Arrow's major influence on my research, other than all of the background welfare stuff that I am too ignorant to understand how to fully make the connection, was as chair of the NOAA Panel on Contingent Valuation. Paul Portney (1994) describes it like this (emphasis added):
Environmentalists insisted that the NOAA rules embrace lost existence values as fully compensable damages and identify the contingent valuation method as the appropriate way to measure them. Not surprisingly, those upon whom these assessments might one day fall—led by the oil companies—pushed hard to exclude existence values and the contingent valuation method from the regulations. Amidst these conflicting pressures, and in recognition of the technical economic nature of the questions at debate, the General general Counsel counsel of NOAA, Thomas Campbell, took an unusual step. He asked Nobel laureates Kenneth Arrow and Robert Solow if they would chair a panel of experts to provide advice to NOAA on the following question: is Is the contingent valuation method capable of providing estimates of lost nonuse or existence values that are reliable enough to be used in natural resource damage assessments? ...
The NOAA panel met eight times between June and November of 1992. This included an extraordinary all-day hearing in August during which it heard statements from 22 experts, including several of the most prominent names in the economics profession, who either extolled the virtues of the contingent valuation method or condemned it. The panel completed its deliberations in December and, on January 11, 1993, submitted its report to NOAA. The report was published in the Federal Register on January 15, 1993.
The NOAA panel may have managed to upset everyone with its report. Those opposed to the use of the contingent valuation method were disappointed by what many took to be the “bottom line” of the panel report. This was the phrase, “ . . . the Panel concludes that CV studies [applications of the contingent valuation method] can produce estimates reliable enough to be the starting point of a judicial process of damage assessment, including lost passive use values.” Not surprisingly, this conclusion was cheered by those government agencies, academic researchers, and others wishing to make continued application of the contingent valuation method in their work.
Nevertheless, the panel reached this conclusion with some reluctance. I believe iIt would be fair to say that none of its members would have been comfortable with the use of any of the previous applications of the contingent valuation method as the basis for actual monetary damage awards. (To reiterate, none of these studies was intended for this purpose.) For this reason, the panel established a set of guidelines to which it felt future applications of the contingent valuation method should adhere, if the studies are to produce reliable estimates of lost existence values for the purposes of damage assessment or regulation.
These guidelines made a number of proponents of the contingent valuation method quite unhappy. In their view, strict adherence to the panel’s guidelines—especially the suggestion that in-person interviews be used to elicit values—would make it very expensive to use the contingent valuation method for damage estimation or regulatory purposes.
The NOAA Panel Report on Contingent Valuation, published as a white paper (in Courier font) and in the Federal Register, has been cited 3914 times according to Google Scholar. I have read and re-read the report a number of times, most recently while preparing Whitehead (2016) where a sub-group of the NOAA Panel makes the basic, yet so very important, point that it is economic, not statistical, significance of the scope test that matters.