as shown in Table 2.
Overall, 86 hospital observations were classified as having both high uncompensated care and being efficient providers. Table 3 provides descriptive statistics for the market and hospital characteristics that are included in the model as independent and control variables.
The parameter estimates that measure the associations of various variables with high levels of uncompensated care and with being an efficient provider are reported in Table 4. The table notes the correlation coefficient (p) from the bivariate probit model, which equals .25 and has a chi-square value of 8.72. Given the latter, we found a positive relationship ( p < .01) between hospital efficiency and uncompensated care performance, suggesting that unobserved factors are positively related to both of these dimensions of hospital performance. Because the correlation is significant, we focused our discussion of results on the bivariate probit model.
The analysis of the factors associated with uncompensated care is presented as follows. In terms of the relationship between environmental munificence and uncompensated care, the negative coefficient for the per capita income ( p < .05) suggests that hospitals provide more uncompensated care in areas where residents have lower income. The coefficient of the unemployment rate, a proxy for the uninsured demand, shows a negative association with the probability of being high uncompensated care providers ( p < .01). This result is not consistent with prior studies (Rosko, 2004b). The estimated coefficient of major safety net in the region is negatively related to the probability of being a high uncompensated care provider ( p < .01). This means that if there is a major safety net hospital within a region, other hospitals located in the same area tend to provide less uncompensated care. This result is consistent with Campbell and Ahern (1993) and Davidoff et al. (2000).
As expected, we found negative relationships between Medicare share ( p < .1) and Medicaid share ( p < .01), with the probability of being a high uncompensated care