Table 3.

Cross-section and panel logit modes. Estimation of income and wealth semi-elasticities of VPHI

Wave	N° of observations	Income	95% CI	Wealth	95% CI
2008	6197	0.064a	(0.023;0.104)	0.015a	(0.006;0.025)
2011	6084	0.080a	(0.050;0.109)	0.008b	(0.001;0.014)
2014	6116	0.116a	(0.094;0.139)	0.003	(−0.002; 0.009)
Panel	1928	0.076c	(−0.014; 0.168)	0.048c	(−0.009; 0.104)

The values in the table report the absolute change in the probability of having voluntary private health insurance if the income or wealth increases by 1%. All the models adjust by age, sex and education level of the head of the family, expected future income, number of children under 14 years of age in the household, and proportion of people in the household with bad or very bad health. The income and wealth of the household are in logarithms and adjusted by family composition according to the OECD scale. For the conditional fixed effects logit model (panel data) the estimates are average (semi) elasticities of P(Y_iy = 1| X_it, α_i), calculated following Kitazawa [22]. The last row contains the estimators of the panel model with household fixed effect

^asignificant at 1%

^bsignificant at 5%

^csignificant at 10%