Crime and immigration: new evidence from England and Wales

Table 6 IV Estimates for A8 and Non-A8 immigrant, separate regressions

	Dependent variable: change in crime rate
	Annual Population Survey, 2004/5-2010/11
	A8 immigrant share		Non-A8 immigrant share
	First stage	Second stage	First stage	Second stage
	(1)	(2)	(3)	(4)
A. England and Wales
Predicted change in A8 immigrant share	0.289 (0.075)
Predicted change in non-A8 immigrant share			0.627 (0.151)
High historical immigrant share			-0.046 (0.014)
Change in A8 immigrant share		-0.429 (0.275)
Change in non-A8 immigrant share				0.045 (0.096)
F-test	14.91		9.35
Controls	Yes	Yes	Yes	Yes
Police force area fixed effects	Yes	Yes	Yes	Yes
Sample size	347	347	347	347
B. London
Predicted change in A8 immigrant share	0.213 (0.091)
Predicted change in non-A8 immigrant share			0.921 (0.262)
High historical immigrant share			-0.054 (0.019)
Change in A8 immigrant share		0.130 (0.469)
Change in non-A8 immigrant share				0.117 (0.130)
F-test	5.37		6.24
Controls	Yes	Yes	Yes	Yes
Police force area fixed effects	Yes	Yes	Yes	Yes
Sample size	33	33	33	33

Notes: Separate population weighted regressions for A8 and Non-A8 immigrants. High Historical Immigrant Share is a dummy variable equal to one if the Immigrant Share in 1991 Census of the local authority is greater than 0.20. For A8 regressions this dummy is not included as there was no local authorities with high historical A8 immigrant shares. Controls are: population growth, the change in the unemployment rate, the change in the share of males aged 15–39 and a dummy variable for the 33 London boroughs. Robust standard errors in parentheses.