The net effects of medical malpractice tort reform on health insurance losses: the Texas experience

Health Economics Review

Table 6 Basic Differences-in-Differences-in-Differences Regression Analyses of Texas Reforms

	New Jersey	Colorado	41 State Subsample	18 State Subsample	9 State Subsample
DDD Estimator	0.6191**	0.2000	0.1505	−0.0424	0.0821
DDD Estimator	[0.270]	[0.305]	[0.245]	[0.253]	[0.271]
Treatment Dummy	0.3992***	0.2950***	0.3064***	0.3221***	0.2984***
Treatment Dummy	[0.038]	[0.039]	[0.037]	[0.037]	[0.038]
Control Dummy	1.2416***	1.5457***	1.5698***	1.6944***	1.6363***
Control Dummy	[0.231]	[0.280]	[0.113]	[0.143]	[0.174]
Control*Treatment	−0.2435	−0.5476	−0.5717**	−0.6964***	−0.6382**
Control*Treatment	[0.320]	[0.342]	[0.236]	[0.246]	[0.264]
Reform Dummy	−0.0138	−0.0741***	−0.0509***	−0.0494***	−0.0527***
Reform Dummy	[0.016]	[0.012]	[0.005]	[0.006]	[0.008]
Treatment*Reform	−0.1916***	−0.1312***	−0.1544***	−0.1560***	−0.1527***
Treatment*Reform	[0.041]	[0.039]	[0.037]	[0.037]	[0.037]
Control*Reform	0.0827	0.5019**	0.5513***	0.7442***	0.6198***
Control*Reform	[0.138]	[0.208]	[0.091]	[0.114]	[0.150]
Constant	0.2608***	0.3650***	0.3536***	0.3379***	0.3616***
Constant	[0.011]	[0.011]	[0.005]	[0.006]	[0.008]
Observations	2447	2873	42,436	21,281	11,603
R-squared	0.2338	0.3113	0.3469	0.4252	0.3794

Notes: This table presents the results of several difference-in-differences-in-differences analyses obtained using the regressions described generally in eq. 2. The dependent variable, Losses per Enrollee (LPE), is defined as the dollar amount of health insurance losses incurred by a given insurer, in a given state, during a given year, scaled by the number of plan enrollees for a given insurer, in a given state, during a given year. LPE is also scaled by 1000. “DDD estimator” is the difference-in-differences-in-differences estimator, “Treatment dummy” indicates firms operating in Texas, “Reform Dummy” indicates years following the enactment of the Texas reform measures, “Control dummy” indicates health insurers, “Control*Treatment” is the interaction of Control dummy and Treatment dummy, “Treatment*Reform” is the interaction of Treatment dummy and Reform dummy, and “Control*Reform” is the interaction of Control dummy and Reform dummy. Each column of output represents a separate analysis that differs only by the subsample of firms used as non-treated groups. Clustered standard errors are presented in parentheses and ***indicates p < 0.01, **indicates p < 0.05, and *indicates p < 0.1

ISSN: 2191-1991