Table 3
Demand Estimation Results Using Upstream HUC12s Instruments by Quarters
| Q1 | Q2 | Q3 | Q4 | |
|---|---|---|---|---|
| First stage | ||||
| Travel costs | −0.005*** | −0.006*** | −0.007*** | −0.005*** |
| (0.000) | (0.000) | (0.000) | (0.000) | |
| Observations | 411,401 | 411,401 | 411,401 | 411,401 |
| Second stage | ||||
| Predicted advisory | −0.353*** | −0.509*** | −0.574*** | −0.375*** |
| (0.108) | (0.191) | (0.206) | (0.140) | |
| Observations | 1445 | 1445 | 1445 | 1445 |
| ASC-by-quarter fixed effects | Yes | Yes | Yes | Yes |
| IVs | Yes | Yes | Yes | Yes |
| Kleibergen-Paap rk Wald F | 29.76 | 29.76 | 29.76 | 29.76 |
| Stock-Yogo | 16.38 | 16.38 | 16.38 | 16.38 |
| MWTP | −73.178 | −73.552 | −87.494 | −69.462 |
Note: The Kleibergen-Paap rk Wald F-statistic is used in instrumental variable (IV) regressions to test for weak instruments. The Stock-Yogo threshold provides critical values to assess weak instrument bias. MWTP are derived as the ratio of the advisory coefficient to the marginal utility of travel cost (i.e., −
). We generated 1,000 bootstrapped values of MWTP for each data sample and constructed 95% confidence intervals by deleting the top 5% and the lowest 5% of the distribution of bootstrapped values. MWTP = marginal willingness to pay.↵* p < 0.1; ** p < 0.05; *** p < 0.01.
