The Standard

Benchmarking 401(k) plans has gone through a metamorphosis of sorts over the last few years—transforming from a process fixated on plan-to-plan comparison to one that is also internally focused on successful participant outcomes. Sponsors have come to realize that knowing what other 401(k) plan sponsors do is not the be all and end all of measuring a plan’s success. Instead, a focus of benchmarking has moved to seeing what works and what does not work for participants, says Douglas Prince, Managing Director for Stifel Nicolaus in Indianapolis, Indiana.

However, while the movement is away from focusing only on participation rates, intra-plan comparisons will not be going the way of defined benefit plans. Plan-level benchmarking is not declining, but drilling down data analysis to the participant level has been added to the benchmarking process, says Prince. Plan sponsors still need to benchmark against other plans to see how their benefit packages stack up compared with competitors’ benefit packages, says Jennifer Flodin, Chief Operating Officer of Plan Sponsor Advisors in Chicago, Illinois. Benchmarking against other plans also allows plan sponsors to see whether they are intelligent buyers of services, notes Mike Alfred, the Chief Executive Officer of Brightscope, Inc., in San Diego, California, which provides data, tools, and analytics to help advisers and sponsors benchmark participant outcomes.

For years, the main focus of benchmarking was basic plan design, with plan participation rates being perhaps the most widely used standard of benchmarking, says Patricia Advaney, Senior Vice President of Participant Solutions for Diversified Investment Advisors, in Purchase, New York. However, just looking at plan-level metrics, such as participation rates, is too simplistic, she says. A plan may have a 98% participation rate, she says, but, if everyone is only contributing 2%, then the plan needs to use other measures to determine whether the plan is serving employees adequately. Furthermore, with the adoption of automatic enrollment and automatic increases, participation and deferral rates may not be as important a focus of benchmarking as before, adds Flodin.

Another problem with the historical method of benchmarking is that it was really only considering past data, and did not focus on the long-term success of the plan, says Prince, such as savings rates and replacement ratios. Previously, there was no measurement of how employees were saving or whether they were saving enough, he says. Looking at savings rates is a start, but trying to quantify how the plan is meeting long-term objectives needs more analysis, says Prince. For example, one client of Prince’s puts in a 4% profit-sharing contribution and has a 50-cent match on the first 4% of pay. The client’s overall philosophy of the plan is for a 30-year employee to have an adequate amount saved for retirement. When you reviewed participation numbers and savings rates, they appeared to be adequate, says Prince. After looking at whether employees would meet the long-term objectives, however, the results were not as good.

Earlier benchmarking data also usually lacked data on the fee side, which sponsors really need to determine the success of their retirement programs, says Ryan Alfred, President of Brightscope, Inc. The Department of Labor has made it clear that sponsors and fiduciaries do not have to select the lowest fee, but they do have to determine if their fees are reasonable. The problem, he says, is that plan sponsors frequently cannot adequately determine what is “reasonable,” because the data was not available. Brightscope now offers data about plan fees through its plan management dashboard and other products, and that will make the plan better for participants, says Ryan Alfred.

A New Approach

Tides shifted and now sponsors are doing more of an internal evaluation of the plan than an external one when benchmarking, says Prince. The new standard of benchmarking includes seeing how your plan is working for your employees, as opposed to just comparing it with other plans, says Prince. The main questions plan sponsors are asking, he says, are “how can we help our employees do a better job?” and “what can make our participants have a successful retirement?”

With the 401(k) becoming the primary retirement vehicle, it is essential to make sure it is doing right by individual participants, says Mike Alfred. Nowadays, Diversified advises sponsors to look at their own situations and determine where they want to be, rather than compare themselves with other plans, says Advaney. While comparisons with other plans are useful, she says, plan sponsors need to measure where they are today and target objectives based on those starting points. A plan could feel good that its participation rate of 55% is better than an industry average of 50%, but maybe a plan’s demographics have something to do with that. “Why stop there?” she asks. “Why assume that is good or admirable? Why not see if they can push 55% to 60%? You have to think about the individuals that are being ‘left out’ or ‘left behind’ rather than always focusing on aggregate plan measures and feeling good about them, when there’s still something left on the table.” The goal is to determine how successful the plan is given the plan sponsor’s objectives, she says.

Benchmarks Beyond Participation

It is not uncommon now for plan sponsors to benchmark a plan by breaking down data to the partic­ipant level. Whereas, in the past, the sponsor may have looked at overall participation rates to see how they compared with other plans, now the sponsor may break those rates down by age or salary within its own plan, says Prince. Prince’s firm has several recordkeepers that help them break down this data but, generally, the firm downloads census and account information and runs the breakdowns itself. Benchmarking has moved to more of a participant outcome-based approach, says Ryan Alfred.

Another focus is participant success in achieving adequate retirement income, by tracking data such as average replacement ratios. Sponsors, for example, are tracking participants to see if they are saving enough to replace 80% of their pre-retirement income in retirement, says Advaney. This is a better measurement of plan success than traditional measures such as plan participation, she says, because it is more individualized. For example, Diversified reports a retirement outlook to participants every time they go online. In addition, Diversified sends out annual statements with similar information. Finally, it reports this information to plan sponsors at an aggregate level, breaking it down by age, salary, and gender, so a plan sponsor can see which segments of the population are most in need of extra support (see “Zooming In,” PLANADVISER, May-June 2009). These second-generation measures are superior to traditional measures, says Advaney, because they take into consideration several different factors, such as age, salary, current savings, to determine how an individual is doing relative to an individualized goal. 

For example, if a worker is 40 years old and needs to have one year’s salary saved to be on track to having a successful retirement, the focus of benchmarking will be seeing if the worker does indeed have one year’s salary saved. If the worker has not met that goal, the next step would be to target communications to that participant to see if the sponsor can get them to that one year’s salary goal. Employers also could use new data to target participants who are not investing properly, says Prince. For example, if George Smith is 100% invested in mid-cap equities, the sponsor can target information to him to persuade him to invest more responsibly.

Data also can be manipulated in new ways to yield even more information. Sponsors can look at data on the participant level and roll it back up to the plan level to look at individual segments—e.g., age, job, etc.—to see how each segment is meeting replacement goals or investing, says Advaney. In this way, sponsors, says Advaney, can see which specific demographics in its employee population need extra attention, and target communications to get to that group. For example, sponsors can view data on how people are invested and, if a certain segment has not diversified investments properly, use that information to communicate to that group to encourage participants in that segment to be more diversified.

Retirement plan benchmarking continues to evolve, says Prince. Technological advances allow advisers and sponsors to view data more easily and to dissect it in new ways. Right now, he says, sponsors may be able to identify a problem, such as workers not saving enough, but can do little about it. In the future, sponsors will be able to break down the data to the point where they can see that Patty Miller is not saving enough and, over the next five years, here is how we can communicate to Patty Miller in a way that will get her on track.

Additionally, the next big focus of benchmarking will be on distributions, predicts Prince. The industry, until recently, focused on the accumulation phase, but analysis down the line will be on how to help employees de-accumulate successfully.

While the role the adviser takes on in benchmarking depends on the adviser, Stifel Nicolaus takes on the role of quarterback, says Prince. “We help the sponsor figure out what to look at and what data to collect,” he says. Recordkeeping organizations are starting to look at this information as well, says Prince. Stifel also uses actuaries to help calculate information. The firm’s role, says Prince, is to be the Chief Retirement Officer for the client. “We are basically a part-time employee of the organization. Our role is to direct the flow of information. If the current providers cannot get us the data, we will help get the project completed,” he says. Moving into the future, he adds, he foresees Stifel doing more and more of this work so that it knows the assumptions being used and has a consistent method of understanding the information and the actions needed. “If we rely on many different vendors, the answers take too much time to analyze versus just being able to understand the same process and methodology each and every time,” he says

Particularly in the micro to mid-size plan market, advisers have a role in insuring that their clients are meeting Employee Retirement Income Security Act (ERISA) fiduciary standards, and benchmarking helps them to see if they are, says Dick Davies, the Head of Product Strategy for AllianceBernstein Defined Contribution Investments in New York. Does the client want to be a best practice plan or just “good enough” to comply with regulations, asks Davies. Advisers should determine the client’s objective. 
Furthermore, plan sponsor data is only as good as the people who collect, input, and analyze it, says Flodin. The quality of the data depends on who is polled, which can be anyone from HR to treasury to risk management, and how well-versed they are on the topic. As an example, if the benchmarking study is regarding fees, her experience is that a CFO or treasurer or SVP of HR would have a better response than other respondents. If the benchmarking data is more specific to plan design or administration, then there are other roles that would produce more accurate responses than those mentioned earlier. “[Data] should not be used as the be all and end all,” she says. How benchmarking data is used should depend on the source of the data, she adds.

“A lot of what benchmarking is used for is philosophical,” says Davies. A client that views its 401(k) plan as a way to attract and retain qualified employees is more apt to want to benchmark to other 401(k) plans. Davies says he has one large client that, up front, says it does not want to have the best 401(k) offering. That client is less interested in knowing participation, contribution, and deferral rates of other employers. Other clients, he says, do care about having best practices, so it is important to them to benchmark to see how their plans rate compared with competitors.