Municipal rental-housing inspection regimes and informal landlord repair investment in Nairobi wards: a difference-in-differences estimation of enforcement-frequency thresholds

Municipal inspection regimes in Nairobi's informal rental market are frequently justified as a means of raising landlord repair investment, yet the enforcement-frequency threshold at which such investment responds has not been formally characterised. The analysis estimates power, minimum detectable effect at 80% power, and the sampling distribution of the BACI interaction estimator. Power reaches conventional levels only at 52 clusters for a 5-percentage-point effect; at 32 clusters the minimum detectable effect at 80% power remains above the 5-point threshold. The results imply that ward-level BACI designs in Nairobi cannot credibly identify enforcement-frequency thresholds below roughly 5 percentage points without cluster counts exceeding those typically available to municipal evaluators, and that the estimand itself is sensitive to the choice of comparison wards.