Short and longer-term impacts of health insurance on catastrophic health expenditures in Kwara State, Nigeria

Table 4 Impact of health insurance on CHE for full sample, by wealth terciles and by chronic illness status [(Average treatment effect on the treated estimates 2011 and 2013)]

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	Full sample -no controls	Full sample-controls	Poor	Middle	Rich	Chronic	Non- Chronic
Panel A: CHE (> 10% of household consumption)
ATET 2011	-0.021	-0.057**	-0.062**	-0.047	-0.049	-0.076**	-0.035
	(0.014)	(0.025)	(0.030)	(0.035)	(0.052)	(0.036)	(0.028)
N	667	667	165	217	273	122	521
Adj. R-sq	0.001	0.014	-0.035	0.065	0.0009	0	0.019
ATET 2013	-0.009	-0.019	0.008	-0.054**	-0.034	-0.02	0.009
	(0.018)	(0.019)	(0.027)	(0.024)	(0.041)	(0.045)	(0.026)
N	709	709	197	218	282	132	553
Adj. R-sq	-0.001	0.019	0.019	0.081	0.177	0.014	0.089
Panel B: CHE (> = 40% of capacity to pay)
ATET 2011	-0.006	-0.026	-0.036	-0.004	-0.024	-0.071	0.003
	(0.011	(0.018)	(0.038)	(0.016)	(0.032)	(0.048)	(0.017)
N	667	667	165	217	273	122	521
Adj. R-sq	0.001	0.018	-0.012	0.068	0.054	0.023	0.002
ATET 2013	-0.001	-0.027*	-0.065	-0.056**	0.025	0.029	-0.037**
	(0.014)	(0.016)	(0.046)	(0.025)	(0.029)	(0.035)	(0.015)
N	709	709	197	218	282	132	550
Adj. R-sq	0.001	0.026	0.019	0.127	0.244	0.03	0.025

Standard errors clustered at EA level (in parentheses); * p < 0.10; ** p < 0.05; *** p < 0.01
Based on the balanced panel of 1,039 households. Covariates included in regressions are baseline values for gender, marital status and education of household head, household size, urban, distance to program health facity, household consumption excluding health expenditure, quality of toilet and water, household health expenditure. Propensity score weighting applied as inverse probability of treatment weights. Wealth index: These are calculated as baseline wealth terciles and were constructed based on the first loading of a principal component analysis of 30 dwelling characteristics and asset ownership indicators measured at baseline

ISSN: 1472-6963