...and the effect seems to persist the rest of the season

Actual versus expected (based on Elo ratings) winning percentage for teams since 1996* who changed managers before their 100th game of the season, over the rest of the season -- following the month immediately after the change

* Excluding 2020 (pandemic) and 2026 (incomplete) seasons. Also, data only includes situations with exactly two managers for the team during the season, for ease of comparison.
Chart: Neil PaineSource: Neil's Substack, Lahman DBGet the dataCreated with Datawrapper