The role of infrastructural and coercive state capacity in explaining different types of electoral fraud

AbstractWhile effective state capacity can reasonably be considered a necessary condition for democratization, strong states do not automatically produce democratic regimes, nor do they guarantee their survival. Far from being sufficient conditions for democracy, strong or capable states are also thought to be indispensable for the maintenance of autocratic rule. The present article puts to the test the hypothesis that a certain level of state capacity is needed to engage in effective electoral malpractice, using general and more specific indicators of electoral fraud. This article proposes two opposing mechanisms through which state capacity can influence the quality of elections: through infrastructural state capacity and coercive state capacity. The article demonstrates that electoral fraud is more likely in countries where infrastructural state capacity is weak and that coercive state capacity plays a more ambiguous role than previously thought. The analyses also reveal that different factors are at work when looking at precise types of electoral malpractice rather than general measures: voter and candidate intimidation, fraudulent tabulation of votes, unfair media coverage of campaigns and vote buying seem to engage different sets of facilitating structures.Keywords: state capacityelectoral fraudpost-communist countriescoercionmeasurement Notes on contributorJessica Fortin-Rittberger is a senior research associate at GESIS – Leibniz Institute for the Social Sciences (Germany). Her areas of research interest include political developments in former communist countries, political institutions and their measurement, the effects of electoral rules, as well as the impact of state capacity on democratization. Her work has appeared in the European Journal of Political Research, International Political Science Review, and Comparative Political Studies.Notes1. Schedler, 'The Logic of Electoral Authoritarianism'.2. Lehoucq, 'Electoral Fraud'.3. Fortin, 'A Tool to Evaluate State Capacity in Post-Communist Countries, 1989–2006'.4. Kelley, Monitoring Democracy; Kelley and Kolev, 'Election Quality and International Observation 1975–2004.5. Hyde and Marinov, 'Which Elections Can Be Lost?'.6. Birch, 'Project Electoral Malpractice and Electoral Manipulation in New and Semi-Democracies'.7. Migdal, Strong Societies and Weak States.8. Katz, 'Intraparty Preference Voting'; Carey and Shugart, 'Incentives to Cultivate a Personal Vote'; Hicken, 'How Do Rules and Institutions Encourage Vote Buying'; Nyblade and Reed, 'Who Cheats? Who Loots?'.9. Persson and Tabellini, The Economic Effects of Constitutions; Kunicova and Rose Ackerman, 'Electoral Rules and Constitutional Structures as Constraints on Corruption'.10. Carey and Shugart, 'Incentives to Cultivate a Personal Vote'; Chang, 'Electoral Incentives for Political Corruption under Open-List Proportional Representation'; Chang and Golden, 'Electoral Systems, District Magnitude, and Corruption'; Birch, 'Electoral Systems and Electoral Misconduct'.11. Although electoral malpractice can be perpetrated by both challengers and incumbents, this article as well as most existing contributions focus on the incumbent side of the equation, one reason being that most comparative data collection projects either focus on incumbents or do not allow differentiating between perpetrators. However, not all literature is limited to this framework: a recent contribution hypothesized that incumbents and challengers will engage in different types of fraud dependent on their relative strength, see Collier and Vincente, 'Violence, Bribery, and Fraud'.12. Cox and Kousser, 'Turnout and Rural Corruption'; Argesinger, 'New Perspectives on Election Fraud in the Gilded Age'; Lehoucq, 'Electoral Fraud'; Lehoucq and Molina, Stuffing the Ballot Box; Mares and Zhu, 'Economic Concentration, Rural Inequality and Electoral Fraud'.13. Simpser, 'Cheating Big'; Donno and Roussias, 'Does Cheating Pay?'.14. Simpser, 'Cheating Big'.15. Pastor, 'The Role of Electoral Administration in Democratic Transitions'; Mozaffar, 'Patterns of Electoral Governance in Africa's Emerging Democracies'; Hartlyn, McCoy, and Mustillo, 'Electoral Governance Matters'.16. Molina and Lehoucq, 'Political Competition and Electoral Fraud'; Lehoucq, 'Electoral Fraud'; Lehoucq and Molina, Stuffing the Ballot Box; Schaffer and Schedler, 'What Is Vote Buying?'; Schaffer, 'Why Study Vote Buying?'; Wang and Kurzman, 'The Logistics'.17. Way, 'Weak States and Pluralism'; Way, 'Authoritarian State Building and the Sources of Regime Competitiveness in the Fourth Wave'; Way and Levitsky, 'The Dynamics of Autocratic Coercion after the Cold War'.18. Although not all: Sarah Birch proposes comparative empirical verifications in Electoral Malpractice.19. Mann, The Sources of Political Power.20. Migdal, Kohli, and Shue, State Power and Social Forces.21. Spruyt, The Sovereign State and Its Competitors.22. Mann, The Sources of Political Power, 59.23. Mann, 'Infrastructural Power Revisited'.24. Harris, 'Building an Electoral Administration'.25. Ades and Di Tella, 'The New Economics of Corruption'; Damania, Fredriksson, and Muthukumara, 'The Persistence of Corruption and Regulatory Compliance Failures'.26. Warren, 'What Does Corruption Mean in a Democracy'; Warren, 'Political Corruption as Duplicitous Exclusion'.27. Fortin, 'Is There a Necessary Condition for Democracy?'.28. Goodwin-Gill, Free and Fair Elections; Pastor, 'The Role of Electoral Administration in Democratic Transitions'; López-Pintor, Electoral Management Bodies as Institutions of Governance; Birch, 'Electoral Management Bodies and the Electoral Integrity'.29. Molina and Lehoucq, 'Political Competition and Electoral Fraud'; Lehoucq, 'Electoral Fraud'; Lehoucq and Molina, Stuffing the Ballot Box; Schaffer and Schedler, 'What Is Vote Buying?'; Schaffer, 'Why Study Vote Buying?'; Wang and Kurzman, 'The Logistics'.30. Way, 'Weak States and Pluralism'; Way, 'Authoritarian State Building and the Sources of Regime Competitiveness in the Fourth Wave'; Way and Levitsky, 'The Dynamics of Autocratic Coercion after the Cold War'.31. Gandhi and Lust-Okar, 'Elections under Authoritarianism', 406.32. Elections both legislative (lower house) and presidential (when presidents are directly elected) are the units of analysis.33. Schedler, 'Elections without Democracy'; Elklit and Reynolds, 'A Framework for the Systematic Study of Election Quality'; Calingaert, 'Election Rigging and How to Fight It'; Birch, 'Electoral Institutions and Popular Confidence in the Electoral Processes'.34. Ziblatt, 'Shaping Democratic Practice and the Causes of Electoral Fraud'.35. Mann, The Sources of Political Power.36. Fortin, 'A Tool to Evaluate State Capacity in Post-Communist Countries, 1989–2006'.37. Despite being high in validity and reliability, the index of state capacity I employ in the following models suffers from missing data in the early years of transition (with some 5% of missing cases). A portion of the missing cases (for tax revenue and contract intensive money) were imputed. In the case of corruption levels, and the quality of private property protection, a more conservative data interpolation technique was employed where starting values are similar to the first year of actual data coverage by the Heritage Foundation (in most cases around 1995). The final index of state capacity therefore contains both actual measurements and estimated measurements of individual indicators and should therefore be handled and interpreted with a measure of prudence in the early years.38. Wintrobe, The Political Economy of Dictatorship.39. Ibid., 34.40. Cingranelli and Richards, 'The Cingranelli-Richards (CIRI) Human Rights Dataset'.41. Singer, Bremer, and Stuckey, 'Capability Distribution, Uncertainty, and Major Power War, 1820–1965'; Singer and Small, 'National Military Capabilities Data. Correlates of War Project'. As Albertus and Menaldo noted in 'Coercive Capacity and the Prospects for Democratization', the variable is right-hand skewed and they chose to correct the distribution through a natural logarithm power transformation. In the present case, analyses show that a logarithm transformation is less efficient than leaving the variable in its original scale.42. Albertus and Menaldo, 'Coercive Capacity and the Prospects for Democratization', 153.43. Gurr, 'War, Revolution, and the Growth of the Coercive State'.44. Readers should note that CIRI's Physical Integrity Rights Index is only weakly associated with the size of military forces per 100 inhabitants. Pearson's r = −0.14 with 209 valid cases.45. Anand, 'Procedural Fairness in Economic and Social Choice'; Brusco, Nazareno, and Stokes, 'Selective Incentives and Electoral Mobilization'.46. Ramseyer and Rosenbluth, Japan's Political Marketplace; Nielson and Shugart, 'Constitutional Change in Colombia'.47. Ziblatt, 'Shaping Democratic Practice and the Causes of Electoral Fraud'. In Birch (Electoral Malpractice) these indicators are associated with patronage relations.48. Schedler, 'The Logic of Electoral Authoritarianism', 1–2.49. Unfortunately this indicator contains few overlapping observations with other variables and could not be included in all models since it significantly reduced the already limited amount of cases.50. Ross, 'Does Oil Hinder Democracy?'.51. A variable containing a lagged value of electoral fraud, as well as a cruder measure capturing previous instance of electoral fraud were considered, but not retained in the final models for three reasons. First, neither yielded statistically significant results. Second, their inclusion did not affect other parameter estimates' size, and last, these variables removed all countries' theoretically important first post-communist elections from the sample.52. Marshall and Jaggers, 'Polity IV Project'.53. Lewis, 'Non-Competitive Elections and Regime Change'.54. Fuller, 'Armenia'.55. The small sample and spread out categories make the estimates fragile. A series of Brant tests (Long and Freese, Regression Models for Categorical Dependent Variables Using Stata), indicated that the parallel lines assumptions underlying ordered logistic regressions were violated. Additional testing reveals that the cut points separating the values of the dependent variables are statistically significantly different from each other in the three models contained in Table 1. The use of clustered standard errors is set to eliminate intra-class correlations.56. Readers should note that while multicollinearity is present in all three models, no variance inflation factor values exceed 2.82 for any variables included in the three models presented in Table 1. Link tests were performed to assess the amount of potential bias from specification errors. The tests were not statistically significant for Models 2 and 3, indicating that they do not suffer from omitted variable bias.57. Ziblatt, 'Shaping Democratic Practice and the Causes of Electoral Fraud'; Birch, Electoral Malpractice.58. Because of the strong correlation between electoral rules and regime type, I was not able to include a regime type variable in these analyses. However, one could hypothesize that the effects of regime types are encapsulated in both electoral rules and election for the most powerful position.59. Lindberg, Democracy and Elections in Africa.60. All dependent variables from IEMD and QED were recorded in ordinal levels of measurement. When ordered logistic regressions were performed, the small amount of cases made it impossible to perform Brant tests, and yielded insignificant thresholds (cut points used to differentiate the adjacent levels of the response variable). These items were therefore transformed into binary variables to allow estimation.61. GDP per capita was dropped in some models due to high multicollinearity with infrastructural state capacity which made both variables impossible to estimate together. In the models where GDP per capita was omitted, readers must be aware that the variable only achieves significance once infrastructural state capacity was pulled from the analyses. Both variables therefore share a large variance that makes it difficult to estimate their independent effects in very small samples. A similar scenario unfolded with the indicator measuring executives with unlimited powers which was dropped from models with smaller sample sizes.62. International Institute for Democracy and Electoral Assistance (IDEA), 'Funding of Political Parties and Election Campaigns'. IDEA's data coverage is limited and excludes 11 out of our 26 countries. As well, it should be noted that the presence of ceilings on political parties' expenditures does not achieve statistical significance in any other model presented in Table 2.63. Wang and Kurzman, 'The Logistics'.64. For example Mexico, see Rigger, 'Grassroots Electoral Organization and Political Reform in the ROC on Taiwan and Mexico'.65. Pastor, 'The Role of Electoral Administration in Democratic Transitions'.66. Albertus and Menaldo, 'Coercive Capacity and the Prospects for Democratization'.