Gooseneckers, Misleading Medians, and the Art of Retirement Alarmism

 Did you hear about the one that claimed an “average” American worker has less than $1,000 saved for retirement?

Well, here’s hoping you didn’t — that your day was occupied with real issues, or perhaps even better that you saw the headline, recognized it for the ludicrousocity[i] of the claim, and scrolled on without clicking, sharing, or commenting. 

But some didn’t. Drawn like a moth to a flame (or perhaps more precisely, gooseneckers at the scene of a horrific accident), some likely did click, if only to see the preposterous assumptions and/or incredulous inverse compounding applied to create such a ridiculous conclusion.

The CBS report cites “research” (and I use that term loosely here) by the National Institute on Retirement Security (NIRS) which — if one has paid attention to its previous outputs might more credibly be called the National Institute on Retirement INsecurity. I say that because the organization — which labels itself “nonpartisan” — nonetheless, and unapologetically definitely has a mission. That mission is the promotion of defined benefit plan designs — and while there’s nothing wrong with that, in the absence of good positive private sector trends to highlight there, they instead tend to find ways to bash what has become the nation’s retirement plan design alternative — the 401(k).

As for this most recent attempt,[ii] you don’t have to dig deep, or delve into footnotes to see just how data was convoluted to derive that click-bait crafted outcome. They simply took data from what appeared to be a reliable government source — though it happens to be a self-reported number of accumulated savings[iii] taken by the government — for all employed adults aged 21-64. Yes, you read that correctly.

Oh, and then picked the median of that wildly diverse range of experiences. 

ad space

So — you take the accumulated savings of a 21-year-old — and mush it together with that of someone on the brink of retirement…. Honestly, you can stop right there, and know that this is a stupid, although ostensibly mathematically accurate, result. Seriously. You might just as well take the mid-day temperature of the Sahara Desert, the midnight temperature of Antarctica, add them together and to derive the temperature in Omaha, Nebraska. 

And yet, that’s the kind of math that is the basis for the headline. 

Then, presumably to provide some “balance,” NIRS produced a median number for those who had some retirement savings — again, though — every worker from age 21-64 — and provided a median of $40,000. 

But again, a mathematically accurate result[iv] that tells us…nothing.[v] 

To its credit, the NIRS report itself (eventually) acknowledges what actually matters: access to a workplace retirement plan, the need to address Social Security’s funding shortfall, and the drag student loan debt places on retirement readiness.

But those realities are buried beneath a click-bait headline built on a median so broad as to be meaningless. That isn’t analysis — it’s alarmism dressed up as math. And while it may generate attention, it amounts to click-bait journalism enabled by irresponsible and misleading “research.”

Don’t fall for it — and by all means, don’t spread it around.

ad space

• Nevin E. Adams, JD

 

---

[i] Yes, it’s a made-up word.

[ii] See Retirement in America: An Analysis of Retirement Preparedness Among Working-Age Americans - NIRS.

[iii] Andrew Biggs notes that in this data sampling from the Survey of Income and Program Participation (SIPP), the bottom quintile of earners have median annual earnings of just about $20,000 even for those aged 35 and over. In other words, these are people who are barely working – very few hours and weeks worked, typically at very low wages. See Does the typical American have only $955 saved for retirement?

[iv] One assumes, but considering the logic in the compilation, it may bear double-checking.

[v] The CBS report went further, of course — just in case you weren’t panicked enough, they juxtaposed the bizarre medians noted above with some generalizations about how much income you’d need in retirement, layered it in with worries about Social Security (and some exaggerated assumptions on how much/many Americans rely heavily on it), and even threw in a reference to Trump Accounts to generate even more clicks.    

Middle "Grounds"

A new report entitled “The Missing Middle” by the National Institute on Retirement Security (NIRS) treads some all-too-familiar ground, myopically focusing on one element of the nation’s private retirement system.

The articulated concern is, of course, the “middle”—an income grouping for which Social Security’s progressive structure doesn’t reach high enough to provide an adequate replacement income, but that lacks the more expansive financial wherewithal of those at the upper end of the income strata.

According to the paper, the tax incentives that arguably existed at the birth of the 401(k) have been muted due to lower marginal tax rates and the expansion of the standard deduction—both of which serve to mitigate the tax burden on lower-income individuals—but in the process also arguably lessen the financial incentive for deferring taxes. And if that were not enough, the authors also argue that the “…tax benefits relating to investment returns may be less in a market with lower returns.”

Of course, the focus of the authors here is the tax incentives for retirement savings[i]—and, unsurprisingly, the premise is that those with lower incomes—and thus less tax liability—get less from the current tax deferrals afforded 401(k) contributions than do those at higher incomes (who pay more in taxes). And, if you look only at that aspect—and that’s where most such critiques stop—it’s a fair point.

Before going into the shortcomings in that analysis, I’ll admit that there are certain legitimate economic realities that the paper highlights—that higher-income (and by this we don’t necessarily mean wealthy) individuals are more likely to have access to a retirement plan through work, that there are racial aspects that correlate to wealth inequities and access in the workplace, that the Saver’s Credit as currently designed (requiring a long-form tax filing to claim and being non-refundable) aren’t available to many who would otherwise be eligible, and that Social Security, though an underlying foundation of private requirement as a whole, and particularly for lower-income individuals, has funding issues of its own to fulfill the current benefit promises.[ii]

‘Missing’ Interactions

Unfortunately, as noted above, these types of analyses always gloss over the interrelationships between the tax incentives and the creation of these plans in the first place. It is assumed (generally implicitly) that employers that want to be considered an “employer of choice” will be forced to offer these plans regardless of the tax preferences to do so. Let’s face it, the tax preferences—though modest at an individual level—do provide an incentive to not only offer the plan, but—in most cases—to provide a matching contribution. A matching contribution that these type critiques always seem to gloss over (it does make their math simpler). And let’s face it, there’s no question that having access to a plan matters—even this NIRS paper acknowledges that those with access are 15 times more likely to save. 

What’s also glossed over is the impact of non-discrimination tests and legal contribution limits—limits that work, and work as designed, to keep an effective balance between the benefits of higher-paid and other workers. In fact, data from the Employee Benefit Research Institute has proven that while higher-income individuals do have higher account balances, those balances are in rough proportion to their incomes. 

In calling for a “recalibration” of what they see as a “fundamentally inequitable system,” the well-intentioned authors are missing the mark. By focusing exclusively on the individual tax preferences, while at the same time ignoring the impact of tax preferences on the decision to offer a plan in the first place, as well as the influence of non-discrimination testing in encouraging an employer match (not to mention the financial impact that has on the retirement prospects of non-highly compensated workers), they—and their purported solutions—turn out to be missing the point—and the very “middle” they claim the current system overlooks. 

- Nevin E. Adams, JD

---

[i] Once again, the trade-off for deciding to defer taking pay now, and depositing it in a trust subject to various restrictions and pre-withdrawal penalties is that you don’t pay taxes on compensation you haven’t gotten access to.

[ii] To address the perceived shortcomings identified, the authors have several solutions. Specifically, they want to boost and expand Social Security, federalize the state-run IRAs, target the Saver’s Credit to participants in those programs, and perhaps introduce a state version of the refundable government credit in place of tax incentives, which would equalize the credit. 

Data "Minding"

Just when you thought retirement plan projections couldn’t get any worse…

Last week the National Institute on Retirement Security released a report – “Retirement in America | Out of Reach for Most Americans?” that claimed that the median retirement account balance among all working individuals is… $0.00. Moreover, that same report claims that “57 percent (more than 100 million) of working age individuals do not own any retirement account assets in an employer-sponsored 401(k)-type plan, individual account or pension.”

It’s not the first time the NIRS has produced reports finding significant problems with the nation’s private retirement system. Indeed, this report “builds on previous NIRS research published in 2015,” though the conclusions presented here seem unusually stark.
Then, as now, the NIRS conclusions rely heavily on self-reported numbers from the Federal Reserve’s Survey of Consumer Finances (SCF) and the U.S. Census Bureau’s Survey of Income and Program Participation (SIPP). The report’s authors acknowledge that the latter “oversamples” lower-income household which, as the report notes, are less likely to be covered by, or to participate in, an employer-sponsored retirement plan.

‘Self’ Sufficient?

The issues with self-reporting have been well-documented – so much so that even the report’s authors acknowledge (albeit in a footnote) that it “…can be problematic for the reporting of account balances and participation in particular types of retirement plans, such as DB pension plans.” In fact, previous research by Irena Dushi, Howard M. Iams and Jules Lichtenstein using SIPP data matched to the Social Security Administration’s (SSA’s) W-2 records found that the DC pension participation rate was about 11 percentage points higher when using W-2 tax records compared with respondent survey reports, “suggesting that respondents either do not understand the survey questions about participation or they do not recall making a decision to participate in a DC plan.” Those authors also found inconsistencies between the survey report and the W-2 record regarding contribution amounts to DC plans. In fact, those researchers found that about 14% of workers who self-reported nonparticipation in a defined contribution (DC) plan had, in fact, contributed (per W-2 records), while 9% of workers self-reported participation in a DC plan when W-2 records indicated no contributions.

Supplementing SIPP survey reports with actual information on tax-deferred contributions in W-2 records, the researchers found that the percentage of employees who were offered a retirement plan increased from 72% in 2006 to 75% in 2012, whereas the participation rate in any retirement plan among all private-sector workers increased from 58% to 61% over this period (a difference that may seem small, but is statistically significant at the 5% level).

Participation, Rated

Compounding the issues, for participation data, the NIRS draws on the Current Population Survey (CPS), even though the report’s own footnotes acknowledge that “the 2014 redesign of CPS produced much lower participation rates for working Americans in years after 2014 (participation in 2014 was at 40.1% but at 31.7% in 2017), which has not been fully explained.” They aren’t the only ones to take note of this aberration (see CPS Needs a New GPS and Commonly Cited Participation Gauge Misses the Mark, Study Says), but decide, for reasons not fully explained, to rely on the data there anyway. Indeed, a June 2018 report on the impact of the changes in the CPS by the nonpartisan Employee Benefits Research Institute (EBRI) cautions that “the estimates from the most recent surveys could easily be misconstrued as erosions in coverage, as opposed to an issue with the design of the survey.”

Moreover, when it comes to assessing retirement readiness, the NIRS analysis extrapolates target retirement savings needs based on a set of age-based income multipliers – income multipliers, it should be noted, that have no apparent connection with actual income, or with actual spending needs in retirement, although this year the report’s authors acknowledge that those are “rule-of-thumb multipliers and are not based on detailed projections of the income needs of individuals,” all of which, in the authors’ words means that their “analysis in aggregate terms, is broadly suggestive rather than definitive.”

Median, Well…

However, much of the data – and key elements, such as wealth, income, participation and retirement savings – that underlie the conclusions is “self-reported” which, as noted above, even the report’s authors acknowledge “…can be problematic…”. It is perhaps a necessary “evil” when seeking to draw broad-based conclusions from limited samplings, but – particularly when sweeping generalizations are posited based on the medians of such samplings, when labels like “typical” are affixed and assumed without explanation – well, let’s just say that caution in applying the conclusions seems well-advised.

Not that the report is completely off the mark; it points out that those with access to retirement plans are significantly better off than those who lack that access, but also that retirement plan coverage has languished at about the 40% level in the private sector for some time now. Clearly solutions that help work to close this coverage gap, and provide more workers with an easier opportunity to save (let’s face it, nothing stops folks from walking down to their local financial services institution and opening an IRA, but the data suggests that those with access to a plan at work are 12 times more likely to do so) are needed.

Ultimately, the recommendations put forth by the NIRS report authors – to strengthen social security, expand access to workplace retirement plans, and expand the saver’s credit – should serve to narrow the gaps in retirement plan access and adequacy.

Even if the data that outlines the situation – and purports to justify those actions – leaves something to be desired.

- Nevin E. Adams, JD

Feel Lucky?

One of my favorite cinematic quotes is from that Clint Eastwood classic, “Dirty Harry.”

It comes at the very beginning of the movie – there’s a bank robbery in process, and Clint Eastwood (in the personage of Inspector “Dirty” Harry Callahan) is trying to wolf down a hot dog when the commotion starts up. Harry, clearly irritated, gets up and heads out, and proceeds to shoot it out with the robbers, and then strolls over to the one who is still breathing, who has a shotgun within reach. Harry proceeds to point out the attributes of his .44 Magnum as he wonders aloud if he still has any bullets left, as the wounded robber contemplates his chances of getting to his shotgun before Harry pulls the trigger. At that point, Harry reminds him: “You’ve got to ask yourself one question: ‘Do I feel lucky?’ Well, do ya, punk?”



When reading the various surveys of workers who seem confident about their retirement prospects – or don’t – when most haven’t even tried to figure out how much they will need – I can almost hear Harry Callahan in the background. They might not feel lucky, but they’re acting as though they will be.

Last week the nonpartisan Employee Benefit Research Institute (EBRI) published an Issue Brief highlighting some of the outcomes that EBRI’s Retirement Security Projection Model (RSPM) has projected in recent years. The RSPM was developed to help state governments figure out if their residents would run short of money in retirement – based on a concern that to the extent they did, the social safety net might have to be patched.

Now such models¹ must, of necessity, rely on certain assumptions, but unlike a number of other models out there, EBRI’s has a significant advantage – being able to draw on actual, anonymized administrative data of tens of millions of 401(k) participants for the retirement savings component, whereas others lean on self-reported samplings. Moreover, rather than rely on the extrapolation of measures like replacement rates, EBRI’s RSPM considers a household to “run short of money” or experience a retirement savings “shortfall” if it doesn’t have enough money to cover actual projected expenses in retirement – including an aspect that most models completely ignore: uncovered long-term care expenses from nursing homes and home health care.

Despite what I would consider to be pretty stringent assumptions on expenses, the RSPM projects that more than half — 57.4% — of all U.S. households (not just those covered by employer-sponsored retirement plans) will not run short of money in retirement. Oh, and that’s assuming that they come up with 100% of those expenses.

But what if you wanted to assume that, confronted with those expenses in retirement, individuals cut back on their spending – say to just 90% of the projected expenses? Well, even if you leave in place the full assumptions about nursing home and home health care costs, under that scenario the percentage of households projected to have sufficient retirement resources increases to more than two-thirds (68.1%). What if you assume that retirees only spend 80% of the cohort average for deterministic expenses? At that point, 82.1% would have enough financial resources. If you are willing to ignore the potential impact of long-term care costs, three-quarters (75.5%) of households will have sufficient financial resources in retirement to meet 100% of projected expenses.

That doesn’t mean that access to a retirement plan doesn’t matter. The report notes that among Gen Xers, those who have 20 or more years of future eligibility (including years in which employees are eligible but choose not to participate) are simulated to have a 72% probability of not running short of money in retirement. Those in that same group with no future years of eligibility are simulated to have only a 48% probability of not running short of money in retirement.

In fact, as you might expect, eligibility for a workplace retirement plan is one of the most – if not the most – important aspects that affect retirement success.

Now, as promising as those results seem, one shouldn’t draw too much comfort. Those who come up short will do so by varying amounts, and some by quite large amounts. Indeed, when you add up all those shortfalls, it amounts to $4.13 trillion in 2014 dollars.

And yet, applying what strike me as remarkably conservative retirement expense assumptions, and drawing from real-world actual administrative savings data, it seems that a good number of us can – assuming we keep doing what we are doing – anticipate a financially successful retirement.

Which brings us back to “Dirty Harry”: “Do you feel lucky? Well, do ya?”

If you are eligible for a retirement plan at work, the answer would seem to be – yes!

- Nevin E. Adams, JD

Two of the most commonly cited projection models are the Center for Retirement Research (CRR) at Boston College and the National Institute on Retirement Security (NIRS). Both rely heavily on self-reported numbers from the Federal Reserve’s Survey of Consumer Finances (SCF), and while both track the progress of American retirement readiness by examining how individuals in the SCF did over time, they fail to acknowledge that doing so compares the balances and readiness of two completely different groups of individuals at different points in time. The NIRS analysis builds on that shaky foundation by incorporating some assumptions about defined benefit assets and extrapolating target retirement savings needs based on a set of age-based income multipliers — income multipliers, it should be noted, that have no apparent connection with actual income, or with actual spending needs in retirement. But then, the math is easier. ↩

The Gaps in Retirement Savings "Gaps"

As spring follows winter, so apparently do dire predictions about the nation’s retirement prospects.

For example, the Center for Retirement Research (CRR) at Boston College now claims we are looking at a $7.7 trillion dollar “retirement gap” for American workers, up from $6.6 trillion five years ago. Meanwhile, a new report by the National Institute on Retirement Security (NIRS) on what it called the “Continuing Retirement Savings Crisis” didn’t cite a specific aggregate gap, but a year ago NIRS employed a similar approach to suggest that that gap was likely somewhere between $6.8 trillion and $14 trillion.

While these outcomes are cited widely (the CRR’s as recently as last week in a U.S. Senate hearing), the underlying methodologies and assumptions are worth understanding. Both rely heavily on self-reported numbers from the Federal Reserve’s Survey of Consumer Finances (SCF), and while both track the progress of American retirement readiness by examining how individuals in the SCF did over time, they fail to acknowledge that doing so compares the balances and readiness of two completely different groups of individuals at different points in time. The NIRS analysis builds on that shaky foundation by incorporating some assumptions about defined benefit assets and extrapolating target retirement savings needs based on a set of age-based income multipliers — income multipliers, it should be noted, that have no apparent connection with actual income, or with actual spending needs in retirement. But then, the math is easier.

There’s no question that there is a gap between what Americans are likely to need to live comfortably in retirement and the resources available to fund it. The non-partisan Employee Benefit Research Institute (EBRI) recently updated its EBRI Retirement Readiness Rating and found that the retirement savings gap was $4.13 trillion for all U.S. households (not just those who have retirement plan balances, though it uses actual 401(k) data for those who do have such plans, rather than relying on self-reported estimates), where the head of the household is between 25 and 64, inclusive. That may be well short of the projections offered by the NIRS and CRR, but it’s a big number, nonetheless.

Indeed, considering the enormity of that gap, policymakers — and Americans generally — might well feel like throwing up their hands and despair of ever closing it (indeed, earlier this month NIRS published a survey indicating that 86% of Americans believe the nation faces a retirement crisis).

What is unfortunately often lost in the trillion-dollar gap headlines (and the concurrent surveys that, unsurprisingly, talk about our deteriorating confidence about our retirement) is that not everyone has a retirement savings gap. Ironically, considering their differing methodologies, the CRR, NIRS and EBRI all put the number at about 50%. (EBRI’s number, which in its assumptions and use of actual data seems more conservative, says the number at risk of running short of money in retirement is in the low 40% range.)

How big is the gap at an individual level? The EBRI analysis breaks it down into manageable numbers: For those on the verge of retirement (Early Baby Boomers), the deficits vary from an aggregate of $19,304 (per individual) for married households, to $33,778 for single males and $62,734 for single females. If you look only at the individuals who do have gaps (EBRI’s projections, which take into account the potential costs of nursing home care and living expenses based on real experience, rather than arbitrary replacement ratios, indicate that about 57% will have sufficient retirement income), the gap for Early Boomers ranges from $71,299 per individual for married households to $93,576 for single males and $104,821 for single females.

Not surprisingly, those eligible to participate in workplace retirement plans fare better. Also not surprisingly, those who have jobs are more likely to have incomes and access to a workplace retirement plan — and those who work for larger employers are more likely to have both larger incomes and access.

However, even EBRI’s estimates include a wide range of personal circumstances, from individuals projected to run short by as little as a dollar to those projected to fall short by tens of thousands of dollars. For those seeking to understand, and perhaps craft solutions for, the current projected shortfalls, this is an important distinction, and one given short shrift by a headline’s focus on the aggregate.

There are, of course, broad policy changes that can, and have, made a big difference, nationally and at a plan level — things like automatic (and immediate) enrollment, contribution acceleration and prudent selection of a default investment option, not to mention ideas (or legislation) that expand access to those plans.

That said, ultimately the retirement savings gap is an aggregation of individual savings gaps. And as advisors well know, you close those one individual at a time.

- Nevin E. Adams, JD