The examination of the causal impact of health insurance coverage on healthcare utilisation is a critical endeavour in both academic research and policy formulation. However, this endeavour faces challenges, notably the endogenous selection into coverage and prevalent misreporting of coverage status. This paper contributes to the literature by providing the first empirical evidence regarding the magnitude and direction of the bias resulting from failing to account for both endogeneity and misreporting in health insurance coverage. It surmounts the above-described critical research hurdles by utilising recently available linked survey and administrative individual data from Australia and employing various methodologies to address potential endogeneity issues in insurance enrolment.
This study offers two primary insights. Firstly, our analysis reveals that individuals with private health insurance (PHI) coverage generally exhibit higher utilisation rates of healthcare services, particularly evident in primary care visits and specialist consultations. However, the extent and significance of these effects demonstrate variability across different types of healthcare services and methodological approaches employed. Secondly, this study emphasizes the crucial role of employing accurate measures of PHI coverage. We observe substantial discrepancies in the magnitude of PHI estimates derived from survey-based and administrative data sources. However, the extent and significance of these discrepancies vary across different healthcare service types and methodological approaches. Notably, our preferred model suggests that utilising a self-reported PHI indicator with a 10% misreporting rate would lead to a significant overestimation of PHI’s impact on the two most commonly used healthcare services.
Overall, the comprehensive consideration of both endogeneity and misreporting in PHI coverage suggests that the positive relationships between PHI coverage and healthcare utilisation may be less pronounced than previously depicted. This research provides valuable insights for studies utilising datasets similar to ours to investigate the relationship between health insurance and health care utilisation.