1990s. Partisan non-conformists disappeared through replacement and conversion. When moderate and cross-pressured members left Congress, their replacements were much more likely to be mainstream partisans in the 1980s and 1990s than they had been in earlier decades. The occurrence of some type of conversion (a shift towards the party's ideological mainstream or a party switch) is also much more common in recent decades. We present evidence that the shrinking middle in Congress resulted from electoral changes.

‘Mirror representation’ or a ‘politics of presence’ presupposes relatively modest levels of diversity among those being represented. If the groups to be represented are too numerous, internally too heterogeneous or too cross-cutting, too many representatives will be required for the assembly to remain a deliberative one where ‘presence’ can have the effects its advocates desire. In those circumstances, what is being represented ought be conceptualized as the ‘sheer fact of diversity’ rather than ‘all the particularities of the diversity among us’. The appropriate response to that is legislative reticence.

This article calls into question the common claim that class identity does not matter for American political behaviour. Using panel-study data spanning thirty-two years and two generations, we investigate the effects of social-class identity on five participatory orientations towards government. As expected, working-class identifiers in both generations consistently display lower levels of involvement in politics than do middle-class identifiers. Significantly, however, these differences typically persist when the analysis controls for objective indicators of class and are always enhanced among those who retain the same class identity over time. Rather than sustaining a conclusion that class identification has little relevance for Americans, the results suggest that class may be particularly important in the present political context.

One of the chief explanations for the decline of social capital in the United States is the passing of the ‘civic generation’, those who came of age during the Depression and the Second World War. These Americans experienced greater government largesse than previous generations, yet we know little about how public programmes influenced their subsequent involvement in public life. This article draws on policy feedback theory to examine how the educational benefits of the GI Bill, through which 7.8 million Second World War veterans attended college or gained vocational training, affected recipients' political participation across three time periods, from 1950 to 1998. We find that initially, interpretive effects of programme implementation produced increased levels of participation among users generally. Later on, resource effects enhanced participation rates selectively, with the strongest effects among those who had attained the highest levels of education. Overall, the study illustrates distinct mechanisms, timing and sequencing through which public policy can shape the interests and capacities of programme recipients to engage in democratic participation.

Low levels of voter turnout in the first election to the National Assembly for Wales in May 1999 brought into question both the ability of devolution to revitalize representative democracy and the legitimacy of the Assembly itself. But drawing wider implications from turnout requires that we understand why electoral abstention was so widespread. We examine three hypotheses about voter turnout in 1999: that non-participation simply reflected a general apathy towards politics; that it was based on a specific apathy towards the new Assembly; or that low voter turnout reflected antipathy towards an unwanted political institution. We find support for the first two hypotheses, but little evidence for the third. Devolution has failed to engage the interest and support of many in Wales, but low turnout has not been prompted by fundamental antagonism to the devolved institution among the Welsh electorate.

Cultural identity has become prominent in studies of world politics in the post-Cold War era. First, a growing literature in world politics has emerged that focuses on the impact of social culture, broadly conceived as the shared religious, racial or ethnolinguistic characteristics of a society. The significance of this aspect of culture is epitomized in studies focusing on ‘ethnic conflicts’, ‘ethnic security dilemmas’, and most prominently in Huntington's ‘clash of civilizations’ thesis, which suggests, among other things, that shared civilization membership is the fulcrum upon which post-Cold War era world politics rests. Secondly, the impact of political culture, broadly conceived as the shared norms and institutions guiding political behaviour in a society, has become increasingly salient, as evidenced by the bourgeoning literature on the democratic peace thesis, which posits that the extent to which a state (or pair of states) is democratic is a major determinant of its war-proneness. Adherents of this view argue either that although democracies are just as war-prone as non-democracies they rarely fight other democracies, or that democracies are more peaceful than non-democracies, in general.

The objective of this study is to assess and compare the relative impact of issues and the economy on the vote in democratic elections. There is a rich and vast literature dealing with issue voting and an equally impressive literature concerning economic voting. For the most part, however, these amount to two separate streams of research. Relatively little attention has been paid to where these literatures overlap and less still to the simple but basic question: which matters most, the issues or the economy?

The main debate in the issue voting literature recently has been between the directional and proximity models. That debate, engaging both technical and conceptual issues, has focused entirely on how issues play in an election, whether voters prefer the party that is closest to their own position or the party that is the strongest defender of their side on an issue. The question of how much issues affect the vote, however, has been neglected. Indeed, both the proximity and directional schools implicitly agree that issues matter, and so challenge the Michigan school's strong scepticism on the import of issues. Given that the difference between the two models is often quite small, a more fruitful line of investigation might be to return to the equally fundamental ‘how much’ question.