Churn Without Fragmentation: How a Party-Label Bug Reversed My Headline Finding

An analysis of English local election data from 2018-2022 initially indicated a massive increase in party system fragmentation, but this finding was reversed after fixing a critical data bug. The error stemmed from treating different ballot labels for the same political party as separate analytical groups, which artificially inflated fragmentation metrics. After properly normalizing party labels into consistent families, the corrected analysis showed that while voter volatility nearly doubled, the party system did not fragment. Instead, the vote share largely consolidated, shifting from the Conservative party to the Labour party, with geographically concentrated gains for third parties.