what the method "doubleb" uses when it calculates standard error of willingness to pay

[Yixin Wang]

Hi all,

Does anyone know what the method "doubleb" uses when it calculates the standard error or confident intervals of willingness to pay?

When we calculate willingness to pay in STATA, we use the code like this:

*WTP for mean values
nlcom (WTP:(_b[_cons]+age_m*_b[age]+female_m*_b[female])), noheader

And we get the results like this:


I know when we calculate standard error or confident interval of WTP we can use the delta method, Krinsky and Robb method and bootstrap method, I wonder what method it uses here?

I saw the source of "doubleb" but it did not help.

[Yixin Wang]

In my opinion, it is more like the delta method as the confidence interval is symmetric.

[Hong Il Yoo]

As the developer of -doubleb-, Alejandro Lopez-Feldman, describes in his background paper (https://mpra.ub.uni-muenchen.de/41018/), -doubleb- parameterizes the model directly in terms of the mean and standard deviation of the assumed WTP distribution, and estimates those parameters directly. If you require any additional information, I recommend Chapter 7.5 of Train's "Discrete Choice Methods with Simulation" (2nd ed.); Chapters 2 and 5 of Haab and McConnell's "Valuing Environmental and Natural Resources"; and Chapter 7.2.4 of Verbeek's "A Guide to Modern Econometrics" (4th ed.).