Obtaining Electoral Fraud Measures through Court Election Petitions

• In Event: The Politics of Judicial Decision-making in the Developing World

Abstract

Electoral fraud is a tricky thing to study. It is often unobserved and is unobserved on purpose. Scholars of comparative politics have made great strides in analyzing fraud, both through deep case study work and though cross-national coding of fraud incidence. However, most of this work has either been based on reports by election observers (Simpser, 2013) or relied on statistical analysis designed to detect irregularities in the results (Beber and Scacco, 2012; Mebane, 2017).
In this paper, we join previous scholarship in arguing that electoral court cases (or petitions) in African elections can provide a valuable source of information about the breadth and depth of electoral malpractice and discuss some of the kinds of information we can glean from studying these cases as well as their limitations. Analyzing election court cases is particularly important in understanding electoral malpractice given the sheer number of electoral cases being brought before African courts. Preliminary evidence points to huge growth in the number of these cases. Across a variety of countries, elections generate hundreds if not thousands (as in the case of Nigeria 2007) of election-related court petitions. These cases, we suggest, can yield much more micro data on both the type and the geographic distribution of fraud allegations than most previous analyses.
To evaluate the use of court cases to measure fraud, we construct a database of every election petition brought before Kenyan courts questioning the validity of an election outcome at either the constituency or county levels in the aftermath of the 2013 Kenyan general elections with linked metadata at the candidate and polling station level. We manually code the text associated with the original High Court decision of each case. Our analysis yields 134 different codes or specific types of electoral fraud and administrative irregularities. We then aggregate our highly specific codes into broader categories of fraud and administrative irregularities, including generalized coding schemes that already exist in the literature. We discuss the descriptive results of this exercise and compare them to other measures of election fraud present in the Kenyan context. We conclude that electoral court cases provide an important new avenue for research on fraud and electoral integrity more generally and consider further steps for building on our work.

Authors

• Aaron S. Erlich, McGill University

• Nicholas Kerr, University of Florida