TABLE 4

Results of Difference-in-Differences (DID) and Probit Marginal Effects (ME)

	Unmatched (1)		Matched^{^a} (2)		Matched without 2 km Buffer (3)
Model	Coeff.	Clustered Std. Err.	Coeff.	Clustered Std. Err.	Coeff.	Clustered Std. Err.
All Groups
DID	−0.02	0.024	−0.042	0.038	−0.046	0.037
Probit ME	0.022^**	0.011	0.004	0.008	0.0098	0.009
Observations	295,001		169,666		134,192
R-squared	0.12		0.15		0.14
Nonresident
DID	−0.04^**	0.02	−0.033	0.02	−0.043^**	0.02
Probit ME	0.003	0.02	−0.014^**	0.007	−0.014^*	0.007
Observations	293,248		90,190		73,159
R-squared	0.13		0.47		0.098
Long-inhabited
DID	−0.043^**	0.02	−0.050^*	0.028	−0.064^*	0.033
Probit ME	−0.018	0.013	−0.022^**	0.011	−0.024^**	0.012
Observations	294,679		94,204		84,471
R-squared	0.13		0.11		0.12
Recently settled
DID	0.026	0.022	−0.078^*	0.047	−0.077^***	0.019
Probit ME	−0.001	0.013	−0.040^**	0.016	−0.015	0.014
Observations	296,262		43,116		25,682
R-squared	0.08		0.25		0.065

Note: Standard errors are clustered at the “concession” region level with 25 clusters for “all groups” and 17 clusters in the subgroups.
↵a Matched samples are based on nearest-neighbor propensity score matching with Mahalanobis distance metric with calipers set to 0.25 standard deviations taken from the sample with 300,000 observations. Results based on a linear probability model where the dependent variable equals 1 if pixel is deforested, 0 otherwise.
↵* Significant at 10%;
↵** significant at 5%;
↵*** significant at 1%.