Audit Results Swing South African By-Elections About Five Percentage Points

🧾 What Was Examined

The connection between information and retrospective accountability is tested in South African municipal politics. Voters are expected to reward improvements and punish poor governance, but prior evidence is mixed. The paper evaluates whether publicly available audit reports change vote shares in local by-elections.

🔎 How Timing Is Used to Isolate Audit Effects

A novel identification strategy leverages the predictable timing of annual audit announcements and a specific class of contests: by-elections triggered by the deaths of sitting local councilors. By comparing by-elections that occur shortly before versus shortly after each year’s audit results are released, the approach isolates the effect of timely audit information on electoral outcomes.

📊 Data Sources and Main Findings

• National set of municipal by-elections in South Africa, focusing on contests caused by councilor deaths.
• Comparison hinges on the annual public announcement schedule of municipal audit results.
• Main result: timely audit information changes the vote share of the party responsible for local governance.
• Magnitude: voters reward improved audits and punish poor audits by about 5 percentage points in vote share.
• Additional individual-level survey evidence corroborates that voters update evaluations and vote intentions in response to audit information.
• The analysis covers universal spatial and temporal variation and uses naturally public, organic dissemination of audit reports.

⚖️ Why It Matters

The findings show that routine audit disclosures can produce meaningful electoral accountability at the municipal level. When audit information is available in time to affect electoral choices, parties are rewarded or penalized in measurable ways, demonstrating a practical pathway through which public audits influence political behavior.