Skip to main content Accesibility Help

Ideology, Party Factionalism and Policy Change: An integrated dynamic theory

  • Ian Budge, Lawrence Ezrow and Michael D. McDonald

Operationalized as a simulation and checked against 1,737 policy shifts in twenty-four post-war democracies, this theory of party position-taking offers both an explanation and specific postdictions of party behaviour, synthesizing some previous approaches and linking up with mandate theories of political representation. These wider implications are considered at the beginning and the end of the article.

Hide All

1 Downs, Anthony, An Economic Theory of Democracy (New York: Harper, 1957), pp. 115118.

2 Downs, , Economic Theory, pp. 122–7.

3 Barry, Brian, Sociologists, Economists and Democracy (London: Collier Macmillan, 1979), pp. 118125.

4 McDonald, Michael D., Mendes, Silvia and Budge, Ian, ‘What Are Elections For?’ British Journal of Political Science, 34 (2004), 126.

5 Downs, , Economic Theory, pp. 96–111.

6 For a critical evaluation, see Barry, Sociologists, Economists and Democracy. For formal modelling, see, for example, Endow, James M. and Hinich, Melvin J., The Spatial Theory of Voting (Cambridge: Cambridge University Press, 1984).

7 The activities of the MRG-CMP and the data produced by them are extensively documented in Budge, Ian, Klingemann, H-D., Volkens, Andrea, Bara, Judith, Tanenbaum, Eric et al. , Mapping Policy Preferences (Oxford: Oxford University Press, 2001), and Klingemann, H-D., Volkens, Andrea, Bara, Judith, Budge, Ian and McDonald, Michael D., Mapping Policy Preferences II (Oxford: Oxford University Press, 2006). Both books are sold with attached CDs containing documentation and data. As we use these for our own analysis, they are described in detail below.

8 The Manifesto dataset and the construction of the left–right scale, which we use in our own analysis, are described in more detail in our data section below. The claim of the scale to provide the best summary representation of public and party policy is buttressed by the spontaneous emergence of a powerful leading left–right dimension from factor analyses of the Manifesto data reported in Budge, Ian, Robertson, David and Hearl, Derek J., eds, Ideology, Strategy and Party Movement (Cambridge: Cambridge University Press [1987], 2008), and Gabel, Matthew J. and Huber, John D., ‘Putting Parties in Their Place’, American Journal of Political Science, 44 (2000), pp. 94103.

9 Similar patterns are found in other countries as reported in the two Mapping Policy Preferences volumes. However, as they mostly have multiparty systems, non-convergence would have been expected from Downs’s arguments about this case in the Economic Theory, pp. 122–5 (see Fig. 2).

10 The studies are listed in the notes to Table 1. The Eurobarometer survey, sponsored by the EU in its member and candidate states, includes a left–right self-placement scale for electors.

11 The rules are cast in a prescriptive form on the basis of the observed correlations between opinion shifts and party changes of position in at least eight European countries.

12 For empirical evidence on this point, see Adams, James, Clark, Michael, Ezrow, Lawrence and Glasgow, Garrett, ‘Are Niche Parties Fundamentally Different from Mainstream Parties? The Causes and Electoral Consequences of Western European Parties’ Policy Shifts, 1976–1998’, American Journal of Political Science, 50 (2006), 513529.

13 For further analyses which support this conclusion, see McDonald, Michael D. and Budge, Ian, Elections, Parties, Democracy: Conferring the Median Mandate (Oxford: Oxford University Press, 2005).

14 Budge, Ian, ‘A New Spatial Theory of Party Competition: Uncertainty, Ideology and Policy Equilibria Viewed Comparatively and Temporarily’, British Journal of Political Science, 24 (1994), 443468.

15 For a more extended analysis of ‘wandering’ and ‘leapfrogging’, see McDonald, M. D. and Budge, Ian, Elections, Parties, Democracy: Conferring the Median Mandate (Oxford: Oxford University Press, 2005), pp. 6273: Klingemann, et al. , Mapping Policy Preferences II, pp. 67–74.

16 Budge, , ‘A New Spatial Theory’, p. 461.

17 Cf. Wlezien’s ‘thermostatic’ model. See also Wlezien, Christopher, ‘Dynamics of Representation: The Case of US Spending on Defence’, British Journal of Political Science, 26 (1996), 81103.

18 Robertson, David, A Theory of Party Competition (London: Wiley, 1976).

19 Of course, some of the other rules fit almost as well.

20 Laver, Michael‘Policy and the Dynamics of Party Competition’, American Political Science Review, 99 (2005), 263281, The ‘New Spatial Theory’ is not, however, cited in his discussion.

21 One should note here the same difficulty as occurred with the NST – parties are characterized by the decision rule which fits them best, in some sense, but it is plausible that other rules fit them almost as well. This provides some grounds for thinking that the rules are not as independent of each other as they appear. We follow up this point below.

22 Tavits, Margit, ‘Principle vs Pragmatism: Policy Shifts and Political Competition’, American Journal of Political Science, 51 (2007), pp. 151165. For a compelling spatial analysis of Democratic and Republican policy movements on economic and social dimensions in the United States since 1896, see Miller, Gary and Schofield, Norman, ‘Activists and Partisan Realignment in the United States’, American Political Science Review, 97 (2003), 245260.

23 Money, Jeanette and Andrews, Josephine, ‘Parties’ Electoral Strategies: An Empirical Analysis’ (unpublished paper, Department of Political Science, University of California, Davis, 2007).

24 Adams, James, ‘A Theory of Spatial Competition with Biased Voters’, British Journal of Political Science, 31 (2001), 210223.

25 Apart from anything else, some electors will not vote.

26 Weak correlations in the 0.20 to 0.40 range have been attained only by controlling for other aspects of the data which seem to characterize them more powerfully, such as reactions to past vote and strong alternation of policy between left and right. The actual findings are sometimes contradictory and often explained on ideological grounds internal to parties rather than purely external vote-seeking ones – (‘niche’ parties, for example).

27 For a formal proof of the disutility of being in office and pursuing repugnant policies, see Budge, Ian and Farlie, Dennis, Voting and Party Competition (London: Wiley, 1977), pp. 150160.

28 Harmel, Robert and Tan, Alexander C., ‘Party Actors and Party Change: Does Factional Dominance Matter?’ European Journal of Political Research, 42 (2003), 409424; Giannetti, Daniela and Benoit, Kenneth, eds, Intra Party Politics and Coalition Governments (London: Routledge, 2008).

29 Cf. Lipset and Rokkan’s observation in their study of West European party systems that the party families in 1964 looked largely as they did in 1918 ( Lipset, S. M. and Rokkan, Stein, Party Systems and Voter Alignments (New York: Free Press, 1967) – substantiated in a careful historical study by Bartolini, Stefano and Mair, Peter, Identity, Competition and Electoral Availability: The Stabilization of European Electorates 1885–1986 (Cambridge: Cambridge University Press, 1990)).

30 See, for example, the review article by Nanestad, Peter and Paldam, Martin, ‘The Cost of Ruling’, in Han Dorussen and Michael Taylor, eds, Economic Voting (London: Routledge, 2002), pp. 1744.

31 We should, however, stress again that there is no unified party decision maker postulated as enforcing these moves, which are rather outcomes of a process of factional alternation.

32 Budge, et al. , Mapping Policy Preferences, pp. 41–7.

33 McDonald, M. D. and Budge, Ian, Elections, Parties, Democracy: Conferring the Median Mandate (Oxford: Oxford University Press, 2005), p. 21.

34 Budge, et al. , Mapping Policy Preferences; Klingemann et al., Mapping Policy Preferencees II. Both publications are sold with a CD containing all the data we use here.

35 The MRG-CMP counted the (quasi-) sentences of all the significant parties’ programmes in each post-war election, first for twenty countries and subsequently for fifty-four (including the twenty-five we focus on here) into one and only one of fifty-six policy categories. ‘A quasi sentence is defined as an argument or phrase which is the verbal expression of one idea or meaning. It is often marked off in a text by commas or (semi) colons’ (Klingemann et al., Mapping Policy Preferences II, p. xxiii). To create the left–right scale, ‘left’ policy categories were identified on theoretical grounds and the percentaged quasi-sentences falling into them added up. ‘Right’ categories were similarly identified and the percentages added up. The sum of the left percentages were then subtracted from the sum of the right percentages to give the overall additive scale running from −100 (left) to +100 (right). As the summed percentages are based on the total number of (quasi-) sentences in a manifesto, scores can also be affected by the number of non left–right references in the document. In this sense, the left–right scale reflects policy tendencies across the whole document. ‘Left’ categories broadly cover government intervention, welfare and peace, while right ones are concerned with freedom, traditional values and military strength. Each party programme can thus be given a single left–right score and movement to the left or right estimated by the difference between programme scores in successive elections.

36 Tavits, M., Principle vs Pragmatism, p. 156; Klingemann et al., Mapping Policy Preferences II, p. 105.

37 Another consideration for choosing the American Republicans is that readers will be reasonably familiar with this example.

38 The website address is

39 The countries included in the analyses are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Israel, Italy, Japan, Luxembourg, the Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom and the United States.

40 The left–right estimates of parties’ positions remained totally unchanged for 125 inter-election shifts in the analysis.

41 For these analyses, Australia and France are classified with the plurality systems (even though they have essentially majority-based rules).

42 The test thus takes on the statistical form of Λ, in terms of the formula: Λb = (Number of errors under alternative form of prediction – Number of errors under rules from integrated theory)/Number of errors under alternative form of prediction. The value varies between 0 and 1.00.

43 Error in this context means that a ‘true’ leftward shift has been inaccurately measured or coded as a rightward shift in the data. We assume – for this limited set of analyses – that these errors are balanced over the 1,737 party-policy shifts analysed.

44 However, within a context of increasing interest in their explanatory potential, cf. Aldrich’s use of activist factions to explain non-convergence under the Downsian two-party model ( Aldrich, John, ‘A Downsian Spatial Model with Party Activism’, American Political Science Review, 77 (1983), 974990): also Harmel and Tan, ‘Party Actors and Party Change’; and Gianetti and Benoit, Intra Party Politics.

* Budge and Ezrow: Department of Government, University of Essex (email: ); McDonald: Department of Political Science, Binghamton University, SUNY. The authors thank Hershbinder Mann for his indispensable work on tracing party movements. In addition, they are grateful to Sarah Birch for her incisive comments, and the anonymous reviewers for their insightful comments on previous versions.

Recommend this journal

Email your librarian or administrator to recommend adding this journal to your organisation's collection.

British Journal of Political Science
  • ISSN: 0007-1234
  • EISSN: 1469-2112
  • URL: /core/journals/british-journal-of-political-science
Please enter your name
Please enter a valid email address
Who would you like to send this to? *


Full text views

Total number of HTML views: 0
Total number of PDF views: 0 *
Loading metrics...

Abstract views

Total abstract views: 0 *
Loading metrics...

* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.

Usage data cannot currently be displayed