ABSTRACT
The above six independent variables were all control variables, and our focus was on the impact of health care reform policy on health service efficiency. The DEA-TOBIT model used in this study was as follows:
Efficiencyit ¼ β0 þ β1Policyit þ β2Densityit þ β3Illiteracyit þ β4GDPit þ β5Urbanizationit þ β6Dk þ ui þ eit
The independent variable Efficiency represents the DEA efficiency scores of provincial health services. To account for the effect of health care reform, the independent variable Policy takes a value of 0 for the period before adoption and 1 after. Regarding the other independent variables, Density, dividing the size of a province’s geographical area by its year-end population, signifies the density of population; Illiteracy, measured by the illiteracy rate, indicates the educational level of each province; GDP is a proxy for income level and is measured by provincial GDP per capita; Urbanization, estimated by the share of urban residents in a province’s total population, indicates the degree of urbanization; Region, divided into three districts, dummy variables D1 and D2 respectively represent central and eastern regions (the western region is the reference group) in China. The random variable u changes with individual data but is uncorrelated with time and the explanatory variables. The random variable e changes independently over time as well as across individual data. The interpretations for the remaining terms are as follows: β0 is the constant term of the regression equation; β1-β6 are the regression coefficients of corresponding independent variables; i stands for DMU, which indicates the number of provinces; and t signifies the year.
