A Two-Stage DEA to Analyze the Effect of Entrance Deregulation on Iranian Insurers: A Robust Approach

Table 5

Second stage—Generalized estimating equations models that capture the most important factors explaining the efficiency score for robust DEA.

Predictors	Models
	Model 1		Model 2		Model 3		Model 4
	𝐵	Hypothesis Test-Sig.	𝐵	Hypothesis Test-Sig.	𝐵	Hypothesis Test-Sig.	𝐵	Hypothesis Test-Sig.
(Intercept)	.681	.006	.681	.005	.779	.000	1.141	.000
Private	−.367	.000	−.370	.000	−.352	.000	−.402	.000
Stock	.118	.143	.118	.138	—	—	—	—
Big	.165	.193	.172	.129	.178	.093	—	—
Marketshare	.001	.751	—	—	—	—	—	—
Capitalstructure	.007	.003	.007	.003	.006	.004	.005	.005
PrmtoSrp	.018	.000	.018	.000	.017	.000	.017	.000
QIC^a	46.367		44.537		41.933		35.142
QICC^a	25.739		23.785		21.424		19.649
Residual normal test^b; Asymp. Sig. (2-tailed)							.224

^aComputed using the full log quasilikelihood function.
^bBeing normal is not rejected at 5% level if the statistic is more than 0.05.