Southern white partisan dealignment persists across many robustness specifications

Evidence

Notes: The twelve coefficients above are each from a separate regression. The first panel take the form Demits = βSouth × After + γXits + µs + ηt + eits and the second panel take the form Demits = βSouth × Y ear + γXits + µs + ηt + eits, where South × After is an indicator variable for being surveyed after April 1963 and residing in one of the eleven Southern states at the time of the survey, South·Y ear is a South-specific linear time trend, µs are state fixed effects, ηt are year fixed effects, and Xits is a vector of controls that we vary to test robustness. Income (five categories) and city-type (three categories) fixed effects are included where specified in the table footer. Where “Interactions” are specified, income and city-type fixed effects have each been interacted with South and (separately) with After (when South × After is the explanatory variable of interest) or Y ear (when South × Y ear is the explanatory variable of interest). The “Restricted” sample used in columns 5 and 6 excludes those younger than 21 years in 1963 and current Southern residents who were not born in the South. (These specifications, therefore, exclude those with missing values for place of birth). Standard errors clustered by state. — Ilyana Kuziemko and Ebonya Washington, Why Did the Democrats Lose the South?…, 1130

Categorization

• XREF: For the threatened-privilege research program this documents the robustness battery applied to the Southern realignment case — a template for checking whether mobility effects survive controls for income, urbanicity, and migration. — connects to: Racial views are analyzed as the central explanatory variable for dealignment — Ilyana Kuziemko and Ebonya Washington, Why Did the Democrats Lose the South?…, 74; Economic development cannot explain Southern Democratic dealignment — Ilyana Kuziemko and Ebonya Washington, Why Did the Democrats Lose the South?…, 798
• QUESTION: Do coefficients attenuate meaningfully from Panel A col 1 (-0.142) to the restricted, interacted specifications (-0.155), and does the stability itself matter for distinguishing realignment from compositional change?

Broader Connections

• Racial views are analyzed as the central explanatory variable for dealignment #responds-to (0.85) — Ilyana Kuziemko and Ebonya Washington, Why Did the Democrats Lose the South?…, 74
• Economic development cannot explain Southern Democratic dealignment #corroborates (0.75) — Ilyana Kuziemko and Ebonya Washington, Why Did the Democrats Lose the South?…, 798