Table 2.

Impact of female leadership on health investment and non-pharmaceutical interventions, RDD estimates.

	Panel A: Health investment
	Pre-outbreak			Post-outbreak
	$\Delta$ Health spending	$\Delta$ Total hosp. beds	$\Delta$ ICU hosp. beds	$\Delta$ Health spending	$\Delta$ Total hosp. beds	$\Delta$ ICU hosp. beds
	(share of total spd.)	per 100k pop.	per 100k pop.	(share of total spd.)	per 100k pop.	per 100k pop.
RD estimator	−0.005	−9.992	−0.372	0.004	−3.212	−0.084
Robust $p$-value	0.678	0.281	0.235	0.749	0.519	0.89
Robust conf. int.	[−0.0273, 0.0178]	[−28.1529, 8.168]	[−0.9858, 0.2418]	[−0.0207, 0.0288]	[−12.9813, 6.5565]	[−1.2766, 1.1078]
CCT-optimal BW	15.95	19.075	8.094	22.881	12.062	11.66
Eff. number obs.	745	871	472	753	626	613
	Panel B: Non-pharmaceutical interventions
	Number of	Face covering	Gatherings	Cordon	Closure of	Public transport
	NPIs	required	prohibition	sanitaire	non-essentials	restriction
RD estimator	0.371	0.08	0.055	0.14	−0.071	0.131
Robust $p$-value	0.057*	0.04**	0.066*	0.083*	0.414	0.221
Robust conf. int.	[−0.0109, 0.7538]	[0.0038, 0.1569]	[−0.0037, 0.1143]	[−0.0184, 0.298]	[−0.2396, 0.0986]	[−0.0783, 0.3395]
CCT-optimal BW	10.254	15.671	9.479	12.224	10.053	10.784
Eff. number obs.	353	499	341	395	349	361

Notes: This table reports our RD estimates of the association between female mayors and several outcomes. The level of observation is the municipality. Panel A reports results on health investment-related outcomes. This panel is divided into pre and post pandemic outbreak outcomes. In the first column of Panel A, the outcome is the variation in the share of municipal spending dedicated to health issues between 2016 and 2019. In the second column, the outcome is the variation of total hospital beds per 100k inhabitants between Jan 2017 and Jan 2020; in the third column, the ICU hospital beds per 100k inhabitants variation between Jan 2017 and Jan 2020. The fourth column reports the estimate of the variation in the share of municipal spending dedicated to health issues between 2019 and 2020. Lastly, the fifth and sixth columns show estimates for the variation of hospital beds per 100k inhabitants between Feb 2020 and Dec 2020 — total beds and ICU beds, respectively. Panel B describes results for the main non-pharmaceutical interventions adopted by mayors until July 2020. The first column outcome is the total number of NPIs adopted. The remaining columns are dummies variables indicating whether a specific NPI was adopted. In any case, we are estimating a first-degree polynomial using a uniform kernel. Optimal bandwidths following Calonico et al. (2014) were chosen to minimize the mean squared error of the local polynomial RD point estimator. Following that same work, we report robust-bias corrected $p$-values and 95% CIs. All estimates account for state fixed-effects following Eq. (1). Coefficients significantly different from zero at 99% (***), 95% (**) and 90% (*) confidence level.