Table 6.

Estimation results for the health plan price elasticity in the HIA market 2006–2015

log (s it) = α log (s i,t−1) + βp it + γ i + δ t + ɛ it, ${\overline{p}}_t$ ranging from 1100 to 1300 euros (Table 2)
(iv) OLS estimation, Number of observations: 577
$\widehat{\mathit{\alpha }}$ = 0.95*** (0.01)	$\widehat{\mathit{\beta }}=-$0.0017** (0.0008), ${\overline{\mathit{\epsilon }}}_t=\widehat{\mathit{\beta }}{\overline{p}}_t$ between −1.9 and −2.2 ${\widehat{\mathit{\tau }}}_t$ between −34 and −40	R 2 = 0.97
(v) GMM estimation, number of observations (including levels): 1013
$\widehat{\mathit{\alpha }}$ = 0.81*** (0.03)	$\widehat{\mathit{\beta }}=-$ 0.0008*** (0.0001), ${\overline{\mathit{\epsilon }}}_t=\widehat{\mathit{\beta }}{\overline{p}}_t$ between −0.9 and −1.0 ${\widehat{\mathit{\tau }}}_t$ between −4.7 and −5.5	R 2 = 0.99

The estimations are performed with the plm-package in R [25], total number of policies used is 155, making distinction between collective and individual policies and using only policies with a minimum of 10,000 enrollees. We use a system GMM estimator with endogenous premiums, including time dummies and fixed effects. Sargan statistics (85.83, D.f. = 68, P value = 0.071), Wald test for coefficients of this model (D.f. = 2) has a P value <0.2 e−16. Additional estimations results are available by the authors upon request

* P value <0.1; ** P value <0.05; *** P value <0.01