Table 2. Summary Statistics for Ecuador-PES.
| Variable | Ecuador- PES | Ecuador-Non-PES | T-test | Normalized Difference in Means | ||
|---|---|---|---|---|---|---|
| Before match | After match | Before match | After match | |||
| Post-treatment deforestation (2011–2013) (Average annual %) | 0.15 (0.31) | 0.73 (1.43) | 7.50*** | 2.96*** | -0.39 | -0.39 |
| Pre-treatment deforestation (2004–2010) (Average annual %) | 0.31 (0.53) | 0.59 (1.1) | 3.365*** | -0.44 | -0.24 | 0.03 |
| Size of parcel (sq km) | 0.64 (0.25) | 0.47 (0.26) | -5.37*** | 0.30 | 0.49 | -0.04 |
| Distance to closest town (km) | 4.81 (3.37) | 3.89 (3.03) | -2.07** | 0.88 | 0.21 | -0.12 |
| Distance to closest road (km) | 4.13 (4.22) | 1.77 (2.32) | -4.35*** | -0.94 | 0.47 | 0.10 |
| Distance to closest river (km) | 8.43 (2.91) | 8.17 (3.35) | -0.65 | -0.70 | 0.05 | 0.07 |
| Distance to closest oil well (km) | 3.90 (3.19) | 2.61 (2.44) | -3.09*** | -0.74 | 0.27 | 0.07 |
| Observations | 63 | 450 | 513 | 112 | 513 | 112 |
*p<0.1;
**p<0.05;
***p<0.01
Standard deviations in parentheses. T-tests test for differences in means assuming unequal variances. “Before matching” uses the full sample; “After matching” is based on 1-to-1 propensity score matching without replacement and limiting the maximum distance between matches with a caliper. Normalized differences in means are calculated as recommended by [5]. There is no test for statistical significance associated with normalized differences in means but a rule of thumb is that sizes larger than 0.25 can bias simple ordinary least squares regression.