Analysis of Economic Effect and Mechanism of Basic Medical Insurance for Urban Employees in China

Table 6

Release effect of normal medical needs.

	Variable	Low-income families		Middle-income families		High-income families
		(1)	(2)	(3)	(4)	(5)	(6)
		Selection equation	Expenditure equation	Selection equation	Expenditure equation	Selection equation	Expenditure equation
Explained variable: logarithm of total medical expenses	Interaction effect	−0.0398	0.00611	−0.0337	0.763	0.00776	−0.0949
	(Treat × time)	[0.0544]	[2.6589]	[0.0312]	[1.4655]	[0.0369]	[1.9022]
	Treatment effect	0.0254	−0.114	0.0157	0.379	0.00504	−1.28
	(Treat)	[0.0347]	[1.2680]	[0.0218]	[0.9869]	[0.0280]	[1.2413]
	Time effect	−0.0095	0.442	0.00968	1.186	−0.0228	1.207
	(Time)	[0.0249]	[0.6081]	[0.0233]	[0.8388]	[0.0302]	[1.4551]
	Constant term		2.706		9.027		35
	(_ cons)		[23.4859]		[13.3360]		[34.0406]
Explained variable: logarithm of self-paid medical expenses	Interaction effect	−0.0398	−1.021	−0.0337	0.492	0.00776	−0.554
	(Treat × time)	[0.0544]	[2.5011]	[0.0312]	[1.4032]	[0.0369]	[1.7911]
	Treatment effect	0.0254	−0.194	0.0157	0.324	0.00504	−1.352
	(Treat)	[0.0347]	[1.2274]	[0.0218]	[0.9537]	[0.0280]	[1.1713]
	Time effect	−0.0095	0.45	0.00968	1.259	−0.0228	0.961
	(Time)	[0.0249]	[0.5956]	[0.0233]	[0.8217]	[0.0302]	[1.2927]
	Constant term		3.438		8.901		26.84
	(_ cons)		[22.7362]		[12.9894]		[30.8461]
	Provinces	Control	Control	Control	Control	Control	Control
	Sample size	453	64	710	82	632	67

Note: the selection equations of models (1), (3), and (5) are estimated by probit model, which gives marginal effect. The standard errors are in brackets, and the standard errors of models (2), (4), and (6) are obtained by bootstrap method. The control variables of the expenditure equation include age, square of age, gender, nationality, marital status, education level, occupation, permanent address, income, previous medical history, smoking, and drinking. The control variables of the selection equation also add whether there are other relatives.