Cream-Skimming And Risk Adjustment in Colombian Health Insurance System:: The Public Insurer Case

This paper analyses the particular case of the bankruptcy of the biggest public insurer in the Colombian Health insurance system (contributory regime) in order to identify some selection atterns within such an insurance market. Using both cross-section and built-in panel data from DANE Quality of Life Survey in two waves (1997 and 2003) and applying an empirical approach based on binary choice models, the paper tries to solve two main questions. The first one is whether exists empirical evidence about advantageous selection in the contributory scheme of Colombian health insurance system. Secondly, I tried to establish whether cream-skimming (if existed) had real influence in the bankruptcy of the Colombian public insurer, which also would imply failures in the risk-adjustment formula. In the final section the principles of a good riskadjustment system suitable for the local scenario are drafted. The results show a strong evidence of dynamic selection from 1997 to 2003, which could take place by favoring both favorable age load" and good socio economic status (income, education, work type and location) for private insurers. No evidence of selection based on household analysis was found, which reinforces the idea of an individual appraisal before enrollment. This situation affected the financial performance of the public insurance, but by no means was the definitive factor of the bankruptcy. On the other hand, the risk adjustment formula used in the Colombian system presents some theoretical flaws and stillis established upon information from fifteen years ago. However, without updated information on morbidity and health care usage (not available) is not possible to measure the quantitative extent of such failures in an accurate manner"Adverse selection, cream-skimming, Colombian health insurancesystem, risk adjustment, binary choice models.