Skip to main content

You are here

Advertisement

Compounding the Problem(s)

Service Providers

Have you heard the one about $1.35 trillion in “forgotten” 401(k) accounts? 

That jaw-dropping number was the headline from a report titled, “The true cost of forgotten 401(k) accounts,” authored by “the Capitalize research team”—and yes, if accurate, that would mean that about a fifth of all 401(k) assets have been “forgotten.” Sound suspicious? Here’s another data point: The report goes on to estimate that these 24.3 million accounts that have been “forgotten” have an average balance of… $55,400 per account

Now, numbers like that are generally reserved for emails regarding a Nigerian prince. Fortunately, these authors show their “math.”

They start by citing 2018 data from Vanguard that suggests that 4.8% of participants changed jobs, but left their account behind (this is actually 48% of the 10% that terminated employment that left their accounts at their former employer’s plan). They then take that 4.8% and apply it to the 55.8 million participants active participants per the DOL’s 5500 database to conclude that there are 2.8 million “additional forgotten 401(k) accounts each year.” That’s right—not only do they claim that there are 24.3 million “forgotten” 401(k) accounts—but that another 2.8 million are added to that tally each year

Attrition Attributions

We’ll set aside for a minute that there might be a difference between an account consciously “left behind” and one that is “forgotten,” but having established this baseline, the authors say they connected with the folks at the Center for Research at Boston College to concoct something they call an “attrition rate.” This, they say, is the percentage of these “forgotten 401(k) accounts” that will be “shut off and/or liquidated,” apparently due to the death of the account owner, or the eventual transfer of the account assets (we’ll also set aside for a moment the realities that both of those events would likely mean the accounts would go to a designated beneficiary). 

Now, there are references to the origins of this assumption—basically they say it derived “based on our consultation with the Center for Retirement Research” and then go on to say that it was “inferred based on a comparison of GAO and Form 5500 records” for 2004-2013. But basically it seems to be “inferred,” And that matters, because it’s an imbedded assumption that underlies the math that is used to make their point on the way to creating their conclusion regarding 24.3 million accounts—that is, in turn, used to “compound” the impact of the assumptions in a whole new “magic” of compounding.

Mathematical Alchemy

But where the mathematical alchemy really enters a whole new dimension is their assertion regarding the average balance of these “forgotten” accounts. They go back to the Vanguard data—this time looking to the assets, rather than the accounts, apply that percentage to Vanguard’s total DC account balance to get an overall average account balance ($92,148)—and then, for some reason determining that Vanguard’s focus on larger plans means that those average balances would also be larger—backed it down to the aforementioned $55,400 as an average balance for these forgotten accounts—and, lest we miss the point, bold that statement. Oh—and despite the “general increases” 401(k) accounts between 2018 and 2021, they decided to let that number stand… “out of conservatism,” though arguably that ship sailed two assumptions ago. 

Account ‘Balance’? 

That said, apparently being “forgotten” isn’t the worst of it—being forgotten in a 401(k) comes in for criticism as well. The paper points out the mathematical damage based on an alleged fee gap between 401(k) accounts and IRAs to cement their argument, and if that weren’t bad enough, comments that (back to Vanguard data) approximately 13% of those employers (still) select a money market or stable value default—yet another reason to favor remembering to put those “forgotten” balances in an IRA (where, the paper notes, again in bold, that “over the past three years, managed IRA (robo-advisor) portfolios have averaged 8.8% annual returns for retirement accounts after fees, though some averaged over 11% in that timeframe.” The point seems to be that these 24.3 million accounts with an average balance of $55,400 are (apparently) languishing in a 401(k) somewhere when they could be prospering in an IRA.

And if that weren’t reason enough, the authors also would also “spare” employers the bother of keeping up with these “forgotten” ex-employee accounts, empathizing that “While these costs are hard to quantify, they are real and add meaningfully to the direct per participant costs incurred by an employer for forgotten 401(k) accounts. What’s worse, these benefits are provided to former employees and divert dollars that could be invested in improving benefits for existing employees,” Oh, and for good measure, they toss in the potential litigation liability these accounts represent, as they might wind up as participants in a class action suit. Once again, they pull out their compounding magic wand, conjure up an assumption of excessive fees, assume an average job tenure of 5 years, and—voila!—conclude that a plan with 500 participants “might end up paying close to $2,500 per participant in a settlement, or about $1.25 million total.”

Source Spots

Now whenever I see some kind of data point or factoid that is both jaw-dropping and intuitively nonsensical, I make it a point to check out the source—and as it turns out, the folks at Capitalize are in the business of capturing IRA rollovers—for “free,” though they acknowledge they “might get compensated by the provider.”[i] And, let’s face it, if you’re in the business of mining rollovers, this report certainly makes that seem a lucrative market.  

Don’t get me wrong; “leakage” is a real concern, and rollovers, for the most part, remain a tedious process for your average participant. Too many smaller (and perhaps some larger) balances do, in fact, get lost or overlooked, and “attribution” via escheatment or force outs does occur. That said, reputable research organizations like the Employee Benefit Research Institute (EBRI) have put some structure around not only the size, but the scope of what’s being “left behind”—they too come up with figures in the $1.5 trillion range—in fact, the casual reader might well make that connection. But EBRI’s projections come from taking every possible rollover transaction (not just the ostensibly “forgotten” ones) and compounding their impact… over a full 40 years.[ii]

Ironically, the white paper concludes by noting that, “By shining a light on the size and cost of the problem, we hope to encourage industry stakeholders to pay closer attention to what we believe is one of the largest, underappreciated reasons for why the aggregate level of retirement savings in America is not what it should be.”

But it seems to me that problems with this kind of compounding is that it actually compounds the problem(s), effectively obscuring reality by multiplying a conglomeration of unrelated, but ostensibly rational data points in what amounts to a classic magician’s mathematical misdirect—a sleight of hand designed to misdirect attention. 

As for the conclusions based on these assumptions about “forgotten” 401(k) accounts? Well, they’re best… forgotten. 


[i] And, at least according to their press releases, they’ve done a good job of lining up… capital.  

[ii] As a stand-alone policy initiative, EBRI has projected that the present value of additional accumulations over 40 years resulting from “partial” auto portability (participant balances less than $5,000 adjusted for inflation) would be $1.50 trillion, and the value would be $1.99 trillion under “full” auto portability (all participant balances). Under partial auto portability, those currently 25–34 are projected to have an additional $659 billion, increasing to $847 billion for full auto portability. But that picks up all potential rollovers, and they certainly aren’t “forgotten.”

Advertisement