What Retirement Plan Fees Really Cost You
Ask a participant what they pay in retirement plan fees and most cannot answer. In my experience reviewing retirement plans since 2009, almost nobody can. The data point in the same direction. A 2021 GAO report found that almost 40 percent of retirement plan plan participants do not fully understand and have difficulty using the fee information plans are required to disclose. The same report found the gap runs deeper than comprehension. 45 percent of participants cannot use their disclosure to figure out what their investment fee actually costs, and 41 percent do not know they are paying a fee at all.
The disclosures are not confusing because people cannot read. They are confusing because they speak a language nobody uses to think about their own money, an expense ratio expressed as a fraction of assets. Nobody budgets in expense ratios. Everybody budgets in salary and in monthly bills. Until fee disclosure speaks one of those two languages, it will keep failing to register.
Why the fee stays invisible
A 2021 Forbes analysis by Chris Carosa goes looking for the reason behind the GAO's numbers and finds the same structural blindness from several different angles. Fees are either absent from the pay stub entirely, or participants assume the employer bears them because a retirement plan reads as an employer run benefit. Financial advisor Hannah Whatley put it plainly in the piece. Participants do not feel the pain of paying the fee from their bank account, and with expense ratios they do not even see the fee come out of the retirement plan account, because the cost is baked into the fund before the participant ever sees a return. Carosa's own summary of the pattern is the line worth remembering. If you do not see it, you do not feel it, and if you do not feel it, does it really exist. That dynamic is a large part of why the GAO comprehension numbers look the way they do.
Employers are not much better informed. A 2012 GAO survey of retirement plan sponsors found nearly six in ten expressed high confidence that their service providers had fully disclosed all fees, yet half admitted they did not know whether they or their workers paid the investment management fee, or believed incorrectly that it had been waived.
The Forbes piece frames fees as invisible in a way taxes are not, and that comparison needs a correction. Taxes are not exactly in your face either. Most people never look at the withholding line, they look at the number that lands in their account. The reason taxes still register at all is not that anyone reads the deduction, it is that the number is sitting there in black and white on the stub if a person ever decided to look. Retirement fees do not get even that. They are not withheld from a paycheck, they are skimmed from a fund's return before the participant ever sees the number, which is a full step more invisible than a line item nobody reads.
Milton Friedman played a significant role in building paycheck withholding into the tax system. He worked on it as a Treasury staffer during World War II, and he spent much of his later career regretting it, telling an interviewer decades later that he had no apologies for the wartime necessity but wished withholding could still be abolished. His specific objection was that when people never see the money in the first place, they never register how much of it there is. Adding fees as a line item on the pay stub, the way taxes appear, would be a real improvement over the current setup, since at least the number would exist somewhere a participant could find it. But a line item is still a withheld number, not a bill. The only version of this that actually forces attention is the one nobody has built, a fee a participant has to write a check for.
The two numbers that break through
Fixing this does not take more disclosure. It takes disclosure in a form people already know how to use. Daniel Kahneman and Amos Tversky, the two psychologists whose work on decision making later won a Nobel Prize in economics, established decades ago that how information is presented can change the decision it produces, independent of the information itself. A TIAA Institute study of university retirement plans found that effect operating inside actual retirement accounts. Faculty members with identical total compensation saved noticeably less on their own when a larger share of their pension contribution was labeled as coming from their own salary rather than from the employer, even though the dollars were fungible either way. Same total pay, different label on part of it, different saving behavior.
Salary comparisons carry a second, sharper effect on top of framing alone, and it explains why people notice a raise that comes in at 3 percent when they expected 5. Kahneman, working with Jack Knetsch and Richard Thaler, ran a series of fairness surveys in 1986 that asked people to judge employer decisions about wages. An employer cutting an existing worker's pay was judged unfair by most respondents. An employer simply not raising that same pay, or hiring a new employee at a lower rate, was judged acceptable. The finding, part of the research that helped establish loss aversion as a durable feature of how people process money, is that a loss from an existing reference point registers far more strongly than a gain that simply fails to arrive. A smaller raise than hoped for is a foregone gain. Money taken from what a participant already has is a loss, and losses are the ones people act on.
That distinction is why fee disclosure should describe a fee as a pay cut rather than stop at a neutral percentage of salary. This is a proposed frame, not an accounting identity. A fee does not appear on a W-2 as a reduction in pay. But it behaves like one closely enough that the comparison earns its keep. A fee is not a smaller raise than hoped for. The money was in the account, and then it was not, which is a loss in exactly the sense Kahneman, Knetsch, and Thaler measured, not a disappointment about a gain that never showed up. Told that a fee equals 1.2 percent of salary, a participant has to do work to feel anything. Told to think of the fee as a 1.2 percent pay cut this year, the participant is standing on ground they already know how to defend, the same ground they stand on every time a raise comes in lower than expected, just with the emotional dial turned up from disappointment to loss.
The second number, a dollar figure billed monthly, works through a different but complementary channel. The National Academies' review of the savings literature attributes much of the shortfall in retirement decision making to limited attention, people simply do not have the bandwidth to work through the consequences of an abstract percentage no matter how it is framed. A monthly number sidesteps that problem entirely by behaving like every other recurring expense already sitting in a household budget, a car payment, a phone bill. It does not ask for interpretation. It asks to be compared to other numbers a person already tracks without effort.
What the two numbers show
Apply this framework to real Chicago-area plans and the numbers stop being abstract. Salary figures below are blended estimates built from BLS wage data and industry compensation benchmarks for each plan's field, not each company's actual payroll, since Form 5500 does not disclose individual pay. A small personal injury law firm with a 1.14 percent fee rate works out to $351 a month per participant, functioning as close to a 3.8 percent pay cut against that blended estimate. A radiology group with a far lower fee rate, 0.33 percent, still comes out to $570 a month, though against physician pay that lands closer to a 1.2 percent cut. A fence contractor comes out to $282 a month, close to a 4.8 percent cut against a blended construction wage. A grocery wholesaler comes out to $104 a month, still a 3.3 percent cut against a modest retail salary.
Two participants at the same fee rate can be living through very different sized cuts depending on how much they have saved relative to what they earn. A 30 year old with roughly a year's salary in the plan and a 55 year old with six or seven times salary saved are paying the same percentage of assets, but the 55 year old's pay cut, converted to salary terms, is several times larger. Nobody explains that gap to either of them today.
Across a dozen or so of these plans, spanning law, medicine, manufacturing, retail, and the trades, the fee rate and the pay cut move independently of each other. A plan with one of the lowest fee rates in the group can still produce one of the largest cuts, because the rate alone says nothing about who is being charged relative to what they earn. An expense ratio hides that. A monthly dollar figure and a pay cut expose it immediately.
Where the average breaks down
Two plans in the dataset are worth walking through because they show the limits of even this framing.
The first is a small mold fabrication shop that has operated since the late 1970s, run under a new comparability profit sharing formula, a design that concentrates employer contributions toward specific participants rather than spreading them evenly. Its headcount fell from 66 participants to 28 over the past eleven years while plan assets nearly doubled. Divide the plan's total fee evenly across every participant and you get an average of $286 a month. That average almost certainly describes nobody, since the owners and everyone else are not accumulating anything close to the same balance.
The IRS caps how much any one person can receive from a plan in a given year, and that cap makes it possible to work backward. Assume the plan's two owners have been maxing out every year, subtract what that would take, and the remaining participants turn out to hold 63 percent of plan assets rather than an even split. Their actual share of the fee comes closer to $194 a month, not $286.
The second is a small family owned manufacturer, running since the early 1980s, with a founder and his adult son both active in the business. Backing out both of their contributions under the same maxed out assumption, including the age fifty catch up the founder has been eligible for through most of the plan's history, the remaining participants hold 66 percent of assets, and their fee share comes out to about $213 a month.
Neither $194 nor $213 comes from a document that discloses individual participant fees, because no such document exists. Both are estimates built backward from the pieces a Form 5500 filing does disclose. What these two plans show is not that the framing is wrong. It shows that a plan average can miss what a real participant experiences by a wide margin, and the framing only tells the truth once it is pointed at the right group of people.
What participants would actually do with the number
Seeing both numbers changes more than awareness. It invites a question nobody currently has reason to ask, which services am I actually using for this.
Cerulli Associates found that 71 percent of pre-retiree retirement plan participants have not sought advice or planning help from their plan provider in the past year, and a separate Cerulli finding put the number even higher across all active participants, 63 percent have no financial advisor relationship on their retirement plan at all. There is no equivalent survey specific to participants under 50, but the advisor side of the relationship points the same direction. A 2024 InspereX survey of financial advisors found that only 18 percent of their clients are under age 50, with 59 percent in their sixties or older. Advisor attention concentrates on older, wealthier clients as a matter of practice economics, which means whatever limited engagement Cerulli found among pre-retirees is very likely the high water mark, not the floor, for the plan as a whole.
The advisory fee embedded in most small plan expense ratios is paying for a relationship the large majority of participants never use. On its own, that is a straightforward waste. It gets worse once the advice itself is examined, because the two most common ways a plan advisor gets paid both build in a reason to steer participants away from what might actually be their best move.
A broker paid through commissions or revenue sharing operates under Regulation Best Interest, which requires the recommendation to serve the client's interest, but the standard does not eliminate the underlying economics. Commission based compensation still creates an incentive for the broker to only recommend investments that generate compensation, since firms build their shelf of recommendable funds around what compensates them, best interest standard notwithstanding. An advisor paid as a percentage of assets under management carries a different but comparably real conflict. The Consumer Federation of America laid this out in a framework paper on adviser conflicts of interest, noting that an AUM based adviser may have an economic incentive to avoid recommending anything that would shrink the managed account, including paying off debt, since the adviser's own compensation falls whenever assets under management do. Kiplinger's coverage of AUM fees makes a similar point from the other side of the relationship, flagging the conflict that can arise when a client asks whether to redirect invested assets toward debt. For a participant carrying high interest credit card or personal loan debt, the mathematically correct move is often to pay that down before maximizing a retirement contribution, a recommendation an AUM based advisor has less financial reason to make than a flat fee advisor does.
None of this is easy to see while the fee stays invisible. It gets much easier once a participant can see $282 a month next to a decision about whether they have spoken to their advisor even once this year, and harder still to ignore once they understand that advisor's compensation may run counter to their own best financial move. A sponsor facing the same two numbers is in an even stronger position, since a plan's advisor is a service provider the sponsor can replace, not a fixture. A 408(b)(2) invoicing requirement already pushes sponsors to see that relationship as an actual dollar cost rather than a bundled percentage. Extending the same idea to participants, a statement that leads with a monthly dollar figure and a pay cut side by side, itemized by service, delivered the way a bill is delivered rather than buried in a once a year notice, gives the people actually paying for that advisor a reason to ask whether they are getting anything for it, and whether what they are getting is even built to serve them.