Booming with Speed: High-Speed Rail and Regional Green Innovation

Table 8

IV estimation.

Panel A: modified 1962 railway network
	(1)	(2)	(3)	(4)	(5)
	L.HSR	GUMA	GIA	GUMG	GIG
IV-modified 1962 railway network	0.0379
	(0.00204)
L.HSR		0.376	0.583	0.339	0.138
		(0.0473)	(0.0629)	(0.0421)	(0.0143)
Controls	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes
City	Yes	Yes	Yes	Yes	Yes
Weak identification test
Kleibergen–Paap rk Wald F statistic		304.951
Stock–Yogo weak ID test critical values (10%)		16.38
First-stage F statistic		157.7
N	3923	3923	3923	3923	3923
Panel B: relief degree of land surface
	(1)	(2)	(3)	(4)	(5)
	L.HSR	GUMA	GIA	GUMG	GIG
IV-relief degree of land surface	−0.0145
	(0.00151)
L.HSR		0.929	1.183	0.801	0.229
		(0.118)	(0.152)	(0.103)	(0.0316)
Controls	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes
City	Yes	Yes	Yes	Yes	Yes
Weak identification test
Kleibergen–Paap rk Wald F statistic		34.735
Stock–Yogo weak ID test critical values (10%)		16.38
First-stage F statistic		137.1
N	3990	3990	3990	3990	3990

Note: the estimation sample includes 285 prefecture level and above cities in China. The sample period is from 2005 to 2018. Dependent variables are indicated by column titles in italics. All control variables are included and lagged by one year. All regressions include city fixed effects and year fixed effects. Robust standard errors in parentheses are clustered at the city level. The symbols , , and indicate significance at the 1%, 5%, and 10% levels, respectively. We use the Kleibergen–Paap rk Wald F statistic to test for weak identification of the endogenous variables [101]. The critical value compiled by Stock and Yogo [102] is 16.38. Our results are also robust to the “rule of thumb” of Staiger and Stock [103], which requires the first-stage F statistic to be larger than 10.