Capital Commerce1 quotes Edward Lazear, the Chair of Bush’s Council of Economic Advisers, answering a question about productivity growth:
I wouldn’t necessarily say 3 percent. But I would expect that we could expect to see high rates, perhaps not quite at the 3 percent level, but somewhere higher than 2 percent. I would expect somewhere closer to 3 percent … If I’m thinking about long-term productivity growth and asking, “Do the fundamentals exist for persistent high productivity growth in the upper 2 percent range?” I think we can still be there, again as long as we continue to maintain policies that are consistent with an open economy.
If Bush’s chief economist is right, then there’s no Social Security shortfall.
To understand why, you have to understand that the 75-year shortfall is a prediction made by the trustees of the Social Security Administration. According to the trustees, we will stop being able to pay 100% of Social Security from payroll tax revenues and the trust fund2 in 2040. According to the Trustees, “This unfunded obligation represents 1.9 percent of future taxable payroll and 0.7 percent of future GDP, through the end of the 75-year projection period.” This means that unless we raise payroll taxes 2%, after 2040 we will be faced with a choice of either paying for some of Social Security from general taxes, or of cutting Social Security benefits.
Current payroll taxes, for those of us who get paid under $97,500 a year,3 are 6.2%. In addition, a matching payment of 6.2% is paid by employers, for a total payroll tax of 12.4%.4 So if we increase the payroll tax 2%, by raising the rate to 7.2% for both employees and employers, the Social Security shortfall would cease to exist. I can’t say that I’m thrilled at the prospect of losing an additional 1% from my paychecks — which for me would be about a loss of about $2.30 a week, or $4.60 a week if we assume that my employer docks my pay to make up for its part of the payroll tax — but it’s not an unbearable expense.
And that’s one way to completely fix the Social Security shortfall. (There are other ways, of course). Not as impossible as we’ve been led to believe, is it?
But in reality, it would probably take considerably less than $4.60 a week out of my paycheck to completely fix the Social Security shortfall.
Predicting the future isn’t simple; the Trustees have to make certain assumptions. One of their assumptions is that productivity increases for the next 75 years will be, on average, 1.7% a year. If they’re right, then the next 75 years will feature the lowest productivity gains in American history. (For comparison, over the last 75 years average productivity has gone up around 3.5% a year).
In other words, the prediction that Social Security will have a shortfall in 2040 is based on pessimistic assumptions about the economy. But what if the next 75 years aren’t actually the worst 75 years in US economic history?5 Well, then, the year in which Social Security has a shortfall will be later than 2040… if it ever comes at all. As the graph (which I swiped from Kevin Drum) shows, the predicted “doomsday” for Social Security keeps receding into the future.
In 1994, they were saying that Social Security would have a shortfall in 2029. Today, it’s predicted for 2040. What happened? A few things, but a major factor is productivity. For 16 years, we simply haven’t had the disaster in productivity that the Trustees’ projections are based on.
That brings us back to Edward Lazear’s prediction of near-3% growth in productivity. If he’s right, then maybe there won’t be any Social Security shortfall. Higher productivity means, at least in theory, that workers earn more; workers earning more means that, without raising taxes, payroll tax revenues go up. If we really have “persistent high productivity growth in the upper 2 percent range,” that would reduce the Social Security shortfall to nothing, or to a trivially low amount.6
So if the Chair of Bush’s Council of Economic Advisers is right, maybe we don’t need to change a thing to save Social Security. But is he right?
Probably not. The reality will almost certainly end up being somewhere between the extremely pessimistic numbers of the Trustees and the extremely optimistic numbers of the Bush administration’s head economist. But the change required to restore Social Security to full solvency will be not be enormous; instead, only minor changes will be required, and those changes don’t have to be made immediately.
(This is the first of an intended series of posts on Social Security.)
- Curtsy: Economist’s View. [↩]
- Full disclosure: I originally forgot to include “and the trust fund” in this sentence. Thanks to Sam in comments for pointing out the error. [↩]
- Only the first $97,500 of income is subject to payroll taxes. So someone who is paid $195,000 a year pays 3.1% a year in payroll taxes, rather than 6.2% a year. [↩]
- Many economists believe that in effect employees pays the entire 12.4%, because employers pay workers 6.2% less than they would if the employer’s matching payment didn’t exist. I think that’s a bit dubious; it assumes that the wage-setting mechanisms of the labor market are a great deal more efficient than they are in real life. [↩]
- Or are the worst, but not by as large a degree? [↩]
- I admit, I’m oversimplifying a bit. For a less simplistic discussion, see Brad DeLong. [↩]
Pingback: Pacific Views
Pingback: Grasping Reality with Both Hands: Brad DeLong's Semi-Daily Journal
Pingback: IST Control Center
Pingback: Creative Destruction
There’s another possibility here. The assumption that productivity growth will rescue social security is based on the notion that the wages of payroll-tax-paying workers (for practical purposes, those earning less than the cap) will rise in concert with productivity, as they did for essentially all of the 20th century.
The 21st century, in contrast, has been market by conditions under which wages have remained flat while measured productivity has increased rapidly. Only the tiniest fraction of the productivity gains since 2001 has been reflected in the wages of payroll-tax-paying workers.
So if you imagine a world where executives get to outsource whatever jobs they choose, coerce workers into “givebacks”, and increase uncertainty by employing only contract and part-time workers while the federal government rolls back protections for union and non-union employees alike, it seems quite possible that the wages available for paying social security taxes will be insufficient to pay mandated benefits. Obviously a situation whose solution involves cutting those benefits further.
According to the trustees, we will stop being able to pay 100% of Social Security from payroll tax revenues in 2040.
Sorry–that’s slightly wrong. (I’m an actuary; it follows that I’m a detail-freak.)
We will stop being able to pay 100% of Social Security benefits from payroll tax revenues AND THE TRUST FUND in 2040. Current-period payroll tax revenues alone are predicted to be insufficient in 2017. And remember, paying off the Trust Fund comes out of general revenue (income tax), and so competes with other spending (national healthcare, more and bigger guns, etc).
In other words, on a cashflow basis, we will need to start using income tax to pay Social Security benefits in 2017.
I’d be interested to see the chart of predicted dates at which on a cashflow basis we will no longer be able to depend purely on payroll taxes. I suspect it has a similar feature to the chart Amp printed (the predicted date moves in step with the prediction date).
Paul’s point holds (but I think Amp is planning to cover it in a future post).
Thanks for your insight on this, Sam.
Of course there is no “trust fund.” As we usually think of that concept, a bunch of money sitting around in a bank account someplace. It doesn’t exist, it got raided out long ago for general fund purposes, and all that’s left is an IOU from an insolvent debtor. (That’s us, folks!)
So, this is what it comes to, if I understand this situation properly. In 2017 (which sounds like the Far Distant Future, but which in fact is only 10 years from now) the payroll tax is projected to be inadequate to pay current liabilities to Social Security recipients as of that year and years subsequent. And that’s if all these rather optimistic growth factors are correct!
So, what happens in ten years? Social Security payments begin to come out of regular income taxes. But wait, we’re ALREADY spending more of this money than we’re taking in, by quite some. This means that we’re going to start running truly gigantic deficits.
This is OK if you think we can just keep doing this forever, on an ever-increasing scale. Not to worry. The whole “crisis” is fictive.
Social Security, as currently implemented, is not sustainable for two reasons:
1. Benefits are indexed to wages, so productivity growth really doesn’t help us that much in the long term. Sure, revenues go up, but expenditures go up by the same factor.
2. Life expectancy is going up faster than the retirement age. As long as this trend continues, the ratio of workers to retirees will continue to decline.
If we address either of these factors (by making benefits grow slower than wages, or by increasing the retirement age faster), we can fix Social Security. Raising payroll taxes is a temporary fix, and an ill-advised one at that.
Also, what SamChevre said. Nothing meaningful changes in 2040.
Paul:
The 21st century has been going on for six years and started at the peak of a huge economic boom. You can’t extrapolate long-term trends from that.
Pingback: Word Munger » Coolest graph ever!
In other words, on a cashflow basis, we will need to start using income tax to pay Social Security benefits in 2017.
As previously noted, while financing Social Security may or may not look like a problem in isolation, it looks like a problem in the context of the larger federal budget.
What remedy? The US is productive enough to pay for quite a safety net. Due to globalization, however, the benefits of US productivity flow disproportionately to a narrow class of Americans. Old formulas that assumed that payrolls would reflect productivity, and therefore we could pay for government services out of payroll taxes, may break down. We need a different formula for capturing the productivity gains in order to fund the safety net.
Of course, wealthier people will scream bloody murder. The wealthier people get, the less self-interest they have in funding the safety net. And the wealthier people get, the more influence they have on public policy. Politically, it’s a train wreck in the making.
It is the same story told over and over. Every so often someone says that Social Security Fund is going bankrupt and needs to be trashed, when all that needs to be done is the fund needs adjustment. Politicians think that most everyone forgets and that if they say the same thing over and over, even though it is untrue, we will accept what they are saying as true. Why is that? Hmmm…..
Your 1st Soc. Sec. post was a treasure! Thanks! Keep up the good work.
I did a calculation a couple of years ago. As I recall, I projected the doomsday pediction shift forward to 2042, and calculated that by then the doomsday date would be 2099.
At which time, I suppose, Prince’s grandkids could party like… well, you get the idea.
Exactly. No matter how the math works out for the individual program, Social Security is dependent on the long-term solvency of the government. Right now we’re counting on imaginary money, and our current economic policies aren’t making it likely to materialize anytime soon. They’re not meant to, either, since the money’s getting funneled to people who have no need for the programs they’re robbing.
Oh and look, there’s a bathtub right there, with such a nice open drain.
Sam, thanks for the correction. I’ve edited the post to fix the error.
Right you are.
Brandon,
No you can’t extrapolate from 6 years, but I’m not sure why Paul set the cutoff for productivity growth producing very poor wage growth for anyone except the rich and the very rich at the turn of the millenium.
Real wage growth has been largely stagnant for a lot longer than that. And that apparent stagnation mostly hides a massively increasing income inequality.
Susan, can you document that the U.S. has been defaulting on paying for treasury bonds? Because that’s what being “insolvent” would mean, in this context.
And no, you can’t document it, because it hasn’t been happening, and no credible economist believes that it will happen anytime in the forseeable future.
US Treasury bonds are quite possibly the safest, most conservative investment in the world; certainly safer than the stock market, which is where money deposited in a bank sometimes winds up being invested.
First of all, the economic predictions you’re talking about are actually very pessimistic. Second of all, by the time 2017 rolls around, the predicted date will have certainly moved forward — probably to around 2027.
No, that’s not true. Currently we take in more in payroll taxes than we pay out in Social Security payments. According to current predictions, that will remain true until 2017 or so; but it’s a safe bet that the 2017 date will be pushed forward.
Another way to look at this is to compare growth under Democratic and Republican presidents. If you do that, it becomes quite clear that one way to avoid any Social Security crisis is to simply stop electing Republicans.
In fact, over the past 75 years, the cumulative growth rate under Dems is roughly 10 times that under Reps.
As I explained in a MyDD diary back in April, 2005, “We’d All Be 3 Times As Rich, If Not For Republicans!” But that’s 3 times as rich as we are NOW:
GDP 1930: $865.2 billion
GDP 2004: $11,000.3 billion
GDP 2004 w/Dem Pres growth rate: $32,236.7 billion
If we’d had nothing but Rep presidents, we’d be more than 3 times poorer:
GDP 2004 w/Rep Pres growth rate: $3,059.7 billion
Put the two together, and a Rep-President-only US over that past 75 years would have a GDP just 9.5% that of a Dem-President-only US.
It was just a quick-and-dirty exercise. But the results are so stark, it hardly matters.
Of course correlation doesn’t prove causality, yadda-yadda-yadda.
But considering all the endless propaganda about how the GOP understands business and the Dems don’t–do you think this comparison would be a well-kept secret is the numbers were reversed???
Or would it be something kids had to learn in order to graduate from high school?
How about, second grade???
I’ll write more about income inequality — which is, far more than longer life spans, the reason for the Social Security shortfall — in a later post.
Nobody Really wrote:
No, that’s not true. It looks like a problem if you combine SS with Medicare, as you did in that comment — because unlike the social security fake “crisis,” there really is a Medicare funding crisis. In fact, there is probably no more essential problem facing US finances right now, which is why it’s especially insane that people are worrying over Social Security, which is perfectly healthy.
The graph at the top of this webpage gives a good idea of what I’m talking about. Social Security spending is projected to go up only slightly; it is Medicaid and especially Medicare that are going to grow up to eat Detroit. (As it were.)
The real solution is to do something about the skyrocketing cost of medical care. And there’s a lot we can do there, in my opinion. But that’s an entirely different subject than Social Security, and conflating SS with Medicare and Medicaid only muddies the real issues.
Dephenistrated wrote:
You’re right, of course. But I don’t think there’s any reason to think the long-term solvency of the US government is in doubt.
And all money is imaginary, in a way. But what else should we use? The alternative systems are even worse. Meanwhile, as imaginary money goes, US Treasury Bonds are about as solid as we could possibly get.
Of course, it’s conceivable that the US will default on those bonds. But anything is conceivable; that doesn’t mean it’s likely to happen.
There’s really nothing pernicious about the lack of an actual trust fund. That excess Social Security revenues were to be used for general government expenditures was stipulated in the original Social Security Act. That doesn’t necessarily make it right, but I don’t see that there was any real alternative.
Government can’t just save money like you or I can. Sure, it can throw cash in a vault and pull it out later, but the actual effect of this, when done on a large scale, is to cause deflation when it puts it in and inflation when it pulls it out.
Inflation being essentially a tax on cash holdings, this isn’t much different in principle from cutting income taxes when SS runs a surplus and raising them when SS runs a deficit. But I do think that in practice the distortionary effects of monetary expansion and contraction are probably worse than the distortionary effects of raising and lowering taxes.
Anyway, the government can’t save money, but it can invest it. One way to do this is to invest it directly in productive enterprises, i.e., by buying stock or starting up companies. People are generally a bit leery of this, IMO for good reason. Another way is to invest in pro-growth policies that will drive future revenue increases. These might include investing in education, infrastructure, or scientific research, but they might also include returning the money to the private sector in the form of tax cuts.
This isn’t to say that I necessarily approve of what the Bush administration is doing (I don’t, because they’re spending way too much money), but there’s nothing inherently wrong with raiding the trust fund.
One of those rare moments…
I’m pretty sure I agree with that.
First, it’s totally unfair that Amp gets to embed graphics and I don’t!
For all the mathematical and astrological conversations about the pending death of Social Security, the easiest way to figure out when Social Security is going to hit the wall is to look at how it works.
Workers’ Salaries * Tax Rate = Money for Benefits
Money for Benefits / Recipients = Benefits per Recipient
Real simple math — I learned about multiplication and division in 2nd grade.
What is “Worker’s Salaries”? That’s the number of workers times their average salary. Keep “number of workers” and “average salary” in mind. What is “Benefits per retiree”. It’s “average salary times replacement rate”. What’s “Replacement Rate”? It’s the percentage of ones salary that Social Security attempts to be. At present it’s about 60% on the low end and 30% on the high end. As someone else pointed out, if salaries rise, so will benefits.
Doing the math, that’s
Average Salary * Tax Rate * Workers = Average Salary * Replacement Rate * Retirees
Remove “Average Salary” from both sides and it reduces to
Tax Rate * Workers = Replacement Rate * Retirees
Solving for Tax Rate,
Tax Rate = (Retirees / Workers) * Replacement Rate
In short, to remain solvent, the tax rate has to be set to the ratio between workers and retirees times the percentage of their salary that they receive from Social Security.
If anyone disagrees at this point, I encourage you to study how Social Security works and perhaps to repeat 2nd grade.
What this shows is that if the Workers to Retirees ratio decreases, either the Tax Rate goes up, or the Replacement Rate goes down.
What insures that Social Security is going to become insolvent is the population charts at this website.
It shows that the ratio between Workers and Recipients is going to get progressively worse from now until forever. There’s no evidence that the age distribution is ever going to improve, from a Social Security solvency perspective.
If you look at Western European nations, their birth rates, and their age distributions, you see what’s coming our way. Italy and Germany are good examples. Right now we have a birth rate of 2.09 children per woman, which is below the replacement rate. We make up our growth on immigration.
Median age will continue to increase — it’s been increasing for quite a while, and it’s just going to keep on increasing.
FCH,
But that is true of any system of retirement support. As you get more retired people relative to the number of workers, the retired people need to take a larger share of the total national income if they are going to keep the same income level relative to the national average income. This is true whether they get their retirement income by owning stocks and bonds or by receiving current payroll tax revenue through social security. So at some imaginary point in time, half the adult population is retired and receiving benefits, and the other half is working, and if we agree that the retired people can retire on 2/3 their working income, then 1/3 of the potential income of each working person goes to support a retired person. This chunk of income can come in the form of payroll taxes, or it can come in the form of decreased wages because of pressure to provide adequate corporate profits to retired shareholders, but it has to come from somewhere. If half the population is not working, then the other half has to either work harder for the same resources or work the same for less resources.
Although working harder may be relative. If the average wage in current dollars were twice what it is now, we could all be paying 50% payroll taxes to support a population in which the majority had retired, and still live as well as we do currently. Productivity increases matter (as does distribution of wealth).
J, H of C writes:
Your math assumes that the “average salary” earned by present-day workers is equal to the “average salary” that retirees earned during their earning years. (If they weren’t equal, it would be illegal to “remove ‘average salary’ from both sides.”)
But you’re wrong; because productivity goes up over time, the “average salary” earned in the present is on average significantly higher than the “average salary” retirees once earned. So the “math” you present here doesn’t work.
With all due respect, Julie, I encourage you to drop the “I’m so superior, anyone who disagrees with me is an idiot” ‘tude.
In the past, Social Security (and our economy) has survived far more extreme hits to the worker/non-worker ratio than anything that’s predicted for the next 75 years. In 1950, the worker-to-retiree ratio was 18 to 1; fifteen years later, it was 4 to 1. That huge change in worker/retiree ratios didn’t destroy Social Security, because (among other reasons) productivity gains meant that the workers in 1965 were adding more to the economy than workers in 1950. Compared to that, the drop from today’s 3 to 1 ratio to 2080’s predicted 2 to 1 ratio is nothing, and can easily be accounted for by a combination of productivity rising and minor increases in taxes.
Well, that and the 140% increase in the FICA tax rate. Historically, Social Security has been kept afloat by tax hikes, not by productivity increases.
Actually, by both tax hikes and productivity gain (what is the increase in productivity from 1950 to the present? certainly not 600%, but certainly not 0%).
Also, the rates have been progressively going up less and less, going up only 50% in the past 37 years. In that time, the real GDP per capita doubled, so a 50% increase in rates meant roughly a tripling of per capita contribution (less because of the growing income inequality). I’m sure someone has done the math for what the rates will look like in 2080 (using realistic productivity rate projections) and how post-FICA tax average incomes will compare to current post-FICA incomes, maybe someone else will find it and post it (I think I’ve seen a Crooked Timber-ite play around with these numbers).
My off the cuff guess is 15% FICA tax (150% growth to cover a ratio changing from 1:3 to 1:2, seems a bit high) on an average income of $160,000 (2% productivity growth for 73 years) in year 2007 dollars. I’ll probably have been dead for a few decades by then, but my heart doesn’t exactly quail at the prospect of the average American (perhaps one of my nieces’ 50 year old kids) only earning $138k pre-income tax rather than $151k pre-income tax (as it would be under the current FICA rate). Lets say a FICA of 33%, and somehow I’m still not weeping for the workers who are forced to get by on an average of only $100k/yr pre-tax income.
Of course, if we continue our current trend of income inequality growth for the next 73 years, the median worker will be making more like $60k, and the 25 percentile worker will be making more like $10k. Very rough guesses, and plenty of time to fix the system, but that (and health care costs) seem like much more real problems than paying for more and more of the population to be on social security. Also, the oceans rising up and drowning New York City.
There are a lot of other things I worry more about my niece’s grand kids having to deal with, but paying for my niece’s retirement really isn’t one of them.
Amp,
“Productivity Increases” are handled by Social Security through “Cost of Living” increases.
I’m half-way between Amp and Susan on the Trust Fund issue.
I don’t think the US is at all likely to default on its bonds. (Here, I’m with Amp.)
BUT–paying off the bonds will come out of current-period cashflows. The bonds are assets of Social Security, and liabilities of the US government. So from a “government as a whole” perspective, they are irrelevant.
An analogy would be the 401-K loan I’m currently paying off. It is an asset in my 401-K, and a liability to my current spending. It’s not an asset to me–it’s my promise to myself to spend some of my income on one thing instead of another.
That just brings me back to my original point. Paying off the Trust Fund comes out of general revenue (income tax), and so competes with other spending (national healthcare, more and bigger guns, etc). In 2017 (or whenever–but since that date is based on demographics rather than productivity it will move less), when we start paying off the Trust Fund, our ability to spend on other things goes down relative to the previous year; in 2040, maintaining current benefit levels will not change spending on other things at all reltive to the previous year.
Sam, I agree with you, but really, I don’t see that as being very different from the situation before 2017. Money is fungible, after all. It’s currently standard practice to talk about the federal budget as if payroll taxes and SS payments were part of the whole budget, rather than separate accounts.
Nor is the spending on SS going to rise to levels that the budget can’t deal with; the rise in SS spending is actually scheduled to be relatively small. As I said earlier, it’s the rising costs of Medicare and Medicaid that ought to have people frightened.
Especially because we’re already seeing the cuts there – they’re not way in the future, nor are they hypothetical.
I really like how Charles put this:
Takes skill to make me laugh at “the oceans rising up and drowning NYC.”
My earlier comment wasn’t specific to SS and the Trust, I guess (and so probably was just plain off-topic), but I still maintain that we can’t run on a deficit indefinitely :(
Thanks!
The oceans rising up and drowning NYC probably happens somewhere in the 22nd century, but I don’t really expect the barrier island my family has gone to for the past 38 years to still be habitable in another 80.
If only our species had some kind of technological or engineering solutions to protect economically critical areas from rising seas…alas, with only three centuries in which to work, I doubt we’ll be able to manage anything for NYC. Better start the evacuations now. I hear Kansas is nice.
True, NYC will merely be diked to hell and back (at some amazing cost, repeated for each coastal city, assuming we can figure out how to budget for it, see New Orleans – how many years have we know that it should have dikes to survive a class 5 hurricane? how many people have argued that it should just be abandoned, rather than being rebuilt to survive a class 5 hurricane? how many years has it had dikes that couldn’t make it through a class 3 hurricane?).
Of course, dikes won’t save Topsail Island or the rest of the outer banks, or the Mississipi delta region.
Sorry, I suppose this belongs on the thread about Amp misspelling Israel. :p
The difficulty faced by New Orleans was the contingent nature of the event. It could have happened in 1856; it could have happened in 2020; it might never have happened at all. That makes it hard to run budget numbers, or to model the cost-effectiveness of various strategies.
But the coastal cities will definitely flood – or so you say – and so we can plan accordingly. Sure, there’s some wiggle room as to exactly when (2210? 2310?) but unlike Nawlins, you know it’s going to happen. Budgeting to deal with a disaster so far in the future, then, is super-trivial: set aside some budgetarily modest chunk of money (say, $10 million per coastal city) and invest it. At a 4% constant rate of return, in 200 years you’ve got about $25 billion per city; that should be plenty, what with the anti-grav technology and the jetpacks and what not.
If Mr. Sun holds off for another century, then you’re even more golden: $1.2 trillion per city. So, no worries there.
There are what, maybe 50 big US coastal cities in a similar situation? That’s only $500 million. Al Gore could raise that in a year.
Now that I’ve solved this one, I’m off to the Middle East. Gotta bust a few heads to get my peace plan adopted.
Except that sea level rise increases the chance of contingent events being disastrous as well as eventually simply directly flooding the cities. NYC will start being vulnerable to being flooded by a vagrant hurricane or a massive Nor’easter long before the daily high tide would run down Broadway.
Look at the Netherlands, which builds its dikes against 1000 year floods (whereas we weren’t willing to protect New Orleans against 100 year floods). Look at the Netherlands, which has recently decided that its 1000 year flood standard is inadequate, because the next century or two of sea level rise is going to completely change what a 1000 year flood looks like.
Given that 1 half of the political spectrum’s response to the question underlying New York City (or Miami) getting flooded by a storm 100 years from now is “La La La La”, my hope for people being willing to put up the money to spend a trillion dollars on dikes for all the low lying cities in the US is pretty low. My bet is that New Orleans is not the last US city to go under water.
Of course, I’ll be dead by then, so who cares, but it looks like a more real problem than the imaginary impossibility of providing for the needs of retired people 80 years from now.