Jane the Actuary Retirement

Forbes Post, “Social Security, The FICA Tax Cap, And Having Your Cake And Eating It Too”

Originally published at Forbes.com on October 25, 2018.

 

Every time I write about Social Security and its financial woes, I inevitably get comments that the entire shortfall can be solved simply by eliminating the cap on FICA taxes, so that the wealthy pay “their fair share.”  After all, the Medicare portion of FICA has already had cap removed, so why not do the same for Social Security?

Here are some key facts:

  • In 2019, the maximum taxable earnings for Social Security will be $132,900, according to recently-released figures.  Below this level, all Americans pay 6.2% of their earned income into Social Security, and their employer pay another 6.2%.  Self-employed workers pay 12.4%.
  • According to The Social Security Game, by the American Academy of Actuaries (it’s fun; you should try it), if the cap were eliminated and earnings above the cap were not credited with Social Security benefit accruals, it would make up for 88% of the shortfall.  If high-income workers did receive accruals based on their above-cap pay, it would only make up for 71% of the shortfall.
  • The Social Security benefit formula is structured to be “progressive” — that is, lower-income earners accrue benefits at a higher rate relative to their income than higher-income earners.  Here’s how it works:  all of your income as recorded by the Social Security Administration is indexed, which means it’s adjusted to 2018 based on national average wage increases since the year you earned it.  Then the highest 35 years are averaged together (if you had less than 35 years of work, there are 0s included in the average), and then your benefit is 90% of the first $11,112 of your average indexed earnings, 32% of the next level of earnings up to $66,996, and 15% of earnings higher than this level.  (See The Motley Fool for these updated-to-2019 figures.)  It’s the same idea as marginal tax rates, except in reverse.  Which means that, while it goes without saying that if high-paid workers simply paid in more taxes without any new accrual, the additional taxes collected simply subsidize everyone else, it’s also the case that even if higher earners accrued benefits on this income, they would still be heavily subsidizing lower earners — otherwise this wouldn’t be remedying the shortfall.
  • Removing the ceiling is consistently popular in polls.  For example, a 2017 poll by the National Committee to Preserve Social Security and Medicare found that 61% of likely voters “strongly” and 13% “not so strongly” favored a proposal to “gradually require employees and employers to pay Social Security taxes on all wages above $127,000, which they don’t do now” and an even higher percentage — 69%/10% — favored a proposal to “increase Social Security benefits by having wealthy Americans pay the same rate into Social Security as everyone else.”  An admittedly-leading question in a  2016 poll found that 72% of respondents supported “increasing — not cutting — Social Security benefits by asking millionaires and billionaires to pay more into the system.”  And a more academic but somewhat older analysis from 2014 found that 39% of Americans strongly favored and 40% somewhat favored eliminating the cap.

Looking at this can make it appear as if it’s a no-brainer to remove the cap.  Only the rich pay, and everyone else benefits.  Yes, they might whine that their taxes go up by 12.4% with nothing to show for it and they’re already paying higher rates, but better that than raise taxes across-the-board or force the elderly to cope with benefit cuts.

But what about the conventional wisdom that says that we need to keep the payroll tax cap (and in addition reject means testing) in order to get broad support of the system as one in which everyone contributes their fair share and has earned their benefits rather than receiving welfare?

How can such large proportions of Americans support making a change that fundamentally undoes this “earned benefits” design to Social Security, especially when the conventional wisdom is that Americans believe that, not only have they earned their benefits, but that the money they contributed was set aside to fund their own personal retirement benefits (or, alternately, would have been had Congress not “stolen” it)?

Is this cognitive dissonance?  Are Americans foolishly, even ignorantly, clinging to their belief that they earn their benefits fair-and-square even when their support of removing the cap says they believe the rich should pay for everyone else?  Do they want to have their cake and eat it too, by collecting subsidies while still insisting they’ve stood on their own two feet all along?

I don’t think so.

After all, a 2017 poll found that 48% of Americans support the idea of a universal basic income, up from only 10% a decade ago, when described as a way to help people who lose their jobs due to AI.  Another poll found 38% somewhat or strongly supported a $1,000/month government check paid for with a tax hike on those earning $150,000 or more.

These UBI supporters are still in the minority, but the idea’s popularity is increasing, and it appears to be just one way out of many in which people are growing increasingly comfortable with the idea that the middle-class should receive government benefits, not (just) the poor.  And once people are comfortable with the idea of “middle class welfare,” then it stands to reason that they’d consider the right solution to the funding deficit to be one requiring the wealthy to top up the system however much is needed?  In such a case, they might be viewing their benefits as “earned” in a more metaphorical/symbolic sense, in which they have a right to them by having been a hardworking American during their working lifetime, regardless of what the math shows.

If the payroll tax cap is removed, it will not be a matter of having the wealthy pay “their fair share.”  It will be a shift towards, or a recognition of (depending on your perspective) FICA taxes being taxes, nothing more or less.  And in that case, why not integrate it all into our regular income tax structure, with the same marginal tax rates, deductions, taxation of investment income, and the rest?

Side note:  I tried to find polling on the extent to which Americans understand that Social Security is fundamentally a pay-as-you-go system with modest reserves having been built up by past surpluses, rather than a genuinely prefunded system, and is a system with significant subsidies from one group to another (not just by the benefit formula, but subsidies from singles to families and from dual-earner to single earner couples, as well), but there’s not much out there.  Polls intending to determine Social Security knowledge, such as this Mass Mutual survey, ask about such practical items as retirement ages and spousal benefits.  One 2010 survey does ask a more basic question and finds that roughly a quarter of Americans believe that benefits are based on contributions plus interest, one half either answer correctly or a rough approximation of the correct answer — that is, either an average of the highest 35 years of earnings or a private pension-like 5 year average pay times working years — and one quarter don’t know; it does not ask whether people think the system is funded or pay-as-you-go, or whether they think their benefits are proportionate to their income.  But without more survey questions, we’re left to draw conclusions from such sources as viral Facebook posts, including one, a variant on a Snopes-fact-checked version, that came across my Facebook feed recently and insisted, “This is NOT a benefit. It is OUR money , paid out of our earned income! Not only did we all contribute to Social Security but our employers did too ! . . .  This is your personal investment.”

 

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, “Yes, Social Security Does Indeed Add To The Federal Deficit”

Originally published at Forbes.com on October 22, 2018.

 

Let’s start with an old joke, attributed to Abraham Lincoln:

How many legs does a dog have if you call his tail a leg? Four. Saying that a tail is a leg doesn’t make it a leg.

and an adage:

If it looks like a duck, swims like a duck, and quacks like a duck, then it probably is a duck,

which has been given the name, the “duck test,” as a form of logical reasoning.

Both of which are useful perspectives for the question of “does Social Security add to the deficit?”, to which fellow Forbes contributor Teresa Ghilarducci asserted the answer is an unquestionable “no.”

After all, she writes,

But Social Security can’t, by law, add to the federal deficit. Medicare and Medicaid can, but not Social Security. Social Security is self-funded.

And Congress did declare in 1990 that Social Security spending and its build-up with reserves are not a part of the federal budget (see this Wikipedia article for background).  As the nonpartisan Tax Policy Center writes,

The budget brings together the spending and receipts of virtually all federal activities, from paying doctors who treat Medicare patients to financing the Environmental Protection Agency to collecting income taxes to selling oil leases on federal land. In two cases, however, Congress has separated programs from the rest of the budget. The Postal Service Fund and the disability and retirement trust funds in Social Security are formally designated as “off-budget,” even though their spending and revenues are included in the unified budget.

Lawmakers created this special accounting to try to wall off these programs. For the Postal Service, the intent was to free the agency to pursue more efficient practices than the conventional budget process allows. But that has not helped it avoid financial difficulties.

With Social Security, the intent was to protect any surpluses from being diverted into other programs. The two Social Security trust funds have accumulated large surpluses since 1983. Those surpluses will eventually be drawn down to pay future benefits. It was therefore argued that those surpluses should be separated from the surplus or deficit of the rest of government. Congress hoped that this separation would induce greater fiscal discipline in the rest of the government.

But think about that joke:  Saying that a tail is a leg doesn’t make it a leg.

Congress decreed that Social Security deficits or surpluses would not be included in its calculations of budgetary spending or calculations of deficits or surpluses whenever the federal government publishes these calculations.  And its motive was well-intentioned enough, but it wasn’t a matter of applying broad accounting principles.

Other “off-budget” federal activities are very different; for instance, Fannie Mae and Freddie Mac are government-sponsored enterprises with private ownership.  And in other instances, the exclusion is because “the government plays a limited role in what is otherwise a private activity,” to quote from the Tax Policy Center again.

But Medicare is on-budget even though its financing, with respect to Part A, is functionally the same as Social Security:  dedicated payroll taxes and a trust fund.  And even an entity such as the PBGC, the provider of “insurance policies” to protect workers’ pensions if their employer goes bankrupt, is on budget, which resulted in the premiums that plan sponsors are required to pay being increased in 2015, at least in part in order to boost government revenue for a budget deal.  And if economists were evaluating the government finances of some other country, they would hardly accept its legislature’s definition of budget deficits or surpluses in performing their analyses.

So just because Congress has decreed that, in its reporting, Social Security finances are to be excluded from budget reporting, doesn’t make it so, in terms of real-world analysis and economic impact.

Which means we need to apply the duck test.

Deficits, after all, don’t matter in isolation.  What matters is the impact of ongoing deficits on the national debt.  How much money does the United States need to borrow?  How does that affect the economy?  Does purchase of government bonds reduce the amount of money going into private investments that would grow the economy?  Can we manage the national debt in such a way as to avoid turning into another Weimar Germany, printing money and producing inflation?  I recognize that there are many people comfortable with the mantra that “deficits don’t matter” and fully confident that politicians can walk that fine line of spending money for programs on their wishlist without crossing over into inflationary spending; that strikes me as risky hubris.

At the same time, what matters is not the total national debt, but the net debt after excluding intragovernmental debt, which is what the Trust Fund is.  Activists might repeat “the government bonds in the Trust Fund are real assets” until they’re blue in the face, but each dollar of FICA surplus, back when it existed, decreased the degree to which the federal government needed to borrow from outside, and each dollar of Trust Fund bond redeemed, is another dollar which the requires the issuing of more bonds.

What’s more, while the Trust Fund bonds are “real” and the government would no more default on them than they would default on any other bonds, a default is wholly unnecessary.  All that’s needed for the government to keep the Trust Fund bonds “unspent” is to reduce Social Security spending, in whatever manner it chooses:  a boost in the retirement age, a benefit phase-out based on other income, an across-the-board haircut, or whatever other mechanism it chooses.  If Congress changed the law tomorrow, all of those beautiful Trust Fund bonds could be kept in perpetuity, never to be redeemed.

Which means that my assessment of the duck test is that in a practical sense based on the plain meaning of words and the impact of the system, regardless of Congress’s labels, Social Security deficits are a real part of the overall governmental spending picture .

So let’s address, then, the final item:  does Social Security add to the deficit, as opposed to simply being a part of government spending?

This is just math, and the Social Security Administration has helpfully done most of the math for us, with projections of the program cost in 2018 dollars (that is, adjusted for inflation) and as a percent of projected GDP.  I’m not going to dispute their calculations — except to observe, as I have in the past, that the projections assume that the U.S. birth rate returns to a “replacement level,” which may or may not happen.

Here are the numbers, based on their intermediate forecast:

From 2007 to 2017, spending on Social Security (old-age and disability combined) increased by an inflation-adjusted 36%.

From 1997 to 2017, the increase was 70%.

From 1987 to 2017, 214%.

From 2017 to a projected 2028 value, it’s another 40% increase.  And to 2033, when this particular table’s projections stop because the trust fund is exhausted, the increase is 60%.

As a percent of GDP, the figures aren’t quite as unsettling, though it’s harder to really grasp the significance of these figures.

From 2007 to 2017, combined spending as a share of GDP increased 19%.  From 2017 to 2027, the projected increase is 15% and from 2017 to 2037, 24% — from 4.91% of GDP to 6.07%.

Now, readers, this is where I stop for today.  I am not interested in making the case today that a move from 4.91% to 6.07% is catastrophic, or no big deal, or somewhere in-between, nor am I keen on arguing about the reasonability of the Trustees’ Report’s assumptions.  And I am furthermore not keen, in this article at any rate, on arguing for the right level of taxation or identifying where Social Security fits in with other spending priorities.

But this is the bottom line:  sometime between now and a forecasted 2034, Congress will need to pass some sort of Social Security legislation.  Most cynically, it can simply declare that general revenue funds will supplement FICA revenues, but Americans would be better off if at some point between now and then we had a real, meaningful re-think of the best way to structure our retirement system.  And when that happens, it is vital that policymakers base their recommendations, and Congress, its laws, on real data rather than rhetoric.

 

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, “Is A Hike In Social Security Retirement Age Really Just A Benefit Cut?”

Originally published at Forbes.com on October 5, 2018 and October 4, 2018.

 

You’ve heard this before, with respect to the prospect of raising the Full Retirement Age in Social Security:

“Raising the retirement age amounts to an across-the-board cut in benefits, regardless of whether a worker files for Social Security before, upon, or after reaching the full retirement age.”  Paul N. Van de Water and Kathy Ruffing, Center on Budget and Policy Priorities.

“[R]aising the full retirement age is nothing more than a benefit cut on future retirees.”   Sean Williams, The Motley Fool.

“Raising the full retirement age may sound innocuous. But it is nothing more than a benefit cut, and one that puts low-paid workers at risk.”  Alicia H. Munnell, director of the Center for Retirement Research at Boston College, writing at MarketWatch.

And these are just the top three search-engine results.

But here’s where my observation yesterday that the United States is in the minority of Western countries with respect to our Social Security benefit structure, comes into play.  Other countries are much more likely to either have a fixed retirement age with no early retirement option, or to allow early retirement only with a very extensive work history and, in that case, without reductions.

Here’s why this matters:

In the U.S. Social Security old-age retirement benefit system, taking into account its existing plan design permitting retirement at a range of ages, and considering the way the hike in the “full retirement age” was implemented, in which both the minimum age and the age of maximum benefit stay unchanged and the benefit level at any given year is reduced, yes, the 1983 “full retirement age” increase was indeed  functionally a benefit cut.

And if we followed the same pattern, with a range of retirement ages from 62 to 70 but with the age for so-called “full retirement” moved up to 68 or even later, then we would be implementing a further benefit cut.  This may nonetheless be a part of an overall Social Security reform package, and may be reasonable and appropriate to the extent that the combination of increasing health of and decreasing physical demands on older workers pair together to mean that individuals are able to work longer without undue hardship.  However, unless workers are able to continue to boost their benefits via late-retirement increases beyond the age of 70, even workers who plan to retire late to make up the difference will be unable to fully do so.

But that doesn’t mean that any such retirement age increase is necessarily a benefit cut.

Consider the Danish system, in which the retirement age increases to 68 in 2038 and is scheduled to increase based on further life expectancy increases since then.  Clearly, there is no way in which a worker reaching the age of 67 in 2038, and obliged to defer retirement another year, has had a “benefit cut” in the sense of a percentage reduction of monthly benefit payments.

At the same time, yes, in principle, one could say that the total value of benefit payments received over one’s lifetime has been reduced.

Consider that a man reaching age 65 today can expect to live 19.3 more years; a woman,  21.7.

This means that, from that age 65 standpoint, ignoring any time-value of money considerations, one can expect to collect 193,000 or 217,000, respectively from a hypothetical $10,000 benefit starting at age 65.  If the retirement age was raised by a year, then, again based on this simplistic calculation, one’s lifetime benefit would be 183,000 or 207,000, about a 5% decrease.

But this sort of “total future benefits” calculation (even disregarding the fact that actuaries the world over are cringing at the math) is simply not a reasonable calculation unless one also takes into account increasing life expectancy.  Most retirement systems do that implicitly in assuming that their retirement age increases pair with historic and forecast future life expectancy; the Danish system explicitly builds this in explicitly, with the intention that the average length of time in retirement stays constant at 14.5 years.

The bottom line is this:  we need a better, more targeted, solution for those unable to work up to their official Full Retirement Age, let alone maximum benefit age, than the existing method of allowing early retirement at the cost of significant benefit reductions.  And creating and implementing this solution will allow us to consider the appropriate retirement age/Social Security benefit commencement age, for the majority of the population, in a more sensible manner.

****

Bonus content:  how does the US retirement age compare globally?  

In the news yesterday:  despite public protests on the matter, Russian President Vladimir Putin signed into law a pension reform bill which increases the retirement age, formerly age 55 for women and 60 for men, to age 60 and 65, respectively.  (See Radio Free Europe for coverage.)

From an American point of view, one might be surprised that the retirement age was ever this low in the first place, or that retirement ages were and still remain different for men and women.  (This is not unusual, as I wrote back in March.)  One might have even a bit of sympathy for Russian men, though, whose life expectancy is a mere 66 years; the Independent (UK) reports that 40% of men will not live to retirement age under the new law.

But here’s something else readers might not notice:  there is no early retirement option available to Russian retirees.  In fact, the Moscow Times reports that the government attempted to mitigate concerns over livelihood in those pre-retirement years by criminalizing the firing of workers in the five years preceding retirement.

In contrast, American workers in the years prior to normal retirement age who exhaust their unemployment benefits, or those who consider themselves likely to die young because of family history or their own poor health, are likely to simply start receiving Social Security benefits with the early retirement penalties.  In fact, our system, despite the official “full retirement age” of 67 for by now most workers, actually provides “full” or maximum benefits at age 70, with reductions or de facto reductions for benefit commencement prior to that age — a provision that’s either a bug or a feature, depending on your perspective:  it provides greater flexibility but at a cost in lost benefits that may create financial hardship down the road.

And here’s what’s worth knowing:  the Russian system with a fixed single retirement age, is actually the norm.  Our system is unusual, as other countries which allow for early retirement generally pair that with substantial work history requirements.  Here’s a listing of Western European countries, from Social Security Programs Throughout the World:

Belgium:  age 65, rising to age 67 in 2030, or age 63 with 41 (42 in 2019) years of coverage (work history).

Denmark:  age 65, rising to age 68 in 2038 with further life-expectancy increases afterwards.  No benefits prior to this age.

France:  the “normal retirement age” is 62, but only with those with 41 years of coverage, rising to 43 years by 2035 (“coverage” also includes 2 years’ bonus per child and unemployment benefit periods); those without sufficient work history can retire at age 62 with a reduction of 5% for each year of missing work history, or can retire at age 67 in any case.

Germany:  age 65 and 7 months, increasing to age 67 in 2029, or age 65 1/2, increasing to age 65 in 2029, with at least 45 years of contributions.

Ireland:  age 66, increasing to age 68 by 2028.  No benefits prior to this age.

Netherlands:  age 66, rising to age 67 and 3 months by 2022.  No benefits prior to this age.

Switzerland:  age 65 for men or age 64 for women; early retirement is available at age 63/62 with a reduction of 6.8% per year.

United Kingdom:  age 65 for men or age 63 for women, rising to 67 for both in 2028.  No benefits prior to this age.

So I invite readers to contemplate this list of retirement ages and imagine that when Congress had increased the normal retirement age from 65 to 67 in 1983, they had likewise increased the early retirement age a corresponding amount, or removed the option entirely, or added a similar work-history requirement.  What would have happened?  Would Americans have adjusted their retirement patterns accordingly, and would employers have adjusted their expectations for when a worker is “too old” to hire or to stay employed?  Does the “safety net” element of early retirement for those who are unable to find work or are in ill health but not to such a degree as to qualify for disability benefits, come at too high a cost, in terms of benefit reductions, compared to other ways of providing those benefits, such as extended unemployment benefits for near-retirees or partial-disability benefits?

 

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.