Valuing Access to our Public Lands: A Unique Public Good Pricing Experiment



As shown in Table 4, the likelihood-ratio statistic testing for the existence of heteroscedasticity is large
and leads us to reject the null of homoscedastic errors in the NPF sample but not in the RDD sample.

Mean and median WTP figures are presented at the bottom of Table 4. The mean and median WTP
values for the RDD sample are approximately $42 and $35, indicating a WTP distribution that is slightly
skewed toward higher WTP values. The mean and median WTP values for the NPF sample are
substantially higher and are both approximately equal to $95.6 Next, we turn to the measurement of
hypothetical bias.

5.2 Hypothetical Bias Model

We now describe the econometric procedure to calibrate estimated WTP to be consistent with stated
rates of GEP purchases from within our survey sample. We intentionally designed our valuation
experiment to account for hypothetical bias by comparing hypothetical decisions to buy the NRP (at a bid
equal to $65) with decisions to purchase the existing GEP, which has been sold for $65 in the
marketplace. Assuming households are accurately stating whether or not they purchased the GEP within
the previous year, this internal calibration should result in WTP and revenue estimates that reflect the
actual purchasing behavior of households in the general U.S. population.

To estimate the potential degree of hypothetical bias in the survey data, we specify a probit model
with a dummy variable to capture the difference between hypothetical and actual purchasing decisions.7
We analyze the RDD and NPF samples separately. Within each of these samples, we pool data from two
distinct subsamples—the
stated preference and the revealed preference subsamples. The revealed
preference subsample includes every household who knew of the $65 GEP. The stated preference

are equally attractive recreation areas, etc. For these reasons, we do not include this spatial variable in our empirical
analysis but leave it as a possible avenue for future research.

6 We also estimated a single-bounded dichotomous-choice (SBDC) WTP model. (The results are not shown here
but are available upon request from the authors.) The predicted NPF WTP distributions and coefficient estimates
from the SBDC model are very similar to the results from the DBDC model. For the RDD sample, the calibrated
mean and median WTP from the SBDC model are approximately $7 lower than from the DBDC model, and the
coefficient estimates are qualitatively similar.

13



More intriguing information

1. FASTER TRAINING IN NONLINEAR ICA USING MISEP
2. The name is absent
3. Om Økonomi, matematik og videnskabelighed - et bud på provokation
4. Tax Increment Financing for Optimal Open Space Preservation: an Economic Inquiry
5. The name is absent
6. Iconic memory or icon?
7. The name is absent
8. The name is absent
9. The name is absent
10. Feature type effects in semantic memory: An event related potentials study
11. The name is absent
12. Wage mobility, Job mobility and Spatial mobility in the Portuguese economy
13. The name is absent
14. The name is absent
15. The name is absent
16. An Efficient Circulant MIMO Equalizer for CDMA Downlink: Algorithm and VLSI Architecture
17. The name is absent
18. The name is absent
19. The name is absent
20. The name is absent