R. J. Rhodes, J. C. Whitehead, T. I. J. Smith and M. R. Denson
Published online: 11 May 2018
Abstract: Recreational saltwater anglers from the mid-Atlantic through the Gulf of Mexico commonly target red drum. Due to concerns about overharvesting within South Carolina coupled with regional management actions, South Carolina explored the technical feasibility of stocking hatchery-produced juvenile red drum as a technique to augment the abundance of South Carolina stock. In order to assess a continued program, in 2005 a mail survey was used to collect data for estimating the economic benefits with the contingent valuation method. The theoretical validity of willingness to pay was assessed by comparison to the value of a change in red drum fishing trips that would result from the program. Benefits were compared to estimated, explicit stocking costs. We illustrate how a certainty recode approach can be used in sensitivity analysis. The net present values (NPVs) for the stocking program are positive suggesting that the program would have been economically efficient relative to no program.
I like this paper for a number of reasons:
- It was first presented over 10 years ago at the Southern meetings. It sat idle for 10 years and JBCA was the first journal where it was submitted. The lesson is that ... don't let your data sit idle for 10 years.
- The paper is a benefit-cost analysis, actually used by the agency (see Stocking South Carolina's Favorite Saltwater Fish). I've been working on benefits forever but rarely get to actually compare them to costs. In addition, I got to work with real-life agency fisheries economists and biologists who had real-world constraints on the survey questions they were able to ask. I maybe learned more from this project than any other.
- I've taught this case study for a number of years in my benefit-cost analysis course. I haven't taught the course in years (department chair duties got int the way) but am scheduled to teach it this fall. I'll be using it again this fall to help explain sensitivity analysis, discounting, horizon values, etc.
- We did revealed preference (travel cost method) and stated preference (contingent valuation method) analyses to estimate benefits. The revealed preference estimates were less informative than the stated preference estimates so the latter were used in the policy analysis. This illustrates my contention that stated preference data can be useful for policy analysis but (at least for recreation) you are better off if you also consider revealed preference estimates (as a validity check -- and see the next bullet point).
- The journal referees were not much impressed with our assertion that an old case study was worthy of publication. So, I went back to my all-time favorite paper for some theory and found convergent validity between single-site recreation demand and contingent valuation. The lesson here may be that if a researcher is looking for a better understanding of behavior and preferences from stated preference data then things can be discovered and learned. On the other hand, if a researcher is looked to break stated preference methods then that can be arranged too.
- Hypothetical bias is an issue that will likely never go away in contingent valuation (or choice experiments if folks doing that sort of stated preference study were targeted by industry-funded critics) regardless of assertions about consequentiality. In this paper we used a certainty follow-up and the analysis in the preceding bullet point to best match the willingness to pay estimates with a revealed preference estimate. From this, we used alternative certainty corrections for sensitivity analysis. As of today, this is how I think benefit-cost analysis with stated preference methods should be done.