Published online by Cambridge University Press: 03 July 2012
What explains local variation in electoral manipulation in countries with ongoing internal conflict? The theory of election fraud developed in this article relies on the candidates’ loyalty networks as the agents manipulating the electoral process. It predicts (i) that the relationship between violence and fraud follows an inverted U-shape and (ii) that loyalty networks of both incumbent and challenger react differently to the security situation on the ground. Disaggregated violence and election results data from the 2009 Afghanistan presidential election provide empirical results consistent with this theory. Fraud is measured both by a forensic measure, and by using results from a visual inspection of a random sample of the ballot boxes. The results align with the two predicted relationships, and are robust to other violence and fraud measures.
Centre for the Study of Civil War, Peace Research Institute Oslo (email: nils.weidmann@gmail.com); and Department of Economics, University of California, San Diego (email: mjcallen@ucsd.edu), respectively. The authors thank Eli Berman, Tiffany Chou, Rex Douglass, Hanne Fjelde, Marjorie Flavin, Susan Hyde, Radha Iyengar, Patrick Kuhn, Jason Lyall, Craig McIntosh, Kyle Pizzey, Gerald Schneider, Jacob Shapiro, Jessica Trisko, Choon Wang, Hal White and their colleagues at UCSD and Princeton for helpful comments and assistance with data. They also thank Israel Malkin for excellent research assistance. This material is based upon work supported by the Air Force Office of Scientific Research (AFOSR) under Award No. FA9550-09-1-0314 and by the European Commission under a Marie Curie Fellowship. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of any of the funders. Replication data will be made available at http://dvn.iq.harvard.edu/dvn/dv/nilsw upon publication of this article. An appendix containing additional information is available at: http://dx.doi.org/10.1017/S0007123412000191
1 Hyde, Susan D. and Marinov, Nikolay, ‘Which Elections Can Be Lost?’ Political Analysis, 20 (2012), 191–210 Google Scholar
Nils Petter Gleditschet al., ‘Armed Conflict 1946–2001: A New Dataset’, Journal of Peace Research, 39 (2002), pp. 615–637 CrossRefGoogle Scholar
2 Berman, Eli, Shapiro, Jacob N. and Felter, Joseph H., ‘Can Hearts and Minds Be Bought? The Economics of Counterinsurgency in Iraq’, Journal of Political Economy, 119 (2011), 766–819 Google Scholar
Kilcullen, David, Counterinsurgency (New York: Oxford University Press, 2010)Google Scholar
3 Collier, Paul and Rohner, Dominic, ‘Democracy, Development, and Conflict’, Journal of the European Economic Association, 6 (2008), 531–540 Google Scholar
4 Authors’ calculations based on the National Elections across Democracy and Autocracy Dataset (Hyde and Marinov, ‘Which Elections Can Be Lost?’) and the Armed Conflict Dataset (Gleditsch et al., ‘Armed Conflict 1946–2001’).
5 Collier, Paul and Vicente, Pedro C., ‘Votes and Violence: Evidence from a Field Experiment in Nigeria’ (Working paper, Department of Economics, Oxford University, 2011)Google Scholar
Ellman, Matthew and Wantchekon, Leonard, ‘Electoral Competition under the Threat of Political Unrest’, Quarterly Journal of Economics, 115 (2000), 499–531 Google Scholar
Wilkinson, Steven, Votes and Violence (Cambridge: Cambridge University Press, 2004)Google Scholar
6 Dahl, Robert, Polyarchy: Participation and Opposition (New Haven, Conn.: Yale University Press, 1971)Google Scholar
7 R. Michael Alvarez, Thad E. Hall and Susan D. Hyde, Election Fraud: Detecting and Deterring Electoral Manipulation, eds (Washington, D.C.: Brookings Institution Press, 2008), pp. 21–36 Google Scholar
8 Alvarez, Hall and Hyde, eds, Election Fraud: Detecting and Deterring Electoral Manipulation, pp. 201–215Google Scholar
9 Ichino, Nahomi and Schündeln, Matthias, ‘Deterring or Displacing Electoral Irregularities? Spillover Effects of Observers in a Randomized Field Experiment in Ghana’, Journal of Politics, 74 (2012), 292–307 Google Scholar
10 For more general treatments of what strategies are chosen under which conditions, see e.g. Collier and Vicente, ‘Votes and Violence’.
11 Birch, Sarah, ‘Electoral Systems and Electoral Misconduct’, Comparative Political Studies 40 (2007), 1533–1556 Google Scholar
Fred Schaffer, ed., Elections for Sale (Boulder, Colo.: Lynne Rienner, 2007)Google Scholar
12 Alvarez, Hall and Hyde, eds, Election Fraud, pp. 99–111Google Scholar
Lehoucq, Fabrice Edouard and Jiménez, Iván Molina, ‘Political Competition and Electoral Fraud: A Latin American Case Study’, Journal of Interdisciplinary History, 30 (1999), 199–234 Google Scholar
Lehoucq, Fabrice Edouard, ‘Electoral Fraud: Causes, Types and Consequences’, Annual Review of Political Science, 6 (2003), 233–256 CrossRefGoogle Scholar
13 Ziblatt, Daniel, ‘Shaping Democratic Practice and the Causes of Electoral Fraud: The Case of Nineteenth-Century Germany’, American Political Science Review, 103 (2009), 1–21 CrossRefGoogle Scholar
14 Collier and Vicente, ‘Violence, Fraud, and Bribery’.
15 Hyde, Susan D., ‘The Observer Effect in International Politics: Evidence from a Natural Experiment’, World Politics, 60 (2007), 37–63 Google Scholar
16 Brancati, Dawn, ‘Democracy Promotion: Perceptions versus Reality’ (Working paper, Department of Political Science, Washington University in St. Louis, 2010)Google Scholar
17 UNDP ELECT Afghanistan. Annual Progress Report – 2009. Available at http://www.undp.org.af/Projects/2009AnnualReports/ELECT_APR09.pdf.2009.
18 We provide some evidence against the second explanation below by showing that our results are robust to the inclusion of the share of planned polling stations that actually operated into the regression.
19 Collier and Vicente, ‘Violence, Fraud, and Bribery’; Collier and Vicente, ‘Votes and Violence’; Ellman and Wantchekon, ‘Electoral Competition under the Threat of Political Unrest’; Wilkinson, Votes and Violence.
20 Gandhi, Jennifer and Lust-Okar, Ellen, ‘Elections under Authoritarianism’, Annual Review of Political Science, 12 (2009), 403–422 Google Scholar
21 See also Ichino and Schündeln, ‘Deterring or Displacing Electoral Irregularities?’
22 For more information about the electoral process, see National Democratic Institute, The 2009 Presidential and Provincial Council Elections in Afghanistan (technical report available at http://www.ndi.org/files/Elections_in_Afghanistan_2009.pdf.2010).
23 National Democratic Institute, The 2009 Presidential and Provincial Council Elections in Afghanistan.
24 See e.g. ‘Afghanistan imposes censorship on election day’, New York Times, 18 August 2009. http://www.nytimes.com/2009/08/19/world/asia/19afghan.html.
25 Bill Greer, Dataset of the Day: Impact of Violence on the Afghanistan Elections (online report, 26 August 2009. Available at http://blog.geoiq.com/2009/08/26/dataset-of-the-day-impact-of-violence-on-the-afghanistan-elections/.2009).
26 Callen, Michael and Long, James D., ‘Institutional Corruption and Election Fraud: Evidence from a Field Experiment in Afghanistan’ (Working paper, Department of Economics, University of California, San Diego, 2011)Google Scholar
27 Sharan, Timor, ‘The Dynamics of Elite Networks and Patron-Client Relations in Afghanistan’, Europe-Asia Studies, 63 (2011), 1109–1127 Google Scholar
28 Chandra, Kanchan, Why Ethnic Parties Succeed: Patronage and Ethnic Head Counts in India (Cambridge: Cambridge University Press, 2004), p. 6Google Scholar
29 Sharan, ‘The Dynamics of Elite Networks and Patron–Client Relations in Afghanistan’.
30 Sharan, ‘The Dynamics of Elite Networks and Patron–Client Relations in Afghanistan’, p. 1122Google Scholar
31 Lehoucq, ‘Electoral Fraud’.
32 Ziblatt, ‘Shaping Democratic Practice and the Causes of Electoral Fraud’.
33 Lehoucq, ‘Electoral Fraud’, p. 234Google Scholar
34 Myagkov, Mikhail, Ordeshook, Peter C. and Shakin, Peter, The Forensics of Election Fraud: Russia and Ukraine (New York: Cambridge University Press, 2009)Google Scholar
35 Alvarez, Hall and Hyde, eds, Election Fraud, pp. 162–181Google Scholar
Jr, Walter R. Mebane, ‘Election Forensics: Vote Counts and Benford's Law’ (Working paper, Department of Political Science, University of Michigan, 2006)Google Scholar
36 Deckert, Joseph, Myagkov, Mikhail and Ordeshook, Peter C., ‘Benford's Law and the Detection of Election Fraud’, Political Analysis, 19 (2011), 245–268 Google Scholar
37 Beber, Bernd and Scacco, Alexandra, ‘What the Numbers Say: A Digit-Based Test for Election Fraud’, Political Analysis, 20 (2012), 211–234 Google Scholar
38 Beber and Scacco (‘What the Numbers Say’) propose additional forensic tests. However, we find that they correlate less well with our recount-based fraud measure (see next section), so we do not use it in the core empirical analysis.
39 Beber and Scacco, ‘What the Numbers Say’.
40 Mebane, ‘Election Forensics’.
41 Lacking information about the precise location of polling stations, we are unable to estimate fraud at lower levels, for example, cities. Moreover, many of the covariates used in our regression analysis below are available only at the district level.
42 Mebane, Walter R., ‘Comment on Benford's Law and the Detection of Election Fraud’, Political Analysis, 19 (2011), 269–272 Google Scholar
43 Beber and Scacco, ‘What the Numbers Say’; Deckert, Myagkov and Ordeshook, ‘Benford's Law and the Detection of Election Fraud’.
44 The stations were divided into five categories, as follows: (A1) 600 or more valid votes were cast, (A2) 600 or more total votes were cast, (B1) more than 100 votes were cast and one candidate received 95 per cent or more of the total votes, (B2) one candidate received 95 per cent or more of the total valid votes cast, and (C1) 600 or more valid votes were cast and one candidate received 95 per cent or more of the total votes.
45 The IEC reports investigating 343 boxes, but categories B1 and C1 described above contain a mutual observation which was overlooked.
46 In the recount, the shares of boxes exhibiting some physical evidence of fraud are as follows. A1: 0.655 (sd = 0.480); A2: 0.7 (sd = 0.466); B1: 0.721 (sd = 0.450); B2: 0.462 (sd = 0.505); C1: 0.939 (sd = 0.241). The sample of 342 stations was drawn randomly and so by the weak law of large numbers the sample mean for each category provides a consistent estimate of the share of fraudulent polling stations in that category at the national level.
47 The IEC publicly posted the data in three waves. They reported returns from 27,163 distinct polling stations on 19 September, 23,300 stations on 10 October and 22,853 polling stations on 20 October. Of the original 27,163, 4,305 are missing or record no votes. This brings us to 22,858 (almost the 20 October number). We use the earliest data release in order to be able to control for the number of missing/closed polling stations.
48 Clionadh Raleigh et al. ‘Introducing ACLED: An Armed Conflict Location and Event Dataset’, Journal of Peace Research, 47 (2010), 651–660 Google Scholar
49 In the case of the WITS dataset, the location of an event is provided by the town or village it occurred in. We added geographic co-ordinates to events by looking up these place names in a gazetteer.
50 US Geological Survey. GTOPO30 Digital Elevation Model. Available at http://edc.usgs.gov/products/elevation/gtopo30/gtopo30.html.2007.
51 Oak Ridge National Laboratory, LandScan Global Population database, 2008.
52 Our results are additionally robust to clustering at the province level; there are a sufficient number of provinces that clustered standard errors are likely to resemble their asymptotic distribution.
53 For examples, see Ellman and Wantchekon, ‘Electoral Competition under the Threat of Political Unrest’; Wilkinson, Votes and Violence.
54 C. Christine Fair and Seth G. Jones, ‘Securing Afghanistan: Getting on Track’ (Working paper, United States Institute of Peace, available at http://library.usip.org/articles/1012068.1022/1.PDF,2009).
55 Collier, Paul, Wars, Guns and Votes: Democracy in Dangerous Places (New York: HarperCollins, 2009)Google Scholar