The Impact of Assumptions on the Impact of Leakage

It has become something of an article of faith in our industry that “leakage,” the distribution of money from retirement accounts prior to retirement, is bad.

Bad, of course, in the sense that those “premature” withdrawals, via hardship withdrawals, loans or the so-called “deemed distributions” that result when a termination occurs when a loan is outstanding, reduce individual retirement savings.

So when the Center for Retirement Research at Boston College published a paper last week entitled “The Impact of Leakages on 401(k)/IRA Assets,” it really wasn’t a question of “if” there was an impact, it was a question of how much. The answer, according to the paper: a reduction in wealth at retirement of a jaw-dropping 25%.

Now by any measure, a 25% reduction in retirement wealth is a massive problem — and one that would, if valid, call for the kind of countermeasures touted by the Boston College researchers.

But consider the findings published recently by the non-partisan Employee Benefit Research Institute (EBRI) based on its massive participant database of actual participant balances and activity. Among participants with outstanding 401(k) loans at the end of 2013, the average unpaid balance was a mere $7,421, and the median loan balance outstanding was $3,973.

Moreover, the report notes — as it has for a number of years examining trend analysis in that database — that overall, loans from 401(k) plans tended to be small, with a sizable majority of 401(k) participants in all age groups having no loan outstanding at all: 88% of participants in their 20s, 73% of participants in their 40s, and 86% of participants in their 60s had no loans outstanding at year-end 2013.

So, how did the Boston College researchers manage to come up with a 25% reduction in retirement wealth? Quite simply, they started by deriving an estimate based on a “host of simplifying assumptions” — their words, not mine.

What are those simplifying assumptions? It starts with a hypothetical participant who initially makes $40,000/year, gets a 1.1% annual pay increase in real terms, assumes that he/she defers 6% of pay, is matched at a rate of 50 cents on the dollar, and whose investments gain a 4.5% real annual return. But the key assumption — and the one that arguably creates the conclusion of the paper — is their further assumption that 1.5% of assets leak out each year!

Under these assumptions, the leakages result in accumulated 401(k) wealth of $203,000 at age 60 compared with $272,000 with no leakage. So their math (and assumptions) lead to a conclusion that these leakages reduce 401(k) wealth by 25%. To “prove” the point, they provide a simple chart of the numbers they just produced, and direct the reader to it as though it provides some sort of independent corroboration.

They go on to clarify: “This estimate represents the overall impact for the whole population, averaged across both those who tap their savings before retirement and those who do not.” That’s right — averaged across everyone, not just those who actually dip into those savings prior to retirement.

Now at that point, it’s arguably just math, not so much a quantification of the impact of leakages as the mathematical impact of a series of simplifying assumptions. As for the so-called impact of leakages? It might more accurately be described as the impact of assumptions.

- Nevin E. Adams, JD

That’s not to say that leakages don’t have an impact on retirement accumulations, and doubtless for some, that impact is significant. EBRI’s analysis found that among loan defaults, hardships and cashouts at job change, cashouts were found to have a much more serious impact on 401(k) accumulation than either plan loan defaults or hardship withdrawals (even with the impact of a six-month suspension of contributions included). The leakages from cashouts resulted in a decrease in the probability of reaching an 80% real replacement rate of 5.9 percentage points for the lowest-income quartile and 4.5 percentage points for those in the highest-income quartile. In fact, the effect from cashouts (by themselves) — not loans or hardship withdrawals — turns out to be approximately two-thirds of the leakage impact.  

It’s also worth noting that it’s one thing to quantify the impact of not allowing early access to these funds — and something else altogether to assume that participants and plan sponsors would not respond in any way to those changes, perhaps by reducing their contribution levels, or by either deciding to continue to participate or to participate in the first place.