Abstract
This article investigates possible bandwagons in actual elections rather than the usual opinion poll data. Until 1918, British general elections were staggered over a fortnight or more. We use the eight general elections between 1885 and 1910 to investigate whether there was a general bandwagon or underdog effect as the election progressed. We find that any bandwagon effect was in favor of the party which eventually won the election, not the party gaining seats compared with last time. We also find that a typical election featured an initial bandwagon effect which peaked about halfway through the election and then declined. Its decline appears to be due both to declining enthusiasm for the leading party and to later polls occurring in places where voters were less prone to get on a bandwagon in the first place. The weakening of the bandwagon was correlated to distance of the constituency from London, although it revived to some extent in Scotland.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Nadeau et al. (1994) is a partial exception to this in that they focus on voters’ expectations over the outcome of the next election. However, one’s expectations are likely to be significantly influenced by the polls.
Chancellor of the Exchequer in the Liberal government defeated at the 1895 election.
Although it has been repeatedly argued that if parties do know what their valences will be at the forthcoming election before they choose their ideological position, then the party with the lower valence will be the one that positions itself further from the center as it tries to get a compensating advantage among at least some voters (or possibly activists) by distancing itself ideologically from its rival (Macdonald and Rabinowitz 1998; Groseclose 2001; Schofield 2004; Bruter et al. 2010).
Another point is that the model implicitly assumes that uninformed voters don’t know the distribution of the electorate around y. If they do, then any \(\hat{y}\neq y\) will be rejected by at least one uninformed voter as impossible, because there must be at least one voter outside the range \((\hat{y}-1, \hat{y}+1)\). Such voters will have to make an estimate of y, and therefore of x, that is closer to the true value, drawing day 2’s results closer to the final results on day 3. If \(\hat{x}<x\), this adjustment will strengthen the bandwagon (if it is in favor of party R) on day 2 but leaves it with less extra mileage to run on day 3. If \(\hat{x}>x\), the bandwagon will now be weaker on day 2 and this will correspondingly weaken (but cannot wipe out) its reversal on day 3. We are grateful to one of the reviewers of the first draft of this article for suggesting both this line of inquiry, and also the point about possible unanimity discussed above.
We exclude Irish and university constituencies as these constituencies did not reflect the political dynamics of the remainder of the United Kingdom.
This exclusion results in 192 observations being dropped. Repeating our analysis treating Labour and the Liberals as one party (on the basis that a vote for either of these parties represents a vote against the Unionists) does not significantly alter our results (results not reported).
Under the Septennial Act of 1715, no parliament could last more than seven years, but by convention the prime minister would ask for a dissolution in the sixth year of a parliament. But he might also call an election if he lost a vote of confidence, and this explains the premature election of 1886. In 1895, it was the incoming government (after the Liberals had lost a vote of confidence and resigned) that called the election. Both elections of 1910 came about during the Liberal government’s clash with the House of Lords over Lloyd George’s budget of 1909: in January they sought a mandate to push the budget through and in December to reform the Lords themselves.
To ensure that serial correlation did not effect our results, a feasible generalized least squares procedure allowing for AR(1) errors was also used to test the robustness of our results. This did not alter the substance of our results (not reported).
To test the robustness of our estimates to alternative samples we re-estimated our baseline model (equivalent to column 1 in Table 2) dropping each of the sample years in turn. Both (d j2±d j1) and \((d_{j2}^{2}\pm d_{j1}^{2})\) retain their significance at the 1 % level in all samples (results not reported).
A third possibility is that, as the eventual winner became more certain, turnout fell as voters downgraded their chances of making a difference (though this would reverse a bandwagon only if voters on the winning side did this disproportionately). We tested this by adding the change in each constituency’s turnout since the previous election as a right-hand side variable. If the slowing of the bandwagon effect was partly due to this kind of turnout effect, the coefficient on this new variable should be negative and significant, and it should also reduce the magnitude of the coefficient on the quadratic term (which would until now have been capturing the effect of the omitted variable). Since neither of these turned out to be the case, we conclude that the decline of the bandwagon was not due to voters’ falling hopes of being pivotal.
Perhaps this was because the Scottish population was so concentrated around the Edinburgh and Glasgow areas. In 1911, 55 % of Scots lived in what is now the Strathclyde and Lothian region, compared with the 20 % of English who lived in the London area.
References
Asch, S. E. (1951). Effects of group pressure on the modification and distortion of judgments. In H. Guetzkow (Ed.), Groups, leadership, and men. Pittsburgh: Carnegie Press.
Asch, S. E. (1952). Social psychology. Englewood Cliffs: Prentice Hall.
Asch, S. E. (1955). Opinions and social pressure. Scientific American, 193, 31–35.
Asch, S. E. (1956). Studies of independence and conformity: a minority of one against a unanimous majority. Psychological Monographs, 70.
Banerjee, A. V. (1992). A simple model of herd behavior. Quarterly Journal of Economics, 107(3), 797–818.
Breuch, T. S., & Pagan, A. R. (1979). Simple test for heteroscedasticity and random coefficient variation. Econometrica, 47(5), 1287–1294.
Bruter, M., Erikson, R. S., & Strauss, A. B. (2010). Uncertain candidates, valence, and the dynamics of candidate position-taking. Public Choice, 144(1–2), 153–168.
Callander, S. (2007). Bandwagons and momentum in sequential voting. Review of Economic Studies, 74(3), 653–684.
Caramani, D. (2000). Elections in Western Europe since 1815: electoral results by constituencies. London: Macmillan.
Ceci, S. J., & Kain, E. L. (1982). Jumping on the bandwagon with the underdog: the impact of attitude polls on polling behaviour. Public Opinion Quarterly, 46(2), 228–242.
Craig, F. W. S. (1974). British parliamentary election results: 1885–1918. London: Macmillan.
Crano, W. D. (2000). Milestones in the psychological analysis of social influence. Group Dynamics, 4(1), 68–80.
Delli Carpini, M. X. (1984). Scooping the voters? The consequences of the networks’ call of the 1980 presidential race. The Journal of Politics, 46(3), 866–885.
Fleitas, D. W. (1971). Bandwagon and underdog effects in minimal-information elections. American Political Science Review, 65(2), 434–438.
Gelman, A., & King, G. (1994). Why are American presidential campaign polls so variable when votes are so predictable? British Journal of Political Science, 23(4), 409–451.
Goot, M. (2009). Underdogs, bandwagons or incumbency? Party support at the beginning and end of Australian election campaigns, 1983–2007. Australian Cultural History, 28, 69–80.
Groseclose, T. (2001). A model of candidate location when one candidate has a valence advantage. American Journal of Political Science, 45(4), 862–886.
Hood, W. R., & Sherif, M. (1962). Verbal report and judgment of an unstructured stimulus. The Journal of Psychology, 54(1), 121–130.
Jackson, J. E. (1983). Election night voting and voter turnout. American Journal of Political Science, 27(4), 615–635.
Knight, B., & Schiff, N. (2010). Momentum and social learning in presidential primaries. Journal of Political Economy, 118(6), 1110–1150.
Lloyd, T. (1968). The general election of 1880. Oxford: Oxford University Press.
Lott, J. R. (2005). The impact of early media election calls on republican voting rates in Florida’s Western Panhandle counties in 2000. Public Choice, 123(3–4), 349–361
Magalhaes, P. C. (2004). In exposure to polls, cognitive mobilization, and voting behavior: the 2002 general elections in Portugal. Research note [online]. Available from: http://www.ics.ul.pt. Accessed 18 May 2011.
McAllister, I., & Studlar, D. T. (1991). Bandwagon, underdog, or projection? Opinion polls and electoral choice in Britain, 1979–1987. The Journal of Politics, 53(3), 720–741.
Macdonald, S. E., & Rabinowitz, G. (1998). Solving the paradox of nonconvergence: valence, position and direction in democratic politics. Electoral Studies, 17(3), 281–300.
Mehrabian, A. (1998). Effects of poll reports on voter preferences. Journal of Applied Social Psychology, 28(1), 119–130.
Morton, R. B., & Williams, K. C. (2000). Learning by voting. Ann Arbor: University of Michigan Press.
Nadeau, R., Niemi, R. G., & Amato, T. (1994). Expectations and preferences in British general elections. American Political Science Review, 88(2), 371–383.
Newspaper Press Directory (1846–1976). London: C. Mitchell.
Rhodes James, R. (1963). Rosebery. London: Weidenfeld and Nicolson.
Schofield, N. (2004). Equilibrium in spatial valence models of politics. Journal of Theoretical Politics, 16(4), 447–481.
Shannon, R. (1999). Gladstone: heroic minister. London: Allen Lane.
Sherif, M. (1935). A study of some social factors in perception. Archives of Psychology, 27, 1–60.
Sherif, M. (1936). The psychology of social norms. New York: Harper & Row.
Sudman, S. (1986). Do exit polls influence voting behavior? Public Opinion Quarterly, 50(3), 331–339.
Triplett, N. (1898). The dynamogenic factors in pacemaking and competition. The American Journal of Psychology, 9(4), 507–533.
Vincent, J. R., & Stenton, M. M. (Eds.) (1971). McCalmont’s parliamentary poll book of all elections, 1832–1918. Brighton: Harvester Press.
West, D. M. (1991). Polling effects in election campaigns. Political Behavior, 13(2), 151–163.
Wooldridge, J. M. (2002). Econometric analysis of cross section and panel data. Cambridge: MIT Press.
Acknowledgements
We would like to thank the Leverhulme Trust for funding the project of which this article is a part; and Jonathan Barry, Andrew Pickering, Adam Zerny and three anonymous referees for Public Choice for their comments on earlier versions of the article.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix 1: Variable definitions
- Swing to Winner: :
-
The swing to the winner is calculated as follows:
(Δ Winner’s percentage share of the vote −Δ Loser’s percentage share of the vote)/2
When there was more than one candidate for a party standing, for example in a two member constituency, the total vote for all candidates of each party was used.
- Swing to Gainer: :
-
As above, with winner and loser replaced by party gaining ground and party losing ground.
- Incumbent: :
-
A trichotomous variable that takes the value of 1 if all of the incumbents are standing for re-election, 0.5 if one but not all incumbents are standing for re-election and 0 if no incumbents are standing for re-election.
- Majority: :
-
The percentage point lead in terms of the total vote in a constituency that the winning party has over the second place party. In two-member constituencies, majority is the percentage point lead of the lowest-polling successful candidate over the highest-polling unsuccessful candidate of a different party.
- Scotland (Wales): :
-
A dummy variable that takes a value of one if the constituency is in Scotland (Wales) and zero otherwise.
- County: :
-
A dummy variable that takes a value of one if the constituency is a County constituency (as opposed to a borough) and zero otherwise.
- Third Party Vote: :
-
The percentage of the total vote in a particular constituency that is not obtained by either the Conservative party, Liberal Unionist Party or the Liberal Party.
- Distance: :
-
The distance in miles as the crow flies using the postcode EC4A 2DY (Fleet Street, London) to the local council offices applicable for that borough or county. Distance as the crow flies was calculated using the website Free Map Tools http://www.freemaptools.com/distance-between-uk-postcodes.htm (accessed 13/10/11). If no council offices exist we used the high street. For counties we chose the town after which the county seat is named. Thus for the county seat Devon Honiton, we use Honiton for the purposes of the distance calculation. Where county seats do not include a town in the name we used the town closest to the center of the constituency. For example, for the county seat Norfolk Northern we used the town of Aylsham.
Appendix 2: Summary statistics
Rights and permissions
About this article
Cite this article
Hodgson, R., Maloney, J. Bandwagon effects in British elections, 1885–1910. Public Choice 157, 73–90 (2013). https://doi.org/10.1007/s11127-012-0027-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11127-012-0027-9