Jane the Actuary Retirement

Forbes Post, “Is Retirement ‘Out Of Reach For Working Americans’?”

Originally published at Forbes.com on September 18, 2018.

 

That’s the claim of a new report by the National Institute for Retirement Security (NIRS), titled, appropriately enough, “Retirement in America:  Out of Reach for Working Americans?”  It contains a number of dire statistics.  But how bad are things, really?

The report starts with historical context.  Over the 34 year period from 1980, prevalence rates for access of private sector workers to retirement plans bounced around a bit, from 55.4% in 1980, dropping down to 51.4% in 1988, growing to 60.4% in 1999, then dropping steadily since then to 50.9% in 2014 (the end point of the data).   Actual workers’ participation in plans followed the same pattern, but with lower rates since some workers choose not to participate; in 2014, only 40.1% of private sector workers participated in a workplace retirement plan.  Why did rates drop?  The authors say this is due to the aftermath of the 2001 and 2008 recessions, but more likely it’s a matter of increasing percentages of Americans working for the sort of contingent, contract, or small employers who have never had a practice of offering retirement benefits.

At the same time, ignoring traditional DB pensions and looking at only at retirement accounts (401ks, 403bs, IRAs, etc.), a mere 40.7% of working-age  Americans had such accounts, as of 2013, though this statistic include working-age individuals rather than more narrowly workers, and rates are considerably lower for the age 21-34 group, 27.5%, than older workers.  At the same time, not surprisingly, higher income working-age adults are far more likely to have retirement accounts:

  • 15.8% of lowest quartile individuals (income less than $15,325),
  • 27.65 of second quartile individuals ($15,325 – $30,660),
  • 52.7% of third quartile individuals ($30,661 – $55,548),
  • and 74.5% of top quartile individuals (income greater than $55,548)

had retirement accounts.

And, again looking at retirement accounts among working-age individuals, only 18.6% of Americans have greater than one times pay in retirement savings, and only 31.7% of Americans near retirement age (55 – 64) have more than one times pay, with an even smaller percentage of this age group exceeding 4 times pay (13.3%).  Separately, the report measures the net worth of working adults; only 46.3% of these near-retirees have net worth (including not just retirement accounts but home equity, college funds, etc.) of greater than 4 times pay.  Considering that Fidelity Investments recommends that one have saved ten times one’s income by the time one reaches a retirement age of 67 (with lesser targets for earlier ages), this suggests that that the vast majority of Americans are or will be unprepared for retirement:  78.2% of 35 – 44 year-olds, 79.8% of 45 – 54 year-olds, and 75.3% of 55 – 64 year-olds will not reach this target.

That sounds pretty dire.  And NIRS has policy prescriptions to remedy the problem, from increasing Social Security benefits, to mandatory auto-enroll IRAs either state-sponsored or with employer mandates, the expansion of tax credits for low-income savers, and even changes in defined benefit funding rules and the creation of hybrid risk-sharing plan types to try to encourage employers to return to DB plan sponsorship.

But there are reasons to think the situation, while concerning, is not quite as bad as all that.

In the first place, the Fidelity targets are based on typical earners, who can expect a Social Security pay replacement of 40%, and they target an 83% of pay replacement level from age 67 to age 93.  But lower-income earners receive relatively greater benefits from Social Security.  In the second place, all of these statistics ignore Defined Benefit plans which do still exist for public sector workforce, and some of the NIRS analysis ignores other sorts of savings as well.  These statistics are also based on individual income, and it’s not clear how a married couple in which one spouse had a retirement plan that provides for both would be treated in their data.  As I discussed in a prior article, in the 2017 Federal Reserve survey, 75% of non-retiree adults (rather than only 40% in the NIRS report) reported having at least some form of retirement savings, including 55% with DC plans, 26% with DB pensions, 32% with IRAs, and 43% with other savings that they considered as retirement savings even though not in a retirement account.

Here’s another data source: the Center for Retirement Research calculates their own measure of retirement readiness, the National Retirement Risk Index (NRRI).  This is based on survey data that’s conducted triennially, so the most recent figures are based on 2016 data.  The figure represents the percent of households which are projected to fall more than 10% short of the CRR’s calculation of income needed in retirement, taking into account specific impacts of Social Security and other factors for different income levels, and assuming that individuals purchase annuities with their retirement account balances.

So, on the one hand, the NRRI reports a less dire situation than the NIRS calculations — only 50%, rather than three-quarters, of Americans are at risk of declining living standards in retirement.

But, in line with the NIRS data, the NRRI shows a generally worsening situation:  the figure was about 30% in the 1980s (that is, 70% of Americans were deemed to be prepared for retirement), grew to the upper 30s and lower 40s in the 1990s and 2000s, and peaked at 53% in 2010, before dropping slightly in 2013 and 2016 to 50% at the time of the last calculation.

But these are all projections and hypothetical calculations.  What has actually been happening?  Forbes contributor Andrew Biggs wrote in the summer that “The Media’s Coverage of Retirement Saving Really is Terrible” and marshaled a number of studies to show that retirees are, at present, on average and even with respect to the lower-income folk, doing better than in the past, with incomes rising, not falling, and that yet another study, the Urban Institute’s Dynasim model

estimates that the median retiree in 2015 had income sources and assets sufficient to support a total annual income of $37,887. By 2025 real median incomes are projected to rise to $40,880, and to $42,165 by 2035. The Urban Institute model also projects that poverty rates in old age will fall, a reflection of high real incomes among the poor.

And yes, the media has no trouble finding retirees in financial straits, and you and I likely know people who, though middle-class, are not saving for retirement.

Ultimately it’s a glass-half-empty/glass-half-full situation.

Do we need to take drastic and urgent action to ensure that the next generation of retirees won’t have to all be eating dog food, per the stories one reads periodically?  No, we don’t.  But while there’s time to make a difference, we can still calmly and deliberately think about the best ways to ensure that Americans don’t fall through the cracks.

 

December 2024 Author’s note: the terms of my affiliation with Forbes enable me to republish materials on other sites, so I am updating my personal website by duplicating a selected portion of my Forbes writing here.

Jane the Actuary Retirement

Forbes post, “Should Poor People Save For Retirement?”

Originally published at Forbes.com on July 21, 2018.

 

Andrew Biggs, American Enterprise Institute scholar (and fellow Forbes contributor) has a new article at the Washington Post, “State-run retirement plans are the wrong way to protect the poor,” in which he addresses the new auto-enrollment state-managed IRA plans.

Five states are launching plans to automatically enroll employees, predominantly lower-income workers, in state-administered individual retirement accounts. More than 20 other states are considering “auto-IRA” programs like those of California, Connecticut, Illinois, Maryland and Oregon. Auto-IRAs seem like an obviously benign effort: Only about 20 percent of low-income workers participate in 401(k) plans, and many low earners depend heavily on Social Security when they retire.

But bureaucratic good intentions sometimes address problems that aren’t problems or end up doing more harm than good. In the case of auto-IRAs for low-income workers, states are likely doing both: These workers are in better shape for retirement than misleading news coverage suggests, and auto-IRAs could saddle them with higher debt while disqualifying them from means-tested government health and welfare programs — thus saving the states a fortune.

The problem, Biggs writes, is that low-income workers have little to no need to save for retirement because Social Security (and, as needed, the relevant means-tested benefits) will provide a sufficient degree of pay replacement, at the standard of living they’re accustomed to.  However, many of these folk are living at the margin to such a degree that even a small loss in income due to (semi-)obligatory savings could increase their debt levels (you’d think they could adjust their savings over time, but Biggs cites a study demonstrating this effect), and the asset tests for many benefits for the poor would mean that they wouldn’t be able to benefit from much of their savings anyway.  Biggs notes as a partial answer that “A good model is Britain’s national saving plan policy of automatically enrolling only workers with salaries above £10,000 pounds (about $13,000).”

So here are a few thoughts to build on this:

What the United Kingdom has done is not simply a matter of a creating an income threshold below which workers do not participate in the autoenrollment plan.  Not only is participation for workers with less than £10,000 in income on an opt-in rather than opt-out basis, but income below £5,876 is not included in the calculation at all.  A typical retirement plan in the Netherlands works similarly, with pensionable salary for employer pension purposes excluding between approximately the first €13,000 – €15,000 of pay, depending on plan type.  Likewise, Switzerland’s Second Pillar plan starts at income of about CHF 25,000.  And in my “MyPlate” retirement savings proposal I had suggested “a rule of thumb such as, ‘save 15% of your annual income above $20,000.'”  (As a reminder, I’d also prefer a Social Security system in which the benefit was a flat benefit for everyone to ensure that every American is protected from poverty, but even absent this, there are multiple layers of supplemental benefits, such as SSI and food stamps, for the elderly whose benefit accrual was low during their working years.)

In our American Social Security system, we don’t have a floor but we do have a ceiling, and that is applied on an annual basis; at whatever point in the year, one reaches that year’s ceiling, Social Security taxes stop.  But this wouldn’t make sense for a floor, and for savings for the poor, which should be regular and predictable.  What would make sense, though, is to apply a floor on an hourly-wage basis.  For example, for simplicity’s sake, a plan could make automatic contributions on income that exceeds, when adjusted to an hourly basis, $10 per hour (though the nice rounding would be lost if the threshold is indexed for inflation) or on income that exceeds the minimum wage.

In addition, there is an income level, for any given family circumstance, at which short-term savings is more important than a retirement fund, and where providing better vehicles for this short-term savings is more valuable than automating retirement savings, especially if those retirement funds end up being withdrawn to cover living expenses between jobs, or even just an unexpected car or home repair bill or a security deposit on a new apartment.

Happily, retirement policy experts and politicians are attempting to address this issue through the concept of a “sidecar” account, so called because it sits alongside a retirement account.  The exact method varies depending on the proposal, though one version exists in recently-introduced bipartisan legislation (more on which later),as detailed recently at Plan Advisor:

Designed either within the retirement plan, as a sidecar next to the retirement plan or completely separate from a retirement plan, the employer establishes an after-tax contribution source, in which the employee contributes up to a certain threshold, such as $1,000 or $1,500, through payroll deductions. Once the employee’s account reaches the specific threshold, future contributions shift to the participant’s formal retirement plan. . . .

David Mitchell, associate director for policy and market solutions at the Financial Security Program of the Aspen Institute, explains, “To ensure a constant savings buffer, the short-term account is automatically replenished as necessary. The hope is that by formalizing the dual role the retirement system currently plays, savers would be in a better position to distinguish between what is available now and what is locked away for retirement. This would allow them to meet both short- and long-term financial goals more easily.”

(Mitchell and the Aspen Institute have published an Issue Brief which provides more detail on the proposal.)

At present, there are various regulatory barriers which prevent employers from offering these, and the proposed legislation aims to remove these barriers; in addition, none of the state-managed programs (to the best of my knowledge) yet include such a short-term savings program.

That being said, I should also admit that I remain, in general, uncertain about whether these state-managed programs are the right path forward in what is very much a transitional period for our retirement system — but it may be quite some time before we really know how they shake out.

 

December 2024 Author’s note: the terms of my affiliation with Forbes enable me to republish materials on other sites, so I am updating my personal website by duplicating a selected portion of my Forbes writing here.

Jane the Actuary Retirement

Forbes post, “Retirement Savings And Pension Funding 101 (Some Actuary-Splaining)”

Originally published at Forbes.com on July 13, 2018.

 

How do employers fund and account for pensions?  How should workers save for retirement?  It occurs to me that going back 20 years to my first days, and first training sessions, as a trainee actuary, and revisiting the basics of pension funding methods, might provide a different way of thinking about these issues.

Back in the day, we learned that there were two general methods by which an employer could calculate their annual contribution, and their accrued liability, for a pension plan:  level-cost methods and unit credit methods.

In the level cost method, or more specifically the Entry Age Normal or EAN method, the actuary calculates an employee’s projected future benefit at retirement, then calculates the amount needed to fund this benefit on a level basis throughout the employee’s career.  If the benefit at retirement would be a flat dollar amount, then you’d calculate this as a level dollar amount throughout the employee’s career, kind of like paying off a mortgage.  If (as in a traditional pension plan) this was based on the employee’s pay at retirement, you’d project the payout to retirement, and calculate the amount needed to fund the benefit as a level percent of pay throughout the employee’s career.  For example, a worker starting at age 25, with a retirement age of 65, and a benefit accrual of 1.5% of pay per year would have accrued 40 x 1.5% = 60% of pay at retirement, so the actuary uses actuarial mathematics to determine a funding percent of pay for the employee’s working lifetime.

Sound familiar?  This is the same sort of calculation, in general principles, as a retirement savings calculator would perform when you enter in your variables, such as “I want to retire at age 65 with enough savings to cover 70% of my future salary,” and ask for the recommended savings amount as a percent of pay — except that the actuary also has to reflect the likelihood that you’ll terminate before retirement (with or without a vested benefit), and your probabilities of dying before and after retirement.

In the Unit Credit method, the actuary’s calculation is different.  To calculate the employer’s liability for any given employee, the actuary calculates the amount of benefit accrued up to the present point in time, then determines its value as a lump sum based on actuarial mathematics, considering the time-value of money as well as probabilities of termination and death.  For a pay-based plan, this is called Projected Unit Credit (PUC) and, just as in the EAN method, pay is projected to retirement to determine the benefit accrual; in the above example, if the worker was age 45 at the time of the valuation, the liability would be based on 20 x 1.5% = 30% of projected pay, and the baseline amount for one year’s funding would be 1.5% of pay.

These hardly seem like two different methods, but there’s a very important difference:  in the PUC method, as each year the employee ages and gets one year closer to retirement, the cost of that benefit increases, because the discounting for the interest rate (and the likelihood of terminating in the meantime) decreases.  If the EAN method is similar to the idealized retirement-calculator method of saving for retirement, the PUC method more closely resembles the clunky way that Americans tend to save in practice, with the amount increasing from one year to the next as retirement grows nearer — except that in the latter case this is due to worry setting in and in the former case this is a consequence of the funding method.

What’s more, in my first paragraph, I referenced learning these funding methods as a trainee actuary.  In practice, new pension funding regulations prescribe the PUC method, and pension accounting has always required this method — which makes sense, because, from an accounting point of view, the PUC method does reflect the true cost to the employer.

Incidentally, here’s another interesting tidbit:  the method of providing Defined Contribution/401(k) benefits at the same level for everyone?  There are two European countries which follow an entirely separate approach.  In The Netherlands, because DC benefits were supposed to provide benefits equivalent at any given age to the traditional pension/Defined Benefit (DB) plans they replaced, employers’ contributions to their employees’ accounts also increase by age:

A typical individual DC scheme uses an age-related table in which contributions increase with age. Since the Dutch tax authorities have set up a maximum DC premium per age bracket (of five years), most companies use a system that corresponds to this.

KWPS provides two sample tables, based on the standard target equivalent defined benefit accrual and two different interest rates (their data, my reformatted table).

 

from KWPS; https://kwps.nl/usr-uploads/files/brochures/Pension%20in%20the%20Netherlands.pdf

own table

And in Switzerland, where employers are required to provide a retirement plan, vaguely similar to a Defined Contribution system in the U.S., with employees contributing about 1/3 of the total contributions, contributions are also based on age, with a minimum total contribution of:

Age 25- 34, 7%

Age 35-44, 10%

Age 45-54, 15%

Age 55-65, 18%

So step back and imagine your employer contributing more to your 401(k) account the older you are.  Hard to imagine, isn’t it?  — Although in some of the earliest American cash balance/hybrid plans, the employer’s contribution schedule did that as well.

So far this all seems very abstract and not of much relevance to American retirement savings or retirement policy.  But consider an article in yesterday’s Wall Street Journal, “Behind on Retirement Savings? It’s Not Too Late to Catch Up.”  It’s paywalled, but much of its content is based off of a blog post by Michael Kitces, “Why The Empty Nest Transition Is Crucial For Retirement Success,” which proposes that, as an alternative to the goal of saving a level percent of your income throughout your working lifetime, you actively plan to cut your spending upon becoming an empty-nester, to the degree necessary to, in fact, fund your retirement, bearing in mind that this requires active efforts to keep your standard of living the same when kids have left the house, rather than boosting your spending to enjoy travel you’ve deferred or otherwise make up for an empty house with more (costly) activities.

Kitces writes:

for those who allow their total family consumption to fall (now that Mom and Dad are only supporting Mom and Dad, and not the kids), a significant uptick in savings can almost entirely make up the lack of savings during the child-rearing phase. In other words, if all the money that used to be spent to raise a family and send kids to college is now redirected into saving for retirement, a couple might find themselves suddenly leap from saving less than 5% of income to being able to save 25% or more.

Is this realistic?  The WSJ article mentions research by the Center for Retirement Research that, on average, empty nesters don’t actually significantly increase their retirement savings when the kids leave home — whether because they’re still supporting those kids (college tuition, weddings, helping out in other ways) or for other reasons, and Kitces himself acknowledges that for who started their families late in life, there’s not much time to save after the kids are grown and before hitting retirement age — which means that, rather than providing a solution to the question of retirement savings, it really just provides a different perspective to take into consideration.

 

 

December 2024 Author’s note: the terms of my affiliation with Forbes enable me to republish materials on other sites, so I am updating my personal website by duplicating a selected portion of my Forbes writing here.