¹ The centrist parties discussed in this article, and the elections they contested in the period analysed, are the Liberal party (1945–79), the Alliance of the Liberal party and the Social Democratic party (1983 and 1987), and the Liberal Democratic party (1992 onward). The Liberals and Social Democrats maintained separate organizations but campaigned under an electoral pact, so their candidates did not oppose each other in any constituency. In 1988, the Liberals and SDP merged to form the Social and Liberal Democrats. In 1990, the merged party simplified its name to Liberal Democrats. For a definitive history of the SDP and the merger, see Crewe, Ivor and King, Anthony, SDP: The Life and Death of the Social Democratic Party (Oxford: Oxford University Press, 1995)Google Scholar. To avoid awkward constructions, we use ‘Liberals’ and ‘Liberal vote’ as generic terms, employing the later names of the centrist party or parties only when the context is restricted to the periods of their existence.

² See also the Appendix for numerical data. Vote shares throughout the article are for Great Britain only. Northern Ireland, which usually records little more than 2 per cent of the overall United Kingdom vote, has a separate party system. Since 1974, the Conservative, Labour and Liberal parties have generally not offered candidates there. Unless otherwise noted, electoral data through 1999 in this article are drawn from Rallings, Colin and Thrasher, Michael, British Electoral Facts, 1832–1999 (Aldershot, Surrey: Ashgate, 2000)Google Scholar. Data for 2001 and 2005 are from the website of the UK Electoral Commission, http://www.electoralcommission.org.uk/election-data/index.cfm. Manifesto scores were provided by Eric Tannenbaum and Judith Bara of the Comparative Manifesto Project. Scores for 1945–97 are also available on a CD included with Budge, Ian, Klingemann, Hans-Dieter, Volkens, Andrea, Bara, Judith and Tanenbaum, Eric, Mapping Policy Preferences: Estimates for Parties, Electors, and Governments 1945–1998 (Oxford: Oxford University Press, 2001)Google Scholar. The measures are described more fully below.

³ Douglas, Roy, Liberals: The History of the Liberal and Liberal Democrat Parties (London: Continuum International Publishing Group, 2005), pp. 281–282. Previously, after the February 1974 election, Labour had formed a minority government with 301 seats, seventeen short of a majority. However, at that time, the Liberals, with fourteen MPs, did not by themselves hold the balance of power. An absolute majority would have required help from the seven MPs of the Scottish National Party (SNP). The Labour government called a new election in October 1974, from which it emerged with an absolute majority of 319 seats. The subsequent Lib–Lab pact of 1977–78 became necessary after Labour lost its majority due to defections and by-election defeats.Google Scholar

⁴ Duverger, Maurice, Political Parties (New York: Wiley, 1954)Google Scholar; Riker, William H., ‘The Two-Party System and Duverger’s Law: An Essay on the History of Political Science’, American Political Science Review, 76 (1982), 753–766.CrossRefGoogle Scholar

⁵ Cox, Gary W., Making Votes Count: Strategic Coordination in the World’s Electoral Systems (Cambridge: Cambridge University Press, 1997)CrossRefGoogle Scholar; Chhibber, Pradeep K. and Kollman, Ken, The Formation of National Party Systems: Federalism and Party Competition in Canada, Great Britain, India, and the United States (Princeton, N.J.: Princeton University Press, 2004).Google Scholar

⁶ Downs, Anthony, An Economic Theory of Democracy (New York: Harper & Row, 1957)Google Scholar.

⁷ Arend Lijphart’s dichtomization of ‘majoritarian’ and ‘consensus’ democracies causes him and many other political scientists to neglect the important differences between majority and plurality rules. See, e.g., Lijphart, , Patterns of Democracy: Government Forms and Performance in Thirty-six Countries (New Haven, Conn.: Yale University Press, 1999)Google Scholar. For a perspective that emphasizes the latter distinction, see Nagel, Jack H., ‘Expanding the Spectrum of Democracies’, in Democracy and Institutions: A Festschrift for Arend Lijphart, edited by Markus Crepaz, Thomas Koelble and David Wilsford (Ann Arbor: University of Michigan Press, 2000), pp. 113–128Google Scholar.

⁸ These results were established through computer simulations by Merrill, Samuel, Making Multicandidate Elections More Democratic (Princeton, N.J.: Princeton University Press, 1988)CrossRefGoogle Scholar and analytically by Cox, Gary, ‘Centripetal and Centrifugal Incentives in Electoral Systems’, American Journal of Political Science, 34 (1990), 903–935CrossRefGoogle Scholar.

⁹ Cox, , Making Votes Count, esp. pp. 231–237.Google Scholar

¹⁰ Hazan, Reuven V., Centre Parties: Polarization and Competition in European Parliamentary Democracies (London: Pinter, 1997).Google Scholar

¹¹ Bingham, Powell G. Jr, Elections as Instruments of Democracy: Majoritarian and Proportional Visions (New Haven, Conn.: Yale University Press, 2000), pp. 240–246Google Scholar; Powell, and Vanberg, Georg S., ‘Election Laws, Disproportionality and Median Correspondence: Implications for Two Visions of Democracy’, British Journal of Political Science, 30 (2000), 383–411.Google Scholar

¹² McDonald, Michael D. and Budge, Ian, Elections, Parties, Democracy: Conferring the Median Mandate (Oxford: Oxford University Press, 2005), pp. 124–130CrossRefGoogle Scholar; McDonald, , Mendes, Silvia M. and Budge, , ‘What Are Elections For?’ British Journal of Political Science, 34 (2004), 1–26CrossRefGoogle Scholar. Despite frequent failure of governments to represent the median voter, public expenditures in Britain were nevertheless fairly responsive to changes in public preference over the 1978–95 time period; Soroka, Stuart N. and Wlezien, Christopher, ‘Opinion-Policy Dynamics: Public Preferences and Public Expenditures in the United Kingdom’, British Journal of Political Science, 35 (2005), 665–689.CrossRefGoogle Scholar

¹³ Cf. Hazan, Centre Parties, pp. 158–60.

¹⁴ For alternative approaches to explaining the dynamics of party ideologies as measured by manifesto scores, see the substantial literature that begins with Budge, Ian, ‘A New Spatial Theory of Party Competition: Uncertainty, Ideology, and Policy Equilibria Viewed Comparatively and Temporally’, British Journal of Political Science, 24 (1994), 443–467CrossRefGoogle Scholar. Recent articles in this tradition include Adams, James, Clark, Michael, Ezrow, Lawrence and Glasgow, Garrett, ‘Understanding Change and Stability in Party Ideologies: Do Parties Respond to Public Opinion or to Past Election Results?’, British Journal of Political Science, 34 (2004), 589–610CrossRefGoogle Scholar; and Somer-Topcu, Zeynep, ‘Timely Decisions: The Effects of Past National Elections on Party Policy Change’, Journal of Politics, 71 (2009), pp. 238–248.CrossRefGoogle Scholar Tests in this literature are based on pooled cross-national data; and do not involve hypotheses that relate parties’ positions to the votes won by a competing party in the preceding election.

¹⁵ For detailed accounts of how the scale is constructed, see Laver, M. J. and Budge, Ian, eds, Party Policy and Government Coalitions (New York: St. Martin’s Press, 1992)CrossRefGoogle Scholar, chap. 2; Klingemann, Hans-Dieter et al. , Parties, Policies, and Democracy (Boulder, Colo.: Westview Press, 1994)Google Scholar, chap. 3; and especially Budge et al., Mapping Policy Preferences.

¹⁶ For examples, see Budge, Ian, ‘Party Policy and Ideology: Reversing the 1950s?’, in Geoffrey Evans and Pippa Norris, eds, Critical Elections: British Parties and Voters in Long-Term Perspective (London: Sage, 1999), pp. 1–22Google Scholar; Budge, Ian, Crewe, Ivor, McKay, David and Newton, Ken, The New British Politics (Harlow, Herts.: Addison Wesley Longman, 1998), p. 410Google Scholar; Norris, Pippa, Electoral Change since 1945 (Oxford: Blackwell, 1997), p. 158Google Scholar; and Bara, Judith, ‘The 2005 Manifestos: A Sense of ‘Déjà vu’?’, Journal of Elections, Public Opinion and Parties, 16 (2006), 265–281.CrossRefGoogle Scholar

¹⁷ Some of these results are reported below; others are available upon request.

¹⁸ An examination of Figure 3 shows that moves by a major party caused all of these anomalies, except in 1964 when the Liberals shifted sharply but temporarily to the left.

¹⁹ This was suggested by a reviewer, to whom we are most grateful.

²⁰ Russell, Andrew and Fieldhouse, Edward, Neither Left nor Right? The Liberal Democrats and the Electorate (Manchester: Manchester University Press, 2005), pp. 90–93.Google Scholar

²¹ The pattern is unchanged when the lagged Liberal vote is included in the equation – its effect is small and insignificant (b = 0.09; s.e. = 0.25). Full-fledged error correction modelling shows further that the effects of the current (time t) changes in Labour and Conservative positions are effects of the corresponding lagged (time t−1) positions. This is important because it strongly suggests that party positions Granger-cause the Liberal vote.

²² The pattern is robust to the inclusion of other variables. Following the suggestions of one anonymous referee, we included the interest rate (following the model of Sanders, David, ‘Economic Performance, Management Competence and the Outcome of the Next General Election’, Political Studies, 44 (1996), 203–231, and doing so did not significantly alter the effects of Labour-Liberal and Conservative-Liberal distance. To probe the nature of voter response, we included lagged changes in our distance measures, and this made little difference to the results as well.CrossRefGoogle Scholar

²³ In these equations, we use the Labour and Conservative left–right scores alone, rather than the differences between them and the Liberal score as in Equation 1. At the beginning of the campaign for election t, each major party controls its own manifesto, but probably cannot calibrate it against the Liberal manifesto, which is published almost simultaneously. For those who would question this premise, we also estimated versions of Equations 3 and 4 using algebraic differences between the major-party and Liberal scores. The results are virtually the same for Equation 3 and far weaker for Equation 4.

²⁴ The effect also holds when the equation includes the lagged Conservative position, the coefficient for which is insignificant (b = 0.27; s.e. = 0.18).

²⁵ In contrast, the occupied-centre relation between the Liberal vote and the Conservative left–right position remained strong during 1966–87, with a correlation of 0.74 (p = 0.06).

²⁶ Adams, James and Merrill, Samuel III, ‘Why Small, Centrist Third Parties Motivate Policy Divergence by Major Parties’, American Political Science Review, 100 (2006), 403–417CrossRefGoogle Scholar. Their analysis in part seeks to explain results first presented in an early version of this article: Nagel, Jack H., ‘Center-Party Strength and Major-Party Polarization in Britain’ (paper presented at the Annual Meeting of the American Political Science Association, San Francisco, 2001).Google Scholar

²⁷ Here is the regression result:

²⁸ As a referee has pointed out to us, Adams and Merrill implicitly assume a uniform distribution of voters across the ideological spectrum, which seems unlikely. Their hypothesis might be salvaged by positing that voters on the extremes have a greater propensity to defect (to non-voting or fringe parties) in response to policy moves than do supporters of the centrist third party.

²⁹ Rallings, and Thrasher, , British Electoral Facts, 1832–1999, pp. 38–66Google Scholar; UK Electoral Commission web site for 2001 and 2005.

³⁰ The CMP codes for statements in favour of decentralization and centralization, but does not include them in the left–right scale (Bara, ‘The 2005 Manifestos,’ Appendix 1).

³¹ Webb, Paul, ‘The Continuing Advance of the Minor Parties’, in Pippa Norris and Christopher Wlezien, eds, Britain Votes 2005 (Oxford: Oxford University Press, 2005), pp. 101–119, at p. 114.Google Scholar

³² Curtice, John and Steed, Michael, ‘Appendix 2: The Results Analysed’, in David Butler and Dennis Kavanagh, The British General Election of 1997 (Basingstoke, Hants.: Macmillan, 1997), pp. 295–325Google Scholar, at p. 305. Curtice and Steed’s analysis of 1997 results suggests that Goldsmith’s campaign had less impact on the Conservatives than was expected.

³³ Norton, , ‘The Conservative Party: The Politics of Panic,’ in John Bartle and Anthony King, eds, Britain at the Polls 2005 (Washington, D.C.: CQ Press, 2006), pp. 31–53Google Scholar, at p. 45. On the minor parties in 2005, see Webb, ‘The Continuing Advance of the Minor Parties’.

³⁴ Although we use a simple linear specification in Equation 4, there is an even closer fit to a logarithmic model.

³⁵ We calculated V*_lib by dividing the Liberals’ vote percentage in Great Britain, V_lib, by the proportion of British constituencies in which a Liberal candidate stood. Owing to variations in constituencies’ population and voter turnout, slightly different results would be obtained by the more laborious method of actually averaging the Liberal vote percentage across the constituencies their candidates contested.

³⁶ We say ‘roughly’ because tactical voting (motivated by fear of ‘wasting’ one’s vote) causes the average vote to give an imperfect measure of popular support for the Liberals.

³⁷ Because the analytic relation is multiplicative (V_lib = C_{lib *} V*_lib), we also estimated a logarithmic version of Equation 5. The results are qualitatively similar.

³⁸ Adams, and Merrill, (‘Why Small, Centrist Third Parties Motivate Policy Divergence by Major Parties’, pp. 411–412)Google Scholar suggest that their policy divergence result can explain the impact of the number of Liberal candidates on major-party polarization, but they do not consider the asymmetry of its impact on Conservatives and Labour.

³⁹ For an account of US political history that similarly depends on the influence of activists and changes in the composition of activist coalitions within parties over an extended time period, see Miller, Gary and Schofield, Norman, ‘The Transformation of the Republican and Democratic Party Coalitions in the U.S.’, Perspectives on Politics, 6 (2008), 433–450CrossRefGoogle Scholar; Schofield, Norman and Miller, Gary, ‘Elections and Activist Coalitions in the United States’, American Journal of Political Science, 51 (2007), 518–531CrossRefGoogle Scholar; and Schofield, Norman, Miller, Gary and Martin, Andrew, ‘Critical Elections and Political Realignments in the USA – 1860–2000’, Political Studies, 51 (2003), 217–540CrossRefGoogle Scholar. Their model of US politics is two-dimensional, whereas we explicitly model only one dimension in British politics, though we do recognize the influence of cross-cutting allegiances and issues.

⁴⁰ Fieldhouse, and Russell, , Neither Left nor Right? p. 90Google Scholar; Alt, James, Crewe, Ivor and Särlvik, Bo, ‘Angels in Plastic: The Liberal Surge in 1974’, Political Studies, 25 (1977), 343–368CrossRefGoogle Scholar; Curtice, John, ‘Who Votes for the Centre Now?’ in Don MacIver, ed., The Liberal Democrats (London: Prentice Hall, 1996), pp. 191–204Google Scholar; Walker, David, ‘The strange case of the Lib Dems’, Guardian, 18 September 2000.Google Scholar

⁴¹ Cyr, Arthur, Liberal Party Politics in Britain (New Brunswick, N.J.: Transaction Books, 1977), p. 76Google Scholar. Cyr emphasizes the role of Britain’s relatively rigid class structure in the twentieth-century decline of the Liberal party. ‘Liberal co-operation with the Labour Party and trade union elements was always feeble and reluctantly undertaken … The Liberals were doubly injured by the class factor … First, they rejected class as a doctrinal source of unity … [and] drew emotional and intellectual sustenance from opposition to class. Second, however, Liberals proved to be most class-conscious in their rejection of working-class allies as equal partners … The Liberals were hurt by their class discrimination while earning no profit from class solidarity’(Liberal Party Politics in Britain, pp. 52, 77). He later notes that Liberals of the 1970s ‘like the old Liberals, are distinctly class-conscious, albeit unintentionally’, p. 236.

⁴² Webb, P.D., ‘The United Kingdom’, in Richard S. Katz and Peter Mair, eds, Party Organizations: A Data Handbook on Party Organizations in Western Democracies, 1960–90 (London: Sage Publications, 1992), pp. 837–870, at p. 847.Google Scholar

⁴³ See Mair, Peter and van Biezen, Ingrid, ’Party Membership in Twenty European Democracies, 1980–2000.’ Party Politics, 7 (2001), 5–21.CrossRefGoogle Scholar

⁴⁴ Our account of Conservative party decision procedures is indebted to Fisher, Justin, ‘Political Parties: Organisational Change and Intra Party Democracy,’Google Scholar in Fisher, Justin, Denver, David and Benyon, John eds, Central Debates in British Politics (Harlow, Herts.: Longman, 2002), pp. 135–156.Google Scholar

⁴⁵ Fisher, , ‘Political Parties’, p. 137.Google Scholar

⁴⁶ Moss, Stephen, ‘I did not call Tony Blair a smarmy git’, Guardian, 21 September 2001, http://politics.guardian.co.uk/bookshelf/story/0,,644724,00.html (accessed 14 December 2007).Google Scholar

⁴⁷ Rallings, and Thrasher, , British Electoral Facts, 1832–1999, pp. 34, 37.Google Scholar

⁴⁸ Joyce, Peter, Realignment of the Left? A History of the Relationship between the Liberal Democrat and Labour Parties (London: Macmillan, 1999), pp. 97–99Google Scholar; Cook, Chris, A Short History of the Liberal Party 1900–1997, 5th edn (London: Macmillan, 1998), pp. 126–129CrossRefGoogle Scholar. The Nuffield study of 1945 notes that Liberals complained bitterly about Churchill’s decision to hold the election in July rather than in the autumn, when they could have been better prepared; see McCallum, R. B. and Readman, Alison, The British General Election of 1945 (London: Oxford University Press, 1947), pp. 15, 69Google Scholar.

⁴⁹ In Equation 1, the 1951 Liberal vote produces the second largest negative residual in the entire postwar period, slightly below 1992.

⁵⁰ Cook, , Short History of the Liberal Party, p. 133Google Scholar; Steed, Michael, ‘The Liberal Tradition’, in Don MacIver, ed., The Liberal Democrats (London: Prentice-Hall, 1996), pp. 41–61, at p. 57Google Scholar. On Liberals as a party of the Celtic, non-conformist and rural periphery, see Tregidga, Garry, The Liberal Party in South-West Britain Since 1918: Political Decline, Dormancy, and Rebirth (Exeter: University of Exeter Press, 2000)Google Scholar.

⁵¹ Steed, , ‘The Liberal Tradition’, p. 58Google Scholar. Fieldhouse and Russell find that the Liberals were dependent on the same ‘heartlands’ from 1929 until 1992. Beginning in 1997, however, they began to capture more seats in areas outside those traditional bailiwicks (see Russell, and Fieldhouse, , Neither Left Nor Right? chap. 7).Google ScholarPubMed

⁵² This is consistent with the model of Taagepera and Grofman, who speculate that the number of parties equals the number of issue dimensions plus one; thus to the extent Britain has a three-party system, its party politics should be shaped by two significant dimensions (Taagepera, Rein and Grofman, Bernard, ‘Rethinking Duverger’s Law: Predicting the Effective Number of Parties in Plurality and PR Systems – Parties Minus Issues Equals One’, European Journal of Political Research, 13 (1985), 341–352CrossRefGoogle Scholar). For an analysis of the continuing influence of religious affiliation, behaviour and belief on voters’ choices in 1992, see Kotler-Berkowitz, Lawrence A., ‘Voting Behaviour in Great Britain: A Reassessment’, British Journal of Political Science, 31 (2001), 523–554.CrossRefGoogle Scholar

⁵³ Douglas, , Liberals, p. 253.Google Scholar

⁵⁴ In both elections, the three won their seats over Labour opponents by margins of less than twenty per cent – in most cases, much less. For an invaluable archive of constituency level election results, see Kimber, Richard, British Governments and Elections Since 1945, http://www.psr.keele.ac.uk/area/uk/uktable.htmGoogle Scholar

⁵⁵ A History of the Liberal Party in the Twentieth Century (Basingstoke, Hants.: Palgrave Macmillan, 2004), p. 177.Google Scholar concurs, Douglas: ‘In the unhappy condition of the Liberal party after 1951, [Davies’ acceptance] could have proved the coup de grace.’ Liberals, p. 259Google Scholar.

⁵⁶ In 1964, the Liberals would be to the left of both major parties, a pattern that did not occur again until 1997 and 2001. From 1969 through 1992, they continuously occupied the middle position.

⁵⁷ The increase in candidates accounts for most of the rise in the Liberal vote nationally. The party’s average vote share per contested constituency improved only a little, from 15.5 per cent to 17.2 per cent; and it elected just six MPs, the same number as in 1955 and 1951. Cook, , Short History of the Liberal Party, pp. 136–140.Google Scholar

⁵⁸ Quoted in Cyr, Liberal Party Politics, p. 158.

⁵⁹ A History of the Liberal Party, pp. 187–188Google Scholar. The reaction to Suez fits what Cyr describes as a longstanding Liberal tradition of ‘taking up an issue of principle in the international field.’ Liberal Party Politics, p. 15Google Scholar.

⁶⁰ The murder attempt failed, Thorpe denied all allegations, and a jury acquitted him shortly after the election. Nevertheless, Thorpe’s constituents denied him re-election. For his own account of these events, see Thorpe, Jeremy, In My Own Time: Reminiscences of a Liberal Leader (London: Politico’s, 1999), pp. 197–206.Google Scholar

⁶¹ Fisher, Stephen D. and Curtice, John, ‘Tactical Unwind? Changes in Party Preference Structure and Tactical Voting in Britain between 2001 and 2005’, Journal of Elections, Public Opinion and Parties, 16 (2006), 55–76; data from p. 64.CrossRefGoogle Scholar

⁶² Whiteley, Paul, Stewart, Marianne C., Sanders, David and Clarke, Harold D. , ‘The Issue Agenda and Voting’, in Norris and Wlezien, eds, Britain Votes, pp. 146–161.Google Scholar

⁶³ Quinn, Thomas, ‘The Conservative Party and the “Centre Ground” of British Politics’, Journal of Elections, Public Opinion and Parties, 18 (2008), 179–199.CrossRefGoogle Scholar

⁶⁴ Liberal parties in Australia and New Zealand established long-term alliances on the centre-right with their countries’ agrarian counterparts of the British Conservatives. Australia’s alternative-vote electoral system enabled its Liberal party to sustain a permanent coalition with the organizationally separate National (formerly Country) party. In New Zealand the United party (successor to the Liberals) merged with Reform in the 1930s to establish the National party, which remains the major party of the centre-right, even though New Zealand replaced FPP elections with PR in the 1990s.

⁶⁵ New York Mayor Michael Bloomberg and other centrist leaders explored such a possibility in early 2008.

⁶⁶ Ashdown, Paddy, ‘We must not allow ourselves to be pushed into reoccupying the ground New Labour has vacated,’ Independent, 13 June 2001.Google Scholar

⁶⁷ Fisher and Curtice in ‘Tactical Unwind?’ find evidence of increased tactical switching from Conservative to Liberal Democratic in 2005, along with a decline in the previously high levels of Labour to Liberal Democrat switches.