Research in the physical sciences is fairly straightforward. When physicists plot the trajectory of a planetary probe, for instance, they might bicker among themselves about methods, but their results are empirical, quantifiable, and either reproducible or not.
Retirement research—which involves the social sciences—is a different animal. Its subject is people, in all their complexity. It has political overtones, as debates over Social Security show. Its results may be used, ignored or abused by the trillion-dollar retirement industry.
Certainly, retirement research itself is a large and growing business. An informal tally of spending on it at universities, at firms like Towers Watson and Mercer, by groups like the Employee Benefit Research Institute, and on countless industry-sponsored surveys, whitepapers and thought-leadership projects, quickly produces a conservative estimate well north of $100 million.
Given that effort, it seems appropriate to reverse roles for once and ask the researchers a few questions. RIJ talked to several retirement researchers—who were either inside industry, outside industry, or had a foot in both worlds—about the conflicts of interest they face and their assessment of the value of what they do. We found a world of grayscale rather than black and white.
The view from Palo Alto
Financial Engines sits astride the academic/industry fault line. Founded by Nobelist William Sharpe, an emeritus professor from Stanford University, the Palo Alto firm provides managed account and advisory services for retirement plan participants. Not surprisingly, its in-house Retirement Research Center focuses on the drawdown phase of retirement planning for 401(k) participants.
Jason Scott, the center’s managing director, says his research is intended to help develop products, but also to be published and even peer-reviewed. “That allows us to collaborate with academics. They can ’kick the tires’ on it. If they like it, that gives us a lift.”
By publishing its research on deferred income annuities, for instance, Financial Engines discovered something it didn’t know. “We published in Financial Analyst Journal, a publication for practitioners, and in the Journal of Risk and Insurance, an academic journal. The study asked: Where do annuities deliver the most bang for the buck? It turned out to be longevity. You buy the annuity at age 65 but don’t start payments until you reach your life expectancy [at ages 83 to 85].
“By publishing, we learned that there’s a regulatory barrier. The IRS requires minimum distributions from retirement accounts starting at age 70½. People who bought a longevity annuity and delayed income to age 85 would have no way to comply.” Just last week, regulators announced that they had addressed that problem. [Interestingly, Financial Engines, as a business, doesn’t recommend the solution that its own researchers found to be optimal for retirees.]
“Academics are good at developing tools for evaluating a problem. Industry research is good at evaluating what individuals want. So the ideal is to put the two types of research together. That’s what I try to do here,” Scott told RIJ. As an example of useful academic research, he cited the work of Harvard’s Brigitte C. Madrian and others in establishing the effectiveness of automatic enrollment in boosting 401(k) plan participation rates. “That kind of research has mushroomed,” he said.
Trade group research, Scott said, Echoing the thoughts of many in both industry and academy, Scott said that trade group research is great for providing data that academics can use. “Organizations like EBRI are able to identify trends,” he said.
The biggest problem, he said, is that “the two groups—academics and industry researchers—don’t talk to each other enough. Academics will solve elegant problems that don’t really help anyone. Practitioners will keep reinventing the wheel rather than learn from the insights and tools developed by academics. The situation is improving, in part because of the growing respect for behavioral economics and finance. That group of academics, at least, is being pulled closer to the practitioners.”
Potential conflicts
John Payne is a professor of psychology and of management and marketing at Duke University as well as a consultant on behavioral finance to Allianz Global Investors. He’s in a position to appraise the value of retirement research—and to consider the potential conflict that arises when academics do paid work for the industry or a particular company.
As long as industry-supported retirement research meets scientific standards—such as being open to testing by other researchers—Payne believes, it is perfectly valid as science.
“I think you have to judge retirement research in terms of the science,” he said. “Are you seeing results that you can imagine replicating? Whoever is producing the research—companies, academics, trade groups or whatever—if they reach that standard, then it is probably worthwhile. I tend to be a little less concerned about the source than the standards. That said, I think academic research, because of the norms of academia, tends to be clearer as to what people did and why they did it.”
He added: “I’m not naive. When academic research is funded by industry, there will certainly be some self-selection [of the researcher]. I think all academics who write for the journals should do full disclosure. For example, I always disclose that I am part of a behavioral advisor group for Allianz.”
Other retirement researchers consider it essential to remain at least arm’s-length from corporate funding. “When I do retirement or health research, it’s funded by a center run by the university,” said Olivia Mitchell, a professor of business and public policy at the University of Pennsylvania’s Wharton School, and director of its Pension Research Council and Boettner Center for Pensions and Retirement Research.
The Boettner Center receives corporate funding, Mitchell said, “but there are no specific funding ties” between its research and any company or product. “I feel very strongly that I do not want to lose my independence. I don’t do much private consulting.
If you do consulting for one group, you get identified with them, and I like to retain my independence. Typically I sign an agreement retaining my right to publish my research. That’s important. And every paper I publish acknowledges the support I received on the front page.”
Mitchell worries, though, that professors will increasingly rely on private support for research as government grants and other public support for pure academic research decline. “How do we set up agreements in the relationship between the academy and industry, so we can help each other, while we researchers retain the ability to publish freely?” she asked.
Theresa Ghilarducci, a professor of economics who specializes in retirement research at the New School of Social Research, says, “The whole question of research on retirement is an important issue.”
Last year, she noted, the American Economics Association established a conflict of interest standard for the first time. It did so after revelations that several prominent economists who testified at hearing on the Dodd Frank Financial Reform Act didn’t disclose that they received payments from some of the major banks.
“And there are still conflicts,” she said. “I can tell when people are funded by the industry, because they focus on financial literacy. They always end up saying, ‘Buyer beware.’ That kind of ‘research’ doesn’t add any value. It’s like looking under the lamppost, because that’s where the light is.”
‘Quasi-academic’
Paul Yakoboski, principal research fellow at the TIAA-CREF Institute, one of the oldest corporate research institutes on retirement issues, asked if the work his institute does is different from academic research.
“Yes and no,” he said. “We view our research as quasi-academic. Some of it is actually done through grants we give to academics. We want our work to be out there, published, so it has impact beyond our own organization. For example, we do a lot of research, understandably, on annuitization, and that research has implications for public policy makers. At the same time, if the research produces information that, at some level, has business relevance to us, we’ll tap it and make use of it.”
Firms that offer retirement products should separate their research work from their marketing work, Yakoboski suggested. “Obviously we each know what the other is doing. It would be foolish to have a wall with no communication going on, but we have distinct missions, and they shouldn’t be confused,” he said.
Leslie Prescott, the chief marketing officer at Thrive Income Distribution System, which sells retirement income planning technology, says that most financial advisors don’t pay much attention to or even read the work of academic researchers.
“They tend to read the trade magazines and the lower end research journals, not the peer-reviewed ones,” she said. “They’re reading Financial Planning, not things published by the Pension Research Council. That’s both good and bad. It’s not peer-reviewed, but it’s more understandable.”
The line between corporate research and academic research can get blurry at times, she said. For example, insurance companies and mutual fund companies consume and fund huge amounts of research, but they’re selective.
“Certain academics, based upon their research, focus on certain product areas, or highlight certain products. Naturally, companies that offer these products are anxious to support those researchers. It’s fairly widely known within the industry which researcher in academia supports which products,” said Prescott, who has a BA from Duke and an MBA from the Wharton School.
As for company research, she says, “There’s a lot of thinking that goes into their products’ design, and in terms of marketing to advisors. But when it comes to the internal research done by firms like Putnam or Vanguard, I’d assume they use that to support their own product lines and their own point of view.”
Regarding academic retirement research, Prescott said, “I don’t think there’s a lot of disclosure in this field. It’s just generally known in the industry who’s got what bias, so in a sense the lack of disclosure here is as serious as it is in medical research.”
Retirement research is tougher to evaluate than, say, pharmaceutical research. “In pharmaceutical research, there’s a clear standard: do people get better or do they die of the side effects? But it’s not so clear with retirement. And unless you have pretty clear standards for measuring success, it’s hard to judge the research,” she said. “You have research that’s very influential and very low quality, and research that’s very high quality but not very influential.”
© 2012 RIJ Publishing LLC. All rights reserved.
