Determinants of the Efficiency of Social Security

國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

5.2 Determinants of the Efficiency of Social Security

As mentioned above, the primary purpose of this study is to examine the

influence of fiscal decentralization on the efficiency of social security in China. Using 31 provincial-level panel data in China during 2000 to 2009, the DEA approach to obtain the efficiency scores as dependent variables and the econometric technique, Tobit model, the estimated results of equation (4.5) are presented in Table 9. There are four specifications of the empirical model used in this study. Based on the

different definitions of dependent variables and primary independent variables, model 1 uses the efficiency score under the assumption of CRS as the dependent variable and the degree of fiscal decentralization calculated by equation (3.1) as the primary independent variable. Model 2 uses the efficiency score under the assumption of VRS as the dependent variable and the degree of fiscal decentralization calculated by equation (3.1) as the primary independent variable. Model 3 uses the efficiency score under the assumption of CRS as the dependent variable and the degree of fiscal decentralization calculated by equation (3.2) as the primary independent variable.

Finally, model 4 uses the efficiency score under the assumption of VRS as the dependent variable and the degree of fiscal decentralization calculated by equation (3.2) as the primary independent variable.

From Table 9, it is obvious that the primary factor, the degree of fiscal

decentralization, which affects the efficiency of social security in China are the same results in all models. This finding suggests that the conclusions made in this study are very robust, no matter which technical efficiency indicators and fiscal decentralization indicators are adopted as dependent variables and independent variables. It is found that the degrees of fiscal decentralization have a significantly positive influence on the

‧

the empirical model. It means that as the degree of fiscal decentralization increases, provincial governments would improve their efficiency of social security because provincial governments possess more information about preferences of residents and costs of providing public goods, such as social security. Therefore, while provincial governments have more resources to use, representing the higher degree of fiscal decentralization, they can improve their efficiency of social security. However, the coefficients of FDSQ are significantly negative in three models, except the

insignificant coefficient in model 4. Although the coefficient of FDSQ is insignificant Table 9: Empirical Results of the Tobit Model

Variables Model 1: CRTEt with FDA t-1 Model 2: VRTEt with FDA t-1

Observations 279 279

Log likelihood 338.739 308.383

Observations 279 279

Log likelihood 336.294 312.102

Source: The empirical data are required by this study.

Notes: 1. Asterisks indicate variables whose coefficients are significant at the 10% (*), 5% (**), and 1% (***) levels.

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

in model 4, there are still three significant results respectively at 1% significance level in model 1 and model 3 and 5% significance level in model 2 supporting this

conclusion. This finding implies there is a non-monotonic relationship between fiscal decentralization and the efficiency of social security. That is to say, the influence of fiscal decentralization on the efficiency of social security is not always positive and it may be negative at some degree of fiscal decentralization. Their distributed

relationship may be kinked curve. This finding can explain why the relationship between fiscal decentralization and the efficiency of provincial government’s social security is controversial in some previous studies.

In addition to the degree of fiscal decentralization, other factors also play important roles in the efficiency of social security in China during 2000 to 2009. It is found that there is a positive relationship at 1% significance level between the

efficiency of social security and the gross regional product. This finding is consistent with previous expectations and is very robust in all specifications of empirical model.

This result implies that provincial residents with higher GRP per capita would be more conscious and pay more attention to their social security, and this benefits the efficiency of social security.

As to the degree of openness, there is an interesting result showed in Table 9.

Under the assumption of CRS, the coefficients of OPEN are significantly positive at 1% significance level, no matter which indicator of fiscal decentralization is chosen. It is consistent with previous expectations. The reason is that as the degree of openness increases, the wage and employment of workers are under the circumstance of higher risk. Provincial governments should bring up policies of incremental expenditures and improve the efficiency of social security to subsidize the victims in labor market.

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

of openness, no matter which indicator of fiscal decentralization is chosen.

Finally, the relationship between the efficiency of social security and the scale of provincial government is significantly positive at 1% significance level in all

specifications of empirical model. This result shows that larger provincial governments in China obtain economies of scale at the social security respect.

Similarly, the coefficients of SOGSQ are significantly negative at 1% significance level in all specifications of empirical model. Therefore, there is also a non-monotonic relationship between the efficiency of social security and the scale of provincial government. In other words, the influence of SOG may transfer from positive to negative at some level. This finding can also explain why the relationship between the scale of provincial government and the efficiency of provincial government’s social security is controversial in some previous studies.

‧