Partisanship and the collapse of presidential approval: evidence from Afrobarometer surveys in Ghana

Between 2017 and 2024, presidential approval in Ghana fell from 82 to 33 per cent and the governing New Patriotic Party (NPP) lost office. This article uses the four Afrobarometer rounds spanning that presidency to ask what happened to the governing party’s own identifiers, a question that cross-sectional comparisons of partisan groups cannot answer. Approval fell within every partisan category, and it fell approximately 24 percentage points less among NPP identifiers than among non-identifiers. We term this gap partisan resilience. Decomposition shows that roughly nine tenths of the aggregate decline came from falling approval within categories, so the approving public did not contract into a partisan remnant. The protection itself changed form. Early in the presidency, NPP identifiers converted adverse economic assessments into withdrawn approval far less readily than non-identifiers. By 2022 this difference had disappeared, while the partisan gap in reported economic conditions continued to widen. Among remaining approvers, the share withholding trust in the president rose from 17 to 45 per cent, concentrated among his own identifiers. Partisan protection is several distinct processes that can end at different times, and a partisan gap can outlast whatever produced it.