Trade and the timing of elections

British Journal of Political Science

Volumn 36, Issue 3, 2006, Pages 437-457

  Kayser, Mark Andreas ^a  

^a UNIVERSITY OF ROCHESTER   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 33646857253     PISSN: 00071234     EISSN: 14692112     Source Type: Journal    
DOI: 10.1017/S0007123406000238     Document Type: Article

References (60)

1. Edward Tufte, Political Control of the Economy (Princeton, N.J.: Princeton University Press, 1978), p. 65.
2. Harvey D. Palmer and Guy D. Whitten, 'Government Competence, Economic Performance and Endogenous Election Dates', Electoral Studies, 19 (2000), 413-26;
3. Steven D. Roper and Christopher Andrews, 'Timing an Election: The Impact on the Party in Government', American Review of Politics, 23 (2002), 305-18;
4. Alastair Smith, 'Election Timing in Majoritarian Parliaments', British Journal of Political Science, 33 (2003), 397-418.
5. Michael D. Bordo, Barry Eichengreen and Douglas A. Irwin, 'Is Globalization Today Really Different from Globalization a Hundred Years Ago?' in Susan M. Collins and Robert Z. Lawrence, eds, Brookings Trade Forum 1999 (Washington, D.C.: Brookings Institution, 1999).
6. Michael J. Artis and Wenda Zhang, 'International Business Cycles and the ERM: Is There a European Business Cycle?' International Journal of Finance and Economics, 2 (1997), 1-16;
7. Michael J. Artis and Wenda Zhang, 'Further Evidence on the International Business Cycle and the ERM: Is There a European Business Cycle?' Oxford Economic Papers, 51 (1999), 120-32.
8. See Mark Andreas Kayser, 'Who Surfs, Who Manipulates? The Determinants of Opportunistic Election Timing and Electorally Motivated Economic Intervention', American Political Science Review, 99 (2005), 17-28.
9. Examples are Takatoshi Ito and Jin Hyuk Park, 'Political Business Cycles in the Parliamentary System', Economic Letters, 27 (1988), 233-8;
10. and Abdur R. Chowdurry, 'Political Surfing over Economic Waves: Parliamentary Election Timing in India', American Journal of Political Science, 37 (1993), 1100-18.
11. See Paul Warwick, Government Survival in Parliamentary Democracies (Cambridge: Cambridge University Press, 1994).
12. This article identifies thirteen of twenty-four OECD members in 1990 as premier-timing; Przeworski et al. list thirty-seven of fifty-nine democracies in 1990 as parliamentary, a government form that typically provides governments with great leeway in choosing when to dissolve parliament for elections (Adam Przeworski, Michael Alvarez, José Antonio Cheibub and Fernando Limongi, Democracy and Development: Political Institutions and Material Well-Being in the World, 1950-90 (Cambridge: Cambridge University Press, 2000)).
13. For example, Nathan S. Balke, 'The Rational Timing of Parliamentary Elections', Public Choice, 65 (1990), 201-16;
14. Roper and Andrews, 'Timing an Election'.
15. Daniel Diermeier and Randolph T. Stevenson, 'Cabinet Terminations and Critical Events', American Political Science Review, 94 (2000), 627-40;
16. Arthur Lupia and Kaare Strøm, 'Coalition Termination the Strategic Timing of Elections', American Political Science Review, 89 (1995), 648-65.
17. Kaare Strøm and Stephen M. Swindle, 'Strategic Parliamentary Dissolution', American Political Science Review, 96 (2002), 575-91.
18. Smith, 'Election Timing in Majoritarian Parliaments'
19. but see André Blais, Elizabeth Gidengil, Niel Nevitte and Richard Nadeau, 'Do (Some) Canadian Voters Punish a Prime Minister for Calling a Snap Election?' Political Studies, 52 (2004), 307-23.
20. Palmer and Whitten, 'Government Competence, Economic Performance and Endogenous Election Dates.'
21. Two examples, both for Japan, are Thomas F. Cargill and Michael M. Hutchinson, 'Political Business Cycles with Endogenous Election Timing: Evidence from Japan', Review of Economics and Statistics, 73 (1991), 733-9;
22. and Takatoshi Ito, 'The Timing of Elections and Political Business Cycles in Japan', Journal of Asian Economics, 1 (1990), 135-56.
23. Chowdhurry, 'Political Surfing over Economic Waves'.
24. By raising the price of imports relative to domestic goods devaluations create an expansionary substitution effect, but this is outweighed by contractionary income effects, especially in the short term. See Pierre-Richard Agenor and Peter Montiel, Development Economics (Princeton, N.J.: Princeton University Press, 1996).
25. Devaluations reduce real wealth held in the local currency and more critically, shift income from wage earners who are likely to spend it to profit recipients who are not. For a similar outcome within the context of a rational political business cycle model in which devaluations, understood as a tax on money balances, are delayed until after elections, see Ernesto Stein and J. M. Streb, 'Political Stabilization Cycles in High Inflation', Journal of Development Economics, 56 (1998), 159-80;
26. and Ernesto Stein and J. M. Streb, 'Elections and the Timing of Devaluations' (Washington, D.C.: Working Paper No. 396, Inter-American Development Bank, Research Department, 1999). Theory aside, empirical work shows that devaluations occur most frequently early in a term and very rarely shortly before elections.
27. See, for example, Sebastian Edwards, 'The Political Economy of Inflation and Stabilization in Developing Countries', Economic Development and Cultural Change, 42 (1994), 235-66;
28. or Jeffrey Frieden, Piero Ghezzi and Ernesto Stein, 'Politics and Exchange Rates: A Cross-Country Approach to Latin America' (Washington, D.C.: Working Paper No. R-421, Inter-American Development Bank, Research Department, 2000).
29. Export credits, insurance and financing, as well as trade agreements, can all boost long-term export growth but are all too unwieldy to ensure export expansion prior to impending elections.
30. 'With the synchronization of electoral calendars in large capitalist democracies, we have a recipe for an international boom and bust cycle' (see Tufte, Political Control of the Economy, chap. 3, especially p. 69).
31. William R. Thompson and Gary Zuk, 'American Elections and the International Electoral-Economic Cycle: A Test of the Tufte Hypothesis', American Journal of Political Science, 27 (1983), 464-84.
32. Ito finds a surprising positive association between elections in the United States and economic growth in Japan. See Takatoshi Ito, 'International Impacts on Domestic Political Economy: A Case of Japanese General Elections', Journal of International Money and Finance, 10 (1991), S73-89.
33. Putnam and Bayne describe such a phenomenon among the G7. See Robert D. Putnam and Nicholas Bayne, Hanging Together: Cooperation and Conflict in the Seven-Power Summits (Cambridge, Mass.: Harvard University Press, 1988).
34. Three OECD members - Greece, Spain and Turkey - allow strategic election timing but have imperfect democratic records. Stable democracy in Greece and Spain since the mid-1970s allows their inclusion. Full military rule from 1980-83 and the continuing influence of the military in government disallows Turkey.
35. King Baudouin refused Premier Wilfried Martens' s bid for governmental resignation after a cabinet session failed to resolve a split in his Christian Democrat-Liberal coalition. The monarchy continues to matter in Belgian politics. Belgium is additionally excluded because of recent co-ordination with Luxembourg to hold elections concurrently with those for the European Parliament.
36. Early elections actually occurred twice in post-war Germany prior to Gerhard Schröder's early re-election bid in 2005. Helmut Kohl engineered exactly such a failed vote of confidence in himself on 17 December 1982 and invoked Article 68 of the German Grundgesetz to dissolve parliament and call early elections. This followed unique circumstances, however, and is unlikely to be repeated. Kohl had recently come to power via a constructive vote of no confidence and sought a popular mandate. The main opposition (SPD) believed that voter disapproval of such political manœuvering would benefit them and thus did not oppose early elections. The Federal Constitutional Court later found this use of the vote of confidence legal but restricted its use. The non-partisan Federal president could also have refused the request for dissolution but did not because of opposition support. A later attempt by Kohl to move up Federal elections by seven weeks in the fall of 1990 to capitalize on reunification euphoria was opposed by the SPD and failed. An early election also occurred in 1972 after the SPD/FDP Bundestag majority eroded. Chancellor Brandt, with agreement from the CDU/CSU and President Heinemann, called a vote of confidence knowing in advance that the opposition would be able to defeat it. It is important to note that both early elections could not have occurred without the co-operation of both the opposition and the federal president. Maurice Duverger, 'A New Political System Model: Semi-Presidential Government', European Journal of Political Research, 8 (1980), 165-87.
37. Shugart and Carey later term such systems as premier-presidential (Matthew S. Shugart and John M. Carey, Presidents and Assemblies: Constitutional Design and Electoral Dynamics (Cambridge: Cambridge University Press, 1992), p. 23).
38. Duverger also identified Ireland as semi-presidential, a classification that I reject on the grounds advanced by Shugart and Carey: the Irish president has no constitutional powers and exerts no political influence (see Shugart and Carey, Presidents and Assemblies, p. 71). I include Austria because its president fails to exercise political influence despite constitutional powers. The primary function of the Austrian president is to name prime ministers but this power has been largely useless in a system that has produced grand coalitions for much of the post-war period. The long-standing domination of Austrian politics by the ÖVP/SPÖ duopoly further eroded this power and probably will again after the recent implosion of the Freiheit party. Somewhat ambiguously, the Austrian constitution provides the premier with the power to dissolve parliament after 'consultation' with parliament.
39. Finland is a president-dominant system in which the president has reserve powers over foreign policy and government formation. The large number of small parties competing in Finland's fragmented political landscape further enhances the president's powers. The abundance of small parties provide her with the de facto ability to include or exclude most parties from government almost regardless of election results - Kekkonen excluded the Finnish Conservative party from every coalition he named after 1966 - and precludes the need for early elections. Frequent cabinet crises and government resignations also rarely leave the premier in a position to consider election timing. Thus, Finland is excluded.
40. In Iceland, like Austria, the president is surprisingly acquiescent to the prime minister and parliament. This weakness is also reflected in the fact that no partisan candidate has ever sought the office. Arend Lijphart attributes this to the unusual ease by which the constitution can be amended - by a pure majority rule within parliament. See Arend Lijphart, Democracies: Patterns of Majoritarian and Consensus Government in Twenty-One Countries (New Haven, Conn.: Yale University Press, 1984), p. 189. An assertive president is likely to have his formal powers curtailed if she were ever to try to use them. Thus Iceland functions quite similarly to the Danish constitutional monarchy on which its post-independence constitution is modelled and is included in the sample. The strength and independence of the presidency in France and Portugal disqualifies them from inclusion.
41. The imposition of narrow bands of allowable exchange rate fluctuation among participating currencies in the European Exchange Rate Mechanism (ERM) in 1979 (2.25 per cent; 6 per cent for Italy) also tied European interest rates more closely together than normal pressures of capital mobility and investment would dictate. As a result, interest rates of ERM members (Belgium, Denmark, France, Germany, Ireland, Italy, Luxembourg and the Netherlands) co-varied, encouraging simultaneous economic expansions or slowdowns among members and even some closely tied non-members.
42. f) and foreign (EX) demand for domestically produced goods and services.
43. 0 is the baseline hazard. Thus the natural log of the hazard is just the log of the baseline hazard, call this γ(t), plus βX. Now if γ(t) = γ we have reproduced the exponential model; γ(t) = γt, the Gompertz model; and γ(t) = γ log t, the Weibull model. This is why Cox models require no a priori assumptions about the form of the baseline hazard.
44. Stratifying the sample in a Cox model, as I do, is equivalent to running separate regressions with the constraint that the coefficients remain equal. Differences between strata are thus attributed to the respective underlying baseline hazards.
45. The Breslow method implicitly assumes that ties really occur at discrete time and has consequently been found to bias coefficient estimates slightly downward. See, for a more complete explanation, Paul D. Allison, Survival Analysis Using the SAS System (Cary, N.C.: SAS Institute, 1995). Ties are not sufficiently common in the present data for this to constitute a major concern. Moreover, the expectation of a slight downward bias should only strengthen confidence in significant findings.
46. As discussed in the System of National Accounts (UN, 1968, §6.131 - 6.135, revised 1982) and IFS (§7, introduction to monthly volume) data descriptions, balance of payment export data is taken from customs sources which record transactions as they cross a border. This represents a substantial lag relative to domestic demand data which is recorded at the time of transaction when ownership changes hands. National Account export figures would circumvent this time inconsistency but would entail losing nearly 40 per cent of the quarterly observations in the panel. Customs based trade data are more complete as missing export data can be inferred from the import data of trade partners, as is often done with IMF Direction of Trade Statistics (DOTS) data.
47. An open question is what role electorally motivated economic manipulation may play. Unlike foreign demand (i.e., exports), which is largely insulated from pre-election manipulation, domestic demand is susceptible to government macroeconomic influence.
48. As Cox hazard models do not assume a functional form hazard, I extract a baseline hazards at each quarter from Model 2.4, plot it with coefficient effects, and assemble them into Figure 1.
49. Δh = [e(0.006tδ)-0.055δ-1]100.
50. The seven additional early elections here are from parliaments that began before the observation window (six) or from one-quarter parliaments censored from the risk set (one).
51. This interpretation of voters' behaviour (but not leaders' inferences) finds support in Harvey D. Palmer and Guy Whitten, 'The Electoral Impact of Unexpected Inflation and Economic Growth', British Journal of Political Science, 29 (1999), 623-39.
52. More specifically, I include all parliaments from the nine European premier-timing states - Austria, Denmark, Greece (post 1974:4), Iceland, Ireland, Italy, the Netherlands, Spain (post1979:1), and the United Kingdom that began after 1 January 1967.
53. Kayser, 'Who Surfs, Who Manipulates?
54. Patrick Brandt and John T.Williams, 'A Linear Poisson Autoregressive Model: The Poisson AR(p) Model', Political Analysis, 9 (2001), 164-84.
55. The Economist, 4 September 2004, 'Australia calls an election: the truth game'.
56. Artis and Zhang, 'International Business Cycles and the ERM: Is There a European Business Cycle?';
57. and Artis and Zhang, 'Further Evidence on International Business Cycles and the ERM'.
58. Denmark, Iceland, Ireland, Italy, Netherlands, United Kingdom. This (including Austria) is the complete set of European states that have had democratic governments for the necessary time-series length and are able to time their elections opportunistically. Thus, Spain and Greece are omitted. The value reported at each year represents the mean of six bivariate correlations between the European countries and Austria's cycle over the current and nineteen preceding years. The correlation mean with the United States includes Austria as well.
59. cf. The Economist, 2 October 2004, 'Economic focus: growing apart'.
60. See Hazard Analysis Variables section for data sources.