¹ Downs, A., An Economic Theory of Democracy (New York: Harper, 1957), pp. 114–41.Google Scholar Even though Downs entitles Part II of his book ‘The General Effects of Uncertainty’, certainty about the location of actors is an essential feature of his spatial model.

² For an up to date review, see Coughlin, Peter J., Probabilistic Voting Theory (Cambridge: Cambridge University Press, 1992), pp. 1–22CrossRefGoogle Scholar: Coughlin himself analyses such situations throughout his book. See also Calvert, Randall, Models of Imperfect Information in Politics (Boston, Mass.: Harwood Academic Publishers, 1986).Google Scholar

³ A notable exception here is provided by Enelow, James M. and Hinich, Melvin J., The Spatial Theory of Voting (Cambridge: Cambridge University Press, 1984), pp. 36–64Google Scholar and passim, who assume that candidates cannot change their ideologically-based policy positions in the course of a single election campaign and indeed can change them only with difficulty over a long time period. Policy movement is confined to changing electors' perceptions of how these general positions translate into specific policies.

⁴ This is the well-known ‘Condorcet’ or ‘Arrow’ problem: given a distribution of preferences over three alternatives of the following type, each supported by a third of electors – A > B > C, B > C > A, C > A > B – no stable majority can be found as A is preferred to B, B to C, but C to A, all by twothirds of electors. The only way to decide is not by democratic voting processes but by arbitrary imposition – a severe critique of democracy if the problem recurs often, as it seems it may.

The problem does not occur if all policies fit along one dimension (say a left-right one) with single-peaked preferences, since then the median position is preferred by a stable majority. But it is very likely to recur in a space of two dimensions and almost certain to occur in higher-dimensional spaces.

On the problem of voting cycles, see Arrow, Kenneth J., ‘Values and Collective Decision Making’, in Laslett, P. and Runciman, W. G., eds, Philosophy, Politics and Society, Third Series (Oxford: Blackwell, 1969), pp. 215–32Google Scholar: McKelvey, R. D., ‘General Conditions for Global Intransitives in Formal Voting Models’, Econometrica, 47 (1979), 1085–111.CrossRefGoogle Scholar

⁵ Simon, Herbert, Administrative Behaviour (New York: Macmillan, 1947)Google Scholar first popularized the idea of sub-optimal but perfectly rational strategies to be undertaken under conditions of circumscribed knowledge and high calculating costs.

⁶ Schelling, Thomas, The Strategy of Conflict (Boston, Mass.: Harvard University Press, 1960), chap. 3.Google Scholar

⁷ McKelvey, R. and Ordeshook, P., ‘Elections with Limited Information: A Fulfilled Expectations Model Using Contemporaneous Poll and Endorsement Data as Information Sources’, Journal of Economic Theory, 35 (1985), 55–85CrossRefGoogle Scholar; McKelvey, R. and Ordeshook, P., ‘Sequential Elections with Limited Information’, American Journal of Political Science, 29 (1985), 480–512.CrossRefGoogle Scholar

⁸ In line with most discussion, Sargent, Lyman Tower, Contemporary Political Ideologies: A Comparative Analysis (Homewood, Ill.: Dorsey Press, 1972), pp. 1–2Google Scholar, defines political ideology as: ‘a value or belief system that is accepted as fact or truth by some group. It is composed of sets of attitudes towards the various institutions and processes of society. It provides the believer with a picture of the world both as it is and as it should be, and in so doing, it organizes the tremendous complexity of the world into something fairly simple and understandable’; ‘[Ideologies] provide a cognitive structure through which to interpret and understand events’ (Christenson, R. M. et al. , Ideologies and Modern Politics (London: Nelson, 1972), p. 8).Google Scholar

⁹ As stressed by the developmental model of party systems, which sees them as evolving from the development of party ideologies emphasizing the outlook and interests of groups differentiated by deep-rooted social cleavages in the society (class, religious, ethnic, agrarian, etc.), for which see Lipset, S. M. and Rokkan, Stein, Party Systems and Voter Alignments (New York: Free Press, 1967), 1–64.Google Scholar

¹⁰ For a development of this point see Budge, Ian and Laver, Michael, ‘Office-Seeking and Policy-Pursuit in Coalition Theory’, Legislative Studies Quarterly, 11 (1986), 485–506.CrossRefGoogle Scholar For a partial exception see Austen-Smith, D. and Banks, J., ‘Elections, Coalitions, and Legislative Outcomes’, American Political Science Review, 82 (1988), 405–22.CrossRefGoogle Scholar

¹¹ A major exception here must be made for Enelow and Hinich's The Spatial Theory of Voting, which recognized the central place of ideology in defining party electoral positions and inhibiting radical policy movement. In order to permit the possibility of some movement within their models, they allow for electors' perceptions of how these general stances translate into specific policy terms to be worked on by party propaganda. The approach adopted here (supported by empirical evidence) simplifies their argument by placing party positions and tracing their constrained movements within a single left-right dimension of contrasting policy priorities (see Section IV).

¹² See the review of contemporary theory in Laver, Michael and Budge, Ian, Party Policy and Government Coalitions (London: Macmillan, 1992), pp. 2–9.CrossRefGoogle Scholar

¹³ Downs, , An Economic Theory of Democracy, pp. 107–13.Google Scholar

¹⁴ Enelow, and Hinich, , The Spatial Theory of Voting, pp. 36–50Google Scholar; Coughlin, , Probabilistic Voting Theory, pp. 172 ff.Google Scholar, review and extend various models which allow for restricted movement.

¹⁵ There may be slight qualifications to be made for closely contiguous parties, see Table 1 below.

¹⁶ See Klingemann, H.-D., Hofferbert, R. I., Budge, Ian et al. , Parties, Policies and Democracy (Boulder, Colo.: Westview Press, 1994)Google Scholar for evidence that this happens in all countries, in regard to at least some government expenditure priorities.

¹⁷ Schelling, , The Strategy of Conflict, pp. 53–67, in particular pp. 62–3Google Scholar, where informal simulation showed two armies stopping at a river, in spite of the fact that this outcome gave one side only a third of the territory.

¹⁸ Budge, Ian, Robertson, David and Hearl, D. J., eds, Ideology, Strategy and Party Change in 19 Democracies (Cambridge: Cambridge University Press, 1987), pp. 44–8.CrossRefGoogle Scholar

¹⁹ Kenneth Shepsle in particular has seen arbitrarily-imposed institutional constraints and divisions of labour as imposing unidimensionality in different areas of policy making and hence facilitating the emergence of stable but artificial equilibria. See Shepsle, Kenneth J. and Weingast, B., ‘Structure Induced Equilibrium and Legislative Choice’, Public Choice, 37 (1981), 503–19.CrossRefGoogle Scholar See also Shepsle, K., Models of Multiparty Electoral Competition (Cambridge, Mass.: Harvard, 1991).Google ScholarEnelow, and Hinich, , The Spatial Theory of Voting, pp. 155–8Google Scholar also follow through this line of argument.

²⁰ Some parties, particularly in Scandinavia and the Low Countries, do not put out a single comprehensive statement of policy but a number of documents – for women, for youth, for the economy, etc. – covering the areas which in Britain and United States are discussed as sections inside the comprehensive manifesto platform. As they do not overlap in policy terms, these separate documents are in no way competing party programmes in the sense used in the text.

²¹ Robertson, David, A Theory of Party Competition (London: Wiley, 1976).Google Scholar

²² Robertson, , A Theory of Party Competition, chaps 4 and 6Google Scholar; Budge, Ian and Farlie, Dennis, Voting and Party Competition (London: Wiley, 1977), pp. 424–33.Google Scholar

²³ McKelvey and Ordeshook explore such a rule with unconstrained two-party movement (‘Sequential Elections with Limited Information’).

²⁴ As Enelow and Hinich point out, votes may be attracted on the basis of candidate appeal or other non-policy based characteristics of a party (A Spatial Theory of Voting, pp. 37 ff.).Google Scholar

²⁵ The Manifesto Research Group, under the direction of Ian Budge, commenced work in 1979, and has published the following reports: Budge, , Robertson, and Hearl, , eds, Ideology, Strategy and Party Change in 19 DemocraciesGoogle Scholar; Laver, and Budge, , eds, Party Policy and Government CoalitionsGoogle Scholar; Klingemann, , Hofferbert, , Budge, et al. , Parties, Policies and Democracy.Google Scholar

²⁶ Budge, Ian and Hofferbert, R. I., ‘Mandates and Policy Outputs: US Party Platforms and Government Expenditures’, American Political Science Review, 84 (1990), 111–31CrossRefGoogle Scholar, and ‘The Party Mandate and the Westminster Model: Election Programmes and Government Spending in Britain’, British Journal of Political Science, 22 (1992), 151–82Google Scholar: Hofferbert, Richard I. and Klingemann, Hans-Dieter, ‘The Policy Impact of Party Programmes and Government Declarations in the Federal Republic of Germany’, European Journal of Political Research, 18 (1990), 277–304CrossRefGoogle Scholar: Klingemann, , Hofferbert, , Budge, et al. , Parties, Policies and Democracy.Google Scholar

²⁷ Budge, , Robertson, and Hearl, , eds, Ideology, Strategy and Party Change, pp. 1–38.Google Scholar To a certain extent, the idea of parties emphasizing different priorities rather than taking up opposing positions on the same issue corresponds to Stokes' distinction between ‘position’ and ‘valence’ issues; see Campbell, A., Converse, P., Miller, W. E. and Stokes, D. E., Elections and the Public Order (New York: Wiley, 1966), pp. 170–1.Google Scholar However, the idea of valency neglects the fact that party emphases set priorities which can be quite directive of policy, especially in the field of expenditures.

²⁸ Quasi-sentences are usually sentences but also include the bits between colons and semi-colons in languages where periods are used.

²⁹ Budge, , Robertson, and Hearl, , eds, Ideology, Strategy and Party Change, pp. 458–67, gives the complete coding scheme.Google Scholar

³⁰ Certain ‘positional’ codings were included experimentally in the scheme (for or against defence expenditures, for example). Rarely do parties want to be seen as against strong defence, however, so the negative category is usually thinly populated. Intentions are signalled by emphasizing other areas such as ‘peace’. Quite as sharp a gap thus opens up between the parties as if they had opposed each other directly on the same policy.

³¹ See McKelvey, , ‘General Conditions for Global Intransitives’.Google Scholar

³² The argument for concentrating on the unidimensional space is reinforced by the finding that for the twenty-dimensional representation distances are best measured by summing them up on each separate dimension, rather than along the shortest line between points – a ‘city-block’ as opposed to ‘Euclidian’ metric. This implies that preferences on each dimension are separate and hence avoids problems of cyclic preferences and instability. See Enelow, and Hinich, , A Spatial Theory of Voting, pp. 18–20Google Scholar; Ordeshook, Peter C., Game Theory and Political Theory (Cambridge: Cambridge University Press, 1986), p. 250.CrossRefGoogle Scholar Such a ‘city-block’ metric is rather a complicated idea to get over in public debate and media comment, which partly accounts for the fact stressed in the text following that parties present policies and the media discuss them in one-dimensional left-right terms, even though in deciding on policy in government they use many separable dimensions.

³³ Enelow, and Hinich, , A Spatial Theory of Voting, pp. 38–40Google Scholar, recognized this tendency to focus public and media discussion on one (left-right) dimension, in the concept of a ‘predictive dimension' on the basis of which electors make inferences about parties’ specific policies.

³⁴ Laver, and Budge, , Party Policy and Government Coalitions, p. 27Google Scholar, words in brackets inserted. Although the dimension derives inductively from data on post-war party programmes, these in fact comprehend all the classical political ideologies. The Gladstonian Liberal mix of free market and peaceful internationalism is well represented by the German Liberals (see Figure 1b), while several Scandinavian populist parties take up positions not dissimilar to those of Ross Perot and Silvio Berlusconi in recent elections.

³⁵ As the number of predictions is equal to the number of cases, their success rate is also their efficiency rate (equivalent to the number of successful predictions divided by total number in the prediction set). Thus the efficiency rates are high by normal standards in social science. See Laver, and Budge, , Party Policy and Government Coalitions, pp. 416–17.Google Scholar

³⁶ As policy areas seem to be separable both in party programmes and government decision making, there is also no implication that spending more on one policy area (such as welfare) implies less on another (such as defence). Welfare can be financed in many other ways than raising taxes – inflation, foreign borrowing, expanding economic activity, etc.

³⁷ Laver, and Budge, , Party Policy and Government Coalitions, p. 417.Google Scholar This result holds whether the same left-right dimension employed here is used as the spatial representation, or whether a twenty-dimensional city-block space is used (in which case the party is that with the median legislator on a majority of the separable dimensions).

³⁸ For a good summary, see Enelow, and Hinich, , A Spatial Theory of Voting, pp. 8–30.Google Scholar