Measurenment of Capital in Cost Efficiency Analysis:

SUMMARY This thesis examines the economic and statistical sensitivity to alternative instruments for capital input in the measurement of hospitals’ cost efficiency. Efficiency is measured through a variable cost function. Capital is treated as the fixed input and measured in three alternative ways. The empirical analysis utilizes the longitudinal data on Dutch general hospitals. The three instruments for the quantity of capital are: the number of beds, the hospital’s floor space and the estimated nominal undepreciated capital stock. The results show that the statistical properties of the cost function are in favor of the nominal undepreciated capital stock, mainly because of the significance of its coefficient. Moreover, its monetary nature constitutes the most precise depiction of capital stock. Estimated mean sector efficiency is very similar between the cost function utilizing the nominal undepreciated capital stock and the number of beds. Their close association is, however, contradicted by the inconsistent rank‐order correlation. The floor space is assessed as the less reliable proxy especially due to its weak link to the hospital’s total capital stock. These findings indicate relatively high sensitivity of the efficiency outcomes to the definition of capital. As a supplement to this analysis an alternative model is developed and presented in the thesis. This model addresses the issue of endogeneity of capital in Dutch hospital sector and thus better reflects the utilization of capital.