We would like to thank the ICI for formally responding to BrightScope’s concerns about their recent fee study. We shared those concernsÂ in our recent blog post. We encourage our readers to read the ICI response in its entirety before proceeding with this post. One of BrightScope’s goals is to bring full transparency to the 401k marketplace and help plan sponsor fiduciaries and participants understand how to best utilize their 401k plans to enhance retirement outcomes. We believe that fee transparency is a very important part of that goal. We hope that by drawing out both sides of the debate we can all agree what needs to be included in calculations of total plan cost and what the magnitudes are for plans of different sizes.
ICI’s response makes it clear that there are still several major fundamental differences between how the ICI calculates fees and how its critics calculate fees which is evidenced by the gap between the ICI’s 0.72% median fee and the 3% fee number used by its critics. While the ICI’s response clarified some areas of concern, it does not resolve the most important issues.
The first major factor that BrightScope questioned with ICI’s fee study was its sampling techniques. In a survey aiming to benchmark 401k fees, it is critical that clear and accurate information about sources and methods is given to those reviewing the study in order for them to make a determination of its utility. It is disappointing that it took our blog post for these techniques to be disclosed. As it turns out the ICI used a nonprobability sampling technique in order to ensure that the survey contained a certain number of plans of each size. While this is a commonly used technique, researchers generally consider random sampling methods to be more accurate and rigorous. We will leave it up to the readers to determine if they think this is an appropriate methodology by which to capture information to compare 401k fees. Most of our thoughts about the resulting sample and its representativeness can be found in our previous blog post. However, we are pleased that we have a better grasp of their sampling techniques and hope that in future reports the ICI will disclose their sampling techniques directly in their survey report.
Are all Costs Included?
We have serious doubts about whether or not the ICI is truly capturing all the fees charged to the plans in its survey. In general within 401k plans there are two different levels at which fees can be charged; directly to the plan or through the funds in the plan. In the report the ICI claimed that the data for the study was gathered from “individuals at the plan sponsor and retirement service provider organization.”Â However, in their response to our blog post the ICI claims that the survey “captured fee information from a variety of service providers including mutual fund companies, insurance companies, banks and third-party administrators.”Â This is a very important distinction. While a mutual fund or insurance company may be aware of the expense ratios of its funds and the revenue sharing it pays to TPA’s and recordkeepers, it may be completely unaware of some of the other charges that come out of plan assets: custodial costs at the plan level, investment advisor fees, and CPA audit fees where the invoice is given to the trustee and paid directly out of plan assets. These are just three prominent examples.Â In order for the fee study to be valid all fees charged to the plan as well as all fees charged by the funds must be included. We would like for the ICI to clarify that all plan level fees were included.
It sounds as if the ICI largely agrees with its critics on the issue of transaction costs. Both sides recognize that fiduciaries should “consider the costs associated with a higher level of trading activity.”Â The major distinction that the ICI makes is that while they agree that trading costs should be “considered”, they don’t want these costs to be “measured.”Â I can see why critics of the ICI are frustrated. How can a fiduciary consider transaction costs without coming up with an estimate of how different levels of transaction costs impact participant outcomes? Doesn’t the participant outcome calculation necessarily require an estimation of what these trading costs are? Is the ICI saying that a consideration of these costs can occur without any knowledge of their likely magnitude?Â This is definitely an area that deserves more exploration and we encourage others to share their opinions on this issue much like Kevin Price has done at the Float blog.
In our previous blog post, in order to demonstrate how adding trading costs on to total plan costs can help reconcile the fee estimates of the ICI and its critics, we used the fund’s expense ratio as a crude proxy for trading costs. This technique is based loosely on the findings of a paper by Evans, Edelin and Kadlec. It is important to note that BrightScope does not recommend fiduciaries use this methodology when estimating trading costs.Â In our original post we called this shorthand technique a “very crude calculation.”Â But, while it is crude, it does help illustrate how large an impact trading costs can have. It is interesting to note that the ICI does not argue that the magnitude of trading costs are lower than the estimates in the academic literature; they just want a more nuanced analysis to be done that takes into account different asset classes and turnover rates.
We would like to take this opportunity to present our current method for calculating transaction costs that we think provides the level of detail that a plan fiduciary needs to estimate the impact of trading costs on their participants. You can read about our Transaction Cost Algorithm by clicking on the image above. We are not holding out this algorithm as an exact measure of these costs, but rather as a tool to enable plan sponsor fiduciaries to estimate these costs so that they can properly “consider” how they impact retirement outcomes. The algorithm takes into account a fund’s disclosed brokerage commissions, its asset allocation based on its quarterly filings, its disclosed turnover rate and the fund’s total net assets. Details about how each of these factors impact our calculations can be found in the attached document. Unlike the ICI, we do not think that “considering” can occur without “measuring.” We have shared the trading costs estimates derived from our algorithm with several large fund companies in order to solicit feedback and are open to sharing our results with others if they are interested. We encourage and appreciate any feedback the ICI or others have on our methods. Our intention in sharing this algorithm is to help advance the dialogue by overcoming ICI’s complaint about using “crude” estimates.
Matt Hutcheson and Other “Critics”
The ICI’s primary complaint about our response is that we attempt to reconcile their study with a commentary by Matthew Hutcheson.Â They say that such a reconciliation is “impossible.”Â Let us take this opportunity to revisit how Matt Hutcheson got involved in this discussion in the first place. In John Murphy’s speech, he compared the ICI’s median fee of 0.72% to the 3% number used by “critics”:
“The bottom line? The median all-in fee for these plans was 72 basis points.Â Less than three-quarters of 1 percent-quite a bargain compared to the 3 percent that some critics cite.”
However, perhaps because of an oversight, John Murphy failed to mention these critics by name. BrightScope’s goal was to shed more light on the comparison that John Murphy made between ICI and its critics. In an effort to further this debateÂ I suggested that Matt Hutcheson might be one of the individuals that they deem a critic, but at the time I had no clue if that was indeed the case.Â Based on the amount of time the ICI spent picking apart Matt’s work in their response, and the fact that they did not mention any other critics by name, I think it is safe to assume that Matt is indeed the critic they are referring to. If this is the case then it is the ICI, not BrightScope, who has decided that Matt’s number can be compared to their numbers even though they now argue that this comparison is “impossible.”Â Thus their arguments are either illogical or they need to name another critic so that we can continue the real debate about why two responsible parties can end up with numbers that are so staggeringly different.
Perhaps we can propose an additional critic here. Barclays suggests “that plans at smaller companies using mutual funds are averaging around 2.5â€“3.5% in total expenses.” This is in direct conflict with ICI’s claim that the 90th percentile of fees in micro plans is 2.2%.Â It is unclear if Barclays uses the same plan size categories as ICI.Â What makes this particularly disconcerting is that Barclays is an ICI member. We find it difficult to reconcile Barclays’ estimates with those of the ICI.
Can a Single Number be Representative?
The ICI will use their 0.72% number going forward as evidence of low fees but for those millions of participants in small plans this number is a gross underestimation and as much as we’d love to think otherwise those millions of participants aren’t going to read far enough into the ICI study to figure out that according to the ICI they have a 10% chance that their fees are over 2.2% (using ICI’s own numbers). The more we listen to both sides on this issue the more we realize that any single number used to describe 401k fees, be it asset-weighted, plan-weighted or participant-weighted is not all that useful in trying to help participants and plan sponsors understand what fees they are paying within their plan. According to an AARP study from 2007 if you ask 100 participants what they are paying in fees within their 401k plan over 80% of them do not know.Â Worse yet, two out of every three participants in the same study responded that they pay no fees.Â Whatever fees participants are actually paying, it is absolutely critical for participants to know what those fees are and more importantly what it means for their retirement.Â We can’t think of a single other marketplace where fees are not disclosed and the industry itself has to commission studies to figure it out. The fact that the ICI had to commission a study to understand fees should be enough evidence that more fee disclosure is desperately needed.
Transparency is the Solution
We think the ICI and its critics could go back and forth all day long about fees in 401k plans, because the bottom line is that the average 401k participant still doesn’t understand what they are paying and now more than ever they lack trust in the financial system. The best way we can think of to restore trust in the system is to provide full fee transparency. If the ICI is so confident that fees are indeed low and reasonable, the best way for them to end this debate is to encourage plan level fee transparency as represented by George Miller’s HR.1984. This is the one surefire way for the ICI to silence its critics. We call on the ICI and its members to support fee disclosure legislation so that everyone knows exactly what they are paying.Â Now that is something I believe we can all get behind.