How do parties update candidate lists after electoral setbacks, and what does this mean for women’s representation? We exploit Turkey’s 2015 back-to-back parliamentary elections as quasi-experimental leverage and implement a difference-in-differences design that compares the governing Justice and Development Party (JDP) to the Republican People’s Party (RPP), whose March 2015 primaries largely fixed the district-level gender composition of slates. Falling short of a single-party majority in June was followed by a roughly 40 % contraction in the JDP’s women candidates and a disproportionate downgrading at electable ranks, interrupting a decade-long upward trend. The contraction is concentrated in conservative strongholds. A rank-weighted decomposition shows that net removals, rather than simple demotions, account for most of the decline; changes outside electable ranks are smaller and imprecisely estimated. Event-time estimates indicate the shock produced a one-off adjustment that reverted by 2018. Taken together, the evidence is most consistent with a mix of statistical discrimination where seats are at stake and attribution bias that overshoots, illustrating how elite responses under compressed timelines can quickly erode representational gains in closed-list systems.