How Good Is Your 401(k)?

The first big hurdle is getting the rating formula right. If BrightScope's data aren't good enough, or its model proves defective, its analysis could be useless, or could even nudge both participants and corporations to make decisions based on faulty assumptions. "It's not as simple as a good plan vs. a bad one," says Lori Lucas, who heads Callan's practice in 401(k)s and other defined-contribution plans. "A mutual fund might have 100 moving pieces. But a 401(k) might have 1,000, making it difficult to come up with a single rating."

On top of these challenges, BrightScope faces other small competitors trying to make a business out of evaluating 401(k)s. Fiduciary Benchmarks in Portland, Ore., tries to gauge retirement readiness based on a more complex income-replacement model—that is, how much of your current income you will need in your post-working years. This includes factors such as estimated payments from traditional pensions and Social Security. The company markets this information and other detailed reports to retirement-plan advisers. London-based PensionDCisions, meanwhile, is gearing up to enter the U.S. market with data on participant behavior within particular 401(k) plans—everything from whether they are using the defaults or choosing their own investments to how often they trade in and out of investments. Neither of these efforts, however, is easily accessible to the general public.

Then there's Morningstar. The firm does not rate 401(k)s and doesn't plan to. But Morningstar Vice-President for Research John Rekenthaler says: "It feels like something we could be doing....It seems like a valuable service."
LIGHTBULB MOMENT

The creative energies behind BrightScope belong to two brothers, Mike and Ryan Alfred, 28 and 26, and a middle-aged techie friend of their family, Dan Weeks. A former Hewlett-Packard (HPQ) engineering manager, the 50-year-old Weeks hatched the idea three years ago after becoming increasingly concerned about his own retirement planning. "I could not make sense of Hewlett-Packard's 401(k)," Weeks says. "Everyone I knew at HP was deciding randomly where to invest."

Neither the Alfred brothers, who were then running a financial advisory business, nor Weeks knew much about 401(k) plans when they sat down to discuss the idea in the fall of 2007. One evening in October, with Monday Night Football playing in the background and glasses of Maker's Mark in hand, Ryan says, a light went off. That's when they realized it was the 401(k) plans themselves that they should be rating. "You know when you start thinking faster and you start talking faster," he recalls. "I started taking notes on my BlackBerry, and soon had five pages of thoughts."

In 2008, Weeks quit his job, and the brothers, who had begun as consultants on the project, signed on full-time. A third brother, Eddie, an undergraduate at Harvard, took a leave of absence to join them. They lined up a first round of investors for about $1 million and raised double that amount in a second round in 2009.

BrightScope hired a team of data entry workers to plug in information on hundreds of plans a day and launched its public Web site in late January, 2009. In July the company launched a product called Plan Management Dashboard for corporate plan sponsors and followed up in September with Advisor Central for retirement-plan advisers. Matthew Hutcheson, an independent fiduciary who works with 401(k) plans and is on BrightScope's advisory board, says he has purchased the Dashboard for all his clients. "BrightScope is chipping through the rock to find this valuable information," says Hutcheson, who says some clients have made changes to their 401(k)s because of the data.

Yet while BrightScope's efforts at rating 401(k)s look a lot like Morningstar's in funds, it's a much more difficult exercise, in large part because the publicly available data is problematic. Companies with 401(k)s are required to file paperwork, called Form 5500, with the Labor Dept. It details the size of the plan, the number of participants, the investments available, and the like. As with other types of government data, though, the information on those forms may be as much as two years out of date. It's also riddled with inconsistencies on basic matters, such as which industrial category a company selects to classify itself. And while consumers can easily look up the symbol of a retail mutual fund, that's often not possible with institutional offerings, which are investment vehicles geared toward large investors and which administrators of the best 401(k)s favor because of their lower fees. "The data is hard to get, and you have to make assumptions," Ryan says. "There's a subset of plans where even if you have the data, they are unratable because there are too many gaps."

BrightScope hopes that as its ratings gain acceptance, those in charge of the 401(k)s will fix errors or misunderstandings that arise from their filings. This happened soon after launching the Web site last January. "We were about to go to the launch party, and Chevron (CVX) e-mailed to say that they believed they had the best match in the industry. They were filing something that no one else files that way," Ryan says. A few tweaks of the data and the information was fixed, giving Chevron a much better rating.

In the long term, rather than waiting for government information, BrightScope would like to obtain data feeds from the companies or the firms that are responsible for managing 401(k) plans on their behalf. Given the tight control that large recordkeepers like Fidelity and Vanguard keep on data, and the security issues involved, that's unlikely to happen on a large scale any time soon. But Invest n Retire, a recordkeeper based in Portland, Ore., is now sending its clients' data to BrightScope on a trial basis. "I hope BrightScope will get momentum," says Darwin Abrahamson, chief executive at Invest n Retire. "I've looked at some plans and said, 'My God, I cannot believe it had such great matching, yet had low account balances, high fees, and low investment quality.'" If the company is going to spend the money on matching, he says, "it needs to correct the other problems, too."
NAILING THE NUMBERS

But what makes a retirement plan good or bad? And how do you account for so many different factors and for the wildly divergent demographics of companies, large and small? BrightScope started with two simple premises: The rating would be based on quantitative factors only, and it would be designed to measure how well the plan was creating retirement income security for its participants. To turn those ideas into a metric, Ryan—with the help of academics like Allan Timmermann, a professor of finance at University of California at San Diego—built a model that simulates "retirement readiness." They fed into it some 200 variables, which fell into two categories: items that are within the sponsor's control, such as the vesting schedule, the eligibility period, the available investments and their associated fees; and items that are up to the employees, such as whether to participate and how much to contribute.

The team then ran some 10,000 simulations of a hypothetical 44-year-old employee's retirement readiness. This analysis gets boiled into the BrightScope rating. Today the highest-rated plan with assets of more than $1 billion, Saudi Aramco, rates a 93; the lowest, one of the plans of manufacturing conglomerate Danaher, scores just 26. "A 100 rating would be, 'You're retired, here's $1 million, and thank you for working for us,'" Ryan says.

The trouble is, just the act of creating a rating raises issues about what is being measured and how. Some retirement experts question BrightScope's decision not to account for retirement benefits outside the 401(k), such as pensions or profit-sharing plans. Intel's 401(k) has a ho-hum overall rating of 75, for example, partly due to low marks for company generosity. Yet Intel (INTC) makes hefty contributions to a separate profit-sharing account, says Stuart Odell, Intel's (INTC) assistant treasurer of retirement investments. All employees receive this contribution, even if they don't sign up for the 401(k), and it has never been lower than 6% of pay—far greater than the typical company match in a 401(k) plan.

Another potential pitfall: BrightScope's list of top 401(k) plans is dominated by industries such as oil, financial services, and pharmaceuticals, where the workforce is well-paid and stable. Higher-paid employees are more likely to participate in their companies' 401(k)s and to contribute at higher levels, both of which are factors that impact the BrightScope rating. Stephen Utkus, head of retirement research at Vanguard, calls the BrightScope analysis "a useful framework," but notes: "If you are ranked by BrightScope with a 60 or 70 and think you have a mediocre plan, you may not. You may just not work in a high-paid industry."

What BrightScope lacks in nuance and complexity, however, it gains in accessibility. Keeping the ratings user-friendly is one of BrightScope's strengths—and it will remain a priority, says Mike Alfred: "I think we're at the early stages. Who's to say we couldn't rate pension plans or health plans?"

Feldman is an associate editor with BusinessWeek in New York.