Regional Correction of The Flis-Słomczyński-Stolicki Formula: The Case of Turkish Elections

This paper proposes a correction to the Flis-Słomczyński-Stolicki (2019, 2020) formula for countries with large variation among their districts in terms of political divisions. The Flis-Słomczyński-Stolicki formula (FSS formula) estimates seat allocations under the Jefferson-D’Hondt method by using national vote shares, as well as other parameters that are often readily available. However, the FSS formula does not yield precise estimates in those countries where there are independent candidates, special rights assigned to minority parties, signifi cant variation in district sizes, or an unequal distribution of votes due to ethnic or other regional divisions. Hence, I propose dividing the national distribution of votes into regions that satisfy the assumptions of the FSS formula within their district borders. By applying the FSS formula to regions consisting of historically and politically homogenous districts, I demonstrate that the formula’s estimates become signifi cantly more precise. For instance, by applying the regional correction to the 2018 Turkish Parliamentary elections, as well as other Turkish elections between 2007 and 2015, I show that the formula with the correction in three separate regions improves the Loosemore-Hanby goodness of fi t estimates from 2.1 to 3.41 percentage points (95% CI). Thus, the correction might signifi cantly improve the estimates of the FSS formula in various countries, including Spain, Peru, and Belgium.