Click for Academy of Management home page

Academy of Management

Second opinions worsen conflicts of interest, study finds

August 15, 2013

For more information, contact: Ben Haimowitz, 212-233-6170,

Aug. 15 - A new client of a financial counselor, seeking to discourage him from exaggerating the virtues of the funds he recommends, asks if it will be okay to solicit a second opinion. After all, it is not unheard of for financial advisors to put their own interests ahead of their clients', and soliciting second opinions has come to be relied upon in all manner of business deals as an antidote to possible advisor conflicts of interest.


But research that was presented at the annual meeting of the Academy of Management (Orlando, Florida, Aug. 9-13) raises cautions about this common practice. While seeking second opinions can be beneficial, the researchers find, it can also have the perverse effect, unrecognized until now, of provoking advisors to be more misleading than they might otherwise be.


Thus, for example, conflicted financial advisors will more greatly exaggerate the investment potential of funds they recommend if told a client will seek a second opinion than if nothing is said.


In the words of the paper, by Sunita Sah of Georgetown University and George Loewenstein of Carnegie Mellon University, "Our experiments point to an adverse consequence of second opinions that has not, to the best of our knowledge, been discussed in prior work: the potential for advisors to give more biased advice when they are aware their advisees have access to a second opinion...People regularly engage in unethical acts without violating their moral self-identity, and our research suggests that the potential availability of second opinions can provide one more rationalization that people, and specifically advisors, can use to rationalize self-interested, unethical behavior."


They add: "Primary advisors could feel morally liberated to give biased advice because they are aware that, with the availability of a second opinion, the advisee now has access to more information...People also feel less responsible for, and are hence less generous towards, aid-recipients who have another potential source of aid."


The findings emerge from a pair of ingenious experiments conducted online with hundreds of anonymous university alumni who were assigned roles as advisors or advisees. Advisors with conflicts of interest - so-called primary advisors -- provided advice to an equal number of advisees, about two thirds of whom had access to a second opinion from unconflicted advisors.


In both experiments advisees were asked to estimate how many filled-in circles there were in a 30-by-30 grid of small circles. Primary advisors were informed that exactly 303 circles were filled in; advisees had only sketchy knowledge based on a 3-by-3 grid, but were aware primary advisors knew the exact number; and secondary advisors were not informed of the exact number but had unlimited access to view the full grid. Advisees and secondary advisors had congruent interests: each received $5 if advisees came within 10 of estimating the number of filled circles. In contrast, primary advisors had interests at odds with advisees', being rewarded if advisees overestimated the number of filled circles -- $5 if they did so by less than 100 and $10 if they overestimated by100 or more.


"This set-up," the authors explain "was designed to simulate a situation in which an advisee received advice from a better informed but conflicted primary advisor and a less well informed but unconflicted secondary advisor." Why less informed? "If the second advisor had a similar level of knowledge but no conflict, there would be little reason to pay attention to, or hire, the first adviser."


In one experiment, primary advisors who were aware a second opinion was available (without knowing what it was) indicated, on average, that there were 474 filled-in circles, which was 171 above the actual number. This was well above the 406 average for those who were either unaware of a second advisor or knew there was none. Seventy-nine percent of the aware group gave advice above 312 (the maximum for which advisees could earn $5) compared to 51% of the others. As the authors put it, "When primary advisors know about the second advisor, they gave more biased advice, and perhaps felt justified in doing so; they openly indicated that they acted more selfishly, gave higher priority to maximizing their own payoff, and stated that they did not want to help the advisee."


The second experiment was similar to the first except that it included an additional group of first advisors: in addition to those who served as solo advisors and those informed of a second advisor, there was a third group who were informed that advisees had access to an optional unconflicted second opinion if they were willing to pay $2 for one. The idea was to test what the effect of the mere possibility of a second opinion would be, and, as far as bias was concerned, the effect was virtually nil. In the words of the study, "Primary advisors in the two conditions in which they were aware of the presence of a second advisor (mandatory or optional) did not differ in the advice they gave but, similar to the first experiment, gave significantly more biased advice than primary advisors acting as solo advisors."


Despite the perversity on the part of primary advisors uncovered by the study, second advisors proved to have clear benefits for advisees. In the two experiments described in the paper, their availability resulted in considerably better estimates by advisees than were achieved in their absence. In the second experiment, for example, 22% of participants with solo advisors made estimates that were accurate within 10 dots, while about double that percentage achieved that level of accuracy when second opinions were available, whether free or for a price.


In the end, the authors posit that "having second opinions available, but unused, might be the worst situation in terms of producing perverse effects with no benefits. One potential policy that could increase the likelihood that second opinions will provide advantages without producing perverse effects would be to provide second opinions confidentially, and, of course, at reasonable costs."


And, best of all, if feasible, they conclude, is not to seek ways to compensate for conflicts of interest but to eliminate them altogether, since "doing so promotes trust in primary advisors and renders the expense (and potential perverse effects) of second opinions obsolete."


The paper, entitled "Second thoughts on second opinions: Conflicted advisors reduce the quality of their advice when they know they will be second-guessed," was among several thousand research reports at the Academy of Management annual meeting, held in Orlando from August 9th through 13th. Founded in 1936, the Academy of Management is the largest organization in the world devoted to management research and teaching. It has some 19,000 members in 110 countries, including about 11,000 in the United States. This year's annual meeting drew more than 9,000 scholars and practitioners for sessions on a host of subjects relating to business strategy, organizational behavior, corporate governance, careers, human resources, technology development, and other management-related topics.


Academy of Management
Member Services
Academy of Management
Online Opportunities
Academy of Management