Table 3.

System GMM estimates for Rural Broadband Availability.

	25/3 - Rural (1)		Fiber - Rural (2)		2+ Competitors - Rural (3)
	Coeff.	S.E.	Coeff.	S.E.	Coeff.	S.E.
Lag (availability) - Rural	0.448	0.015***	0.818	0.025***	0.803	0.019***
% Bachelor's Degree or More – Rural	0.264	0.029***	0.095	0. 022***	0.104	0.018***
% Poverty – Rural	−0.401	0.046***	−0.091	0.028***	−0.112	0.192***
% Housing after 2010 - Rural	−0.152	0.096	0.382	0.065***	0.059	0.075
ln (Median Household Income – Rural)	−0.035	0.004***	−0.009	0.330***	−0.016	0.002***
ln (Rural Population)	0.045	0.003***	−0.010	0.168***	0.015	0.001***
Rural % of Population	0.019	0.014	0.050	0.009***	−0.008	0.006
Topography	−0.001	0.001	−0.000	0.000	0.000	0.001
Year F.E.	Yes	***	Yes		Yes
Policy Variables
State funds	0.018	0.007***	0.021	0.005***	0.014	0.004***
State office	−0.007	0.008	0.015	0.006**	0.003	0.005
Municipal restrictions	−0.037	0.011***	−0.016	0.006**	0.003	0.006
Constant	0.175	0.042***	0.208	0.032***	0.037	0.024
Wald Chi Squared	15,432***		7,215***		17,635***
# Instruments	54		53		52
# Groups	3,028		3,028		3,028
Hansen J-test	0.251		0.304		0.275
AR(1)	0.000***		0.000***		0.000***
AR(2)	0.201		0.621		0.042**
# Observations	18,159		18,159		18,159

*, **, and *** represent statistical significance at the p < .10, .05, and 0.01 levels, respectively.

Hansen J-test represents p-values for the null hypothesis of valid instruments (overidentification).

AR(1) and AR(2) represent p-values for null hypotheses of no 1st and 2nd-order autocorrelation.