Table 2

Effect of the old-age pension programme on healthcare utilisation, health expenditures and insurance uptake

	(1)	(2)	(3)	(4)	(5)
Panel A: healthcare utilisation
	Formal care	Informal care	Hospital stays	Medication	Any access
Treatment*Post	0.090 (0.015 to 0.164)	0.009 (−0.032 to 0.050)	0.014 (−0.019 to 0.047)	0.077 (0.020 to 0.134)	0.068 (−0.006 to 0.143)
Post	0.042 (−0.261 to 0.345)	0.016 (−0.188 to 0.220)	−0.078 (−0.220 to 0.065)	−0.146 (−0.392 to 0.100)	0.009 (−0.300 to 0.319)
Observations	3818	3818	3817	3815	3818
Individuals	2397	2397	2396	2396	2397
Panel B: healthcare expenditures (log)
	Formal care	Informal care	Hospital stays	Total OOP
Treatment*Post	−0.120 (−0.311 to 0.071)	0.024 (−0.040 to 0.089)	−0.013 (−0.109 to 0.082)	−0.100 (−0.305 to 0.105)
Wave 2	0.841 (0.179 to 1.503)	0.227 (−0.248 to 0.702)	−0.145 (−0.396 to 0.106)	0.975 (0.181 to 1.768)
Observations	3818	3818	3818	3818
Individuals	2397	2397	2397	2397
Panel C: health insurance (HI) uptake
	Social security	Private HI	Public HI	Any HI
Treatment*Post	−0.006 (−0.032 to 0.020)	−0.000 (−0.008 to 0.007)	0.042 (0.015 to 0.069)	0.031 (−0.002 to 0.063)
Wave 2	−0.021 (−0.089 to 0.046)	−0.008 (−0.028 to 0.011)	−0.011 (−0.077 to 0.055)	−0.048 (−0.136 to 0.041)
Observations	3751	3751	3751	3751
Individuals	2367	2367	2367	2367

Confidence intervals were obtained using bootstrapped SEs. The coefficients are estimated with an individual fixed-effects difference-in-difference models controlling for changes in age fixed effects, education years, marital status, household size, labour force status, chronic conditions and number of limitations in daily living activities. Valladolid is the treatment village where the old-age pension programme was implemented in December 2008. The models use data from the baseline and follow-up survey.
OOP, out of pocket.