Table 3.

IV Two-Stage Least Squares Instrumental Variables Regression Results.^a

DV: Performancei , t + 1	Model 1	Model 2	Model 3 (IV)
(A) All cases
Performancei,t	0.323***	0.248***	0.238**
Diversificationi,t			9.414
Firm brand breadthi,t			158.210
Brand intensityi,t			−368.824*
Ln assetsi,t		13.484	13.652
Ln costsi,t		−130.984***	−128.776***
${\text{Ln costs}}_{i,t}^2$		6.481***	6.521***
Dummy missing costs		−591.880***	−569.108***
N	264	263	218
R 2	.100	.266	.304
Cragg-Donald Wald F			28.921
Sargan-Hansen test  (p value)			0.9591
(B) Pure-service cases
Performancei,t	0.223	0.228*	0.121
Diversificationi,t			203.918*
Firm brand breadthi,t			−5,001.265*
Brand intensityi,t			−1.6e+03
Ln assetsi,t		−10.627	−48.683
Ln costsi,t		−182.851***	−119.929***
${\text{Ln costs}}_{i,t}^2$		8.546***	3.328
Dummy missing costs		−872.497***	−891.949***
N	82	81	80
R 2	.042	.592	.572
Cragg-Donald Wald F			7.664
Sargan-Hansen test  (p value)			0.2897
(C) Cases with products
Performancei,t	0.264**	0.086	0.084
Diversificationi,t			−17.948
Firm brand breadthi,t			203.798
Brand intensityi,t			−136.817
Ln assetsi,t		34.661*	33.491*
Ln costsi,t		6.737	7.488
${\text{Ln costs}}_{i,t}^2$		0.606	0.615
Dummy missing costs		176.584	184.481
N	138	138	137
R 2	.072	.188	.199
Cragg-Donald Wald F			5.422
Sargan-Hansen test  (p value)			0.809

^aThe estimation uses prestige rank, patents, and age as instruments for brand breadth. We report the estimated coefficients and their significance.

*p < .05. **p < .01. ***p < .001.