Using Vote Counts' Digits to Diagnose Strategies and Frauds: Russia
61 Pages Posted: 12 Aug 2013
Date Written: 2013
Tests of the digits of vote counts have been proposed to diagnose election fraud. Both the second-digit Benford's-like Law (2BL) and the idea that the last digits should be uniformly distributed have been proposed as standards for clean elections. Many claim that election fraud is rampant in recent Russian federal elections (since 2004), so Russia should be a good setting in which to see whether the digit tests add any diagnostic power. Using precinct-level data from Russia, I first use a randomization test to identify sets of precincts (called UIKs in Russia) in which vote counts for candidates are augmented compared to vote counts in a comparison sets of UIKs. These are a subset of UIKs in which turnout percentages or the percentage of votes for Putin (or United Russia) are divisible by five. Then I run tests of the second and last digits of the UIK vote counts both for the entire set of UIKs in an election year and separately for various sets of UIKs. The digit tests produce surprising and on balance implausible results. For example, they suggest that none of the votes for Putin in 2004 and 2012 or for United Russia in 2011 were fraudulent, while votes for Medvedev in 2008 were fraudulent. The usefulness of simple and direct application of either kind of digit tests for fraud detection seems questionable, although in connection with more nuanced interpretations they may be useful.
Keywords: election forensics, digit tests, randomization inference, Russian elections
JEL Classification: C00
Suggested Citation: Suggested Citation