used to produce other forms of health care output. OLD is the proportion of the population
above 75 years of age, and our hypothesis is that an older population will require more health
care being produced. As such, the proportion of old inhabitants in a local county council will
affect the amount of resources in health care production and, thus, the efficiency of the county
council.
Furthermore, three variables, CIRCULATORY, NERVOUS SYSTEM and PSYCHIATRIC,
have been included to control for differences in case mix between county councils. These
variables are measured using the Nordic Diagnosis Related Group (NDRG) system. The
original DRG system was invented at Yale University during the 1960ties, and the system was
developed in order to measure case mix in hospitals and/or local county councils in a practical
manor.
NON-SOCIALIST is a dummy variable equalling one if the local county council has a non-
socialist government. Previous studies (Gerdtham et al, 1999) have shown that non-socialist
county councils have higher economic efficiency than other county councils, and we thus
expect that a non-socialist local government will have a positive impact on efficiency.
TREND is a time trend included to capture possible time trends in efficiency.
Finally, an inspection of the efficiency scores also showed that one county council,
Norrbotten, was a clear outlier when it comes to efficiency scores. As such, in addition to the
model presented above, a model including an indicator variable equalling one if the
observation was related to the county council of Norrbotten, DNORRBOTTEN, was also
estimated. Descriptive statistics for the variables used in the empirical analysis are presented
in Table 2, and the results from the estimations of the two different Tobit models are
presented in Section 3 below.
Table 2 about here.
3. Results
Turning to the results from the estimation of the Tobit regression model, parameter estimates
and standard errors are presented in Table 3. Note that the parameter estimate relating to the
indicator variable for Norrbotten is negative and statistically significant at the 5 percent level.