Municipal elections in the Czech Republic are problematic from the standpoint of invalid votes and the suspicion of electoral frauds. Government agencies do not make public any information about the quantity of invalid votes. There exist strong indications pointing to the unwarranted invalidation of votes during the counting of votes.
The paper presents a method to identify municipalities and districts in which strong anomalies in the quantity of invalid votes occurs. The probable proportion of invalid votes are estimated using OLS regression (about 67,000 districts, elections from 2002 to 2014). On the basis of the residuals from the model, we can identify the problematic electoral districts where there are an unaccountably high proportion of invalid votes (up to 60%). Through other methods we can deduce whether there exists a standard explanation or if there is an error in the counting, or if there could be suspicion of electoral fraud.
