Adverse selection or beneficial selection in dual western medicine-traditional Chinese medicine insurance markets

As pluralistic health systems increasingly integrate traditional and biomedical care, there is a need for empirical evidence on how add-on Traditional Chinese Medicine (TCM) coverage shapes behavioural responses and thus informs benefit design. Hong Kong offers such a case: 10.7% of the population opt into voluntary plans with TCM add-ons. If adverse selection dominates, higher-need individuals may be more likely to choose TCM-inclusive plans, biasing non-IV estimates upwards. Conversely, if TCM add-ons attract more health-conscious enrollees with lower expected needs, non-IV estimates would be biased downward (attenuated toward zero). We therefore use an Instrumental Variable (IV) approach to control for endogenous selection bias. We find that IV estimates are two to four times larger than non-IV estimates, a pattern consistent with beneficial selection. After accounting for endogeneity, TCM-inclusive coverage is associated with higher TCM use and WM use across both extensive (any use) and intensive (number of visits) margins, and with lower OOP spending for recent outpatient care. Overall, voluntary adoption of TCM-inclusive coverage may proxy a lower baseline propensity to utilize care, even though coverage itself is associated with increased utilization and improved financial protection.