Page 187 - ES 2020-21_Volume-1-2 [28-01-21]

170 Economic Survey 2020-21 Volume 1

family floater) of health insurance from years 2015-2019. The bivariate patterns between per-
capita premium and per-capita claim amount for the nine combinations (three insurer types
for each of the three insurance types) are illustrated in Figure 20. Figure 20 also includes a
linear trend line for each of the scatter plots. An upward sloping trend line is noticeable for all
scenarios but one (i.e., private insurer underwriting group-insurance schemes).

Figure 20: Correlation Patterns Between Per-Capita
Premium and Per-Capita Claim amount

5.33 To rule out any unobservable differences between insurance providers and time-varying
aggregate shocks that may systematically impact both per-capita premiums and per-capita claim
amount, we examine the same correlation after including a fixed effect for each insurer and
for each year. Figure 21 shows that the results shown in Figure 20 remain unaltered even after
controlling for these unobservable differences.

table 2: Conditional Pearson Correlation Between Per-Capita Premium and
Per-Capita Claim amount after controlling for insurer fixed effects and year fixed effects

Conditional Correlations
Government Group Individual/
type of Insurer
Sponsored Insurance Family
Private 0.968 *** -0.040 0.361***
Public Sector 0.544** 0.935*** 0.850***

Standalone 0.916*** 0.990*** 0.226
Note: ∗p<0.01; ∗∗p<0.05; ∗∗∗p<0.01