The system effect and group benefit equity of long-term care insurance from the perspective of short-term policy pilot

Abstract

Methods: This study focuses on the long-term care insurance (LTCI) policy pilot, using the CHARLS database to continuously track survey data. It constructs a difference in-difference model based on city, time, coverage, and beneficiaries to accurately identify policy coverage and empirically examine the institutional effect of the long-term care insurance policy pilot and the fairness of group benefits.

Results: The results indicate that the policy pilot has a significant positive impact on the overall medical consumption of disabled older adults, with impacts on monthly outpatient consumption, annual hospitalization consumption, annual hospitalization times, and last hospitalization days of 0.7064, 0.4142, 0.0887, and 1.5607, respectively. In addition, the LTCI policy pilot significantly and positively affected disability-related health indicators such as individual self-assessment health, ADL disability, and the number of serious diseases, with effect sizes of 0.8677, 1.0854, and 0.6668, respectively.

Discussion: The results regarding group benefit equity show that the LTCI policy pilot can improve the equity of medical consumption and disability-related health among groups; however, over time, it may exacerbate the inequality of medical consumption and disability-related health among disabled older adults in the treatment group. Based on this, the study finds that the LTCI policy pilot has effects on medical consumption and disability-related health for disabled older adults, primarily driven by the moral hazard associated with the assessed individuals obtaining LTCI treatment due to the short-term policy pilot.

Keywords: disability; inequality; institutional effect; long-term care insurance; policy pilot.

Figure 1

Test of the parallel trend assumption. (a) Test 1 of the parallel trend assumption (Monthly outpatient consumption). (b) Test 2 of the parallel trend assumption (Monthly outpatient times). (c) Test 3 of the parallel trend assumption (Annual inpatient consumption). (d) Test 4 of the parallel trend assumption (Annual inpatient times). (e) Test 5 of the parallel trend assumption (Last hospitalization days). Each vertical axis in the figure represents the trend of core variables affected by the policy change. The horizontal axis represents the time points before and after the implementation of the policy, with pre_3 and pre_1 indicating the third and first phases before the policy’s implementation, corresponding to 2013 and 2015, respectively; post_2 refers to the second period after the policy’s implementation, corresponding to 2018 in this article. The data in this work do not include 2014, 2017, and 2019, so the corresponding pre_2, post_1, and post_3 do not exist.

Figure 2

Density distribution of regression coefficient of medical consumption. (a) Distribution of monthly outpatient consumption coefficient. (b) Distribution of monthly outpatient times coefficient. (c) Distribution of annual inpatient consumption coefficient. (d) Distribution of annual inpatient times coefficient. (e) Distribution of last inpatient days coefficient.

Figure 3

Density distribution of regression coefficients of disability-related health indicators. (a) Distribution of self-assessed health coefficient. (b) Distribution of ADL disability coefficient. (c) Distribution of number of serious diseases coefficient.