The Economics of Raising the Social Security Maximum Taxable Earnings Amount

A RBC reader named Mike M. just taught me several things about the Social Security tax.  While I have published in the Journal of Public Economics, I didn’t know the details of the 2012 Social Security tax.   Until I read these details about the non-linear taxation, I was unaware that the maximum taxable earnings cap is set at $110,100.     So, permit me to do some algebra and then I will make a progressive tax point.  Consider Mr. A and Mr B.    Mr. A earns $110,100 while Mr. B. earns $220,200.   Given the Social Security tax rate of 4.2% and given current rules, they both pay $4624.2 per year to Social Security and their employer pays 6.2% or  $6826.2 per year to Social Security.    This would appear to be a superstar subsidy.   Suppose that UCLA must choose between hiring a $500,000 Nobel Laureate or 5 Assistant Professors.   If I’m reading these rules correctly,  UCLA would face a choice between paying $6826.2 per year for the Laureate’s Social Security or a rough total of $30,000 for the 5 rookies.  That’s what I meant by the “Superstar Subsidy”.

What would happen to the Social Security System’s stream of revenue if the tax rate was cut to 4% but there was no cap?   I think that revenue would rise sharply, I don’t see what unintended consequence would be triggered and I would support this policy change.    I do see that firms who used to face a zero marginal tax on employees making over $110,000 might reduce their hiring of this group but this is an empirical question that merits research.    Who is at the margin? Facing a 6.2% marginal rate on such $110,000 workers who would demand fewer and how many fewer jobs?


Author: Matthew E. Kahn

Professor of Economics at UCLA.

39 thoughts on “The Economics of Raising the Social Security Maximum Taxable Earnings Amount”

  1. Do you mean that you didn’t know an earnings cap existed or you didn’t know at what earning level the cap is applied? If the former then I’m jaw-droppingly gobsmacked and will have trouble taking you seriously in the future. I’ve found your posts consistently interesting up till now. Even this one works pretty well (after the woeful admission), that “superstar subsidy” is a real-world application of the cap which never occurred to me.

    Keep up the good work, but please don’t scare me with such scathing self-revelatory honesty again.

    1. C’mon Karl. When I was a thirty-something faculty member I had no idea about what my retirement package was. It was not a factor in my taking the job, but I’m sure glad there was one!

      I had asked Matt to look into this because, now that SS apparently on the chopping block, I wondered how much revenue would be generated if the cap were increased (or, as Matt hypothesizes, eliminated). And in my query to Matt, I explicitly said that I know that this is not his area of expertise, so don’t fault him for not knowing.

      Perhaps someone who has the US salary distribution data can estimate how much additional revenue would be generated as the cap increases.

      I’d also like to know the history of the cap: what it was initially and how (if at all) it has changed over they ensuing decades, especially wrt inflation. Do any of you know this, or know where the data resides?

      1. It looks like Table I of this covers it; I think you’ll have to look further or adjust for inflation yourself (I glanced at 1982, it’s gone up faster than inflation since then, but then I think I recall that Social Security benefits have been set by a factor that is larger than the inflation rate, and that might be the better comparison).

      1. The cap at $110k was originally set to cover 90% of total income earned. These days it covers about 86%. The cap should go to at least the same threshold today.

  2. The math of your hypothetical is surely worse: you’ve overlooked health benefits. I don’t know whether aggressively recruited Full Professors in Named Chairs with Fancy Parking Spaces get a better health insurance package than the newly hired Assistant Professor, but I actually rather doubt it. So there’s another significant cost likely to be multiplied by five.

    On the other hand, your numbers seem to me to be way off, on both sides. I’ve known a number of assistant professors (including some at UCLA), and although I’ve only discussed it in general terms, so far as I know none gets a starting salary of $100K (a quick glance at Google says $84K at UCLA, which would be a bit generous for a new assistant professor from the conversations I’ve had). My acquaintances are in the Biological Sciences, where pay is typically higher than in the Humanities (I don’t know if UCLA makes this distinction); it may be higher still in some Engineering disciplines or elsewhere. I also know a couple of Nobel Laureates (I know none at UCLA, which has a few), and they do quite nicely, thank you – but I don’t think their University salaries are typically boosted, and certainly not three-fold above Full Professor as you suggest. They might get more funding for their lab (they definitely will if they’re recruited already having won or already tipped to win), and they can make money consulting, speaking, and serving on boards, but their University salaries typically remain modest.

    Also cutting the other way, in Biology and I suspect in any of the hard sciences a prominent, successful person likely to be aggressively recruited as you describe pays their own salary through grants, and through overhead on those grants is a major profit center for the University. There is the size of the startup package to be considered (though I suspect the multiple may be lower than you suppose there, too), but a bigshot is expected to already have a track record of successfully obtaining the funding to run a large lab, while the newly hired assistant professor may be struggling to get their first smaller grant proposals funded.

    1. If your comment on “health benefits” referred to Medicare, Medicare taxes are not capped, only Social Security taxes are capped.

      Further consider that the higher the salary, the larger the incentive.

      Outside of academics, patent and product generating research & development and medicine, there are no objective measures of accomplishment. A Nobel Laureate may be worth a half million but a CEO, really almost never on ability. Showmanship, ruthless generation of unsustainable profits and connections in the Old Boy Network count for bona fides. I worked with many engineers who earned MBA’s to move into middle management and universally they held the skills and knowledge of the business and finance students in contempt.

      The managers most highly prized by financial “analysts” demonstrate why the root of “analyst” in this case is “anal”. The classic case of Albert Dunlap, widely hailed as a turnaround genius before being convicted as a multiple felon, seems to have not cooled “analysts” on psychopath CEOs:

      1. My comment on “health benefits” referred to health benefits – i.e. medical and dental insurance. They are part of the cost of employing someone, which Kahn was supposedly exploring, and unlike some other costs (such as Medicare taxes) they don’t scale well with salary, especially at large public-minded employers like the University of California that often give all employees the same health insurance.

        As to the rest of my comment, I remain rather appalled that Kahn – who, after all, is a professor at UCLA and must know more about the pay and benefits than I do – decided to illustrate the situation he was describing using an example that so completely falls apart: none of the salaries he describes are accurate, and in most cases the people he describes pay their own salaries off of grants and indeed make a profit for the University, such that the cost to be considered in their recruitment is not their (payroll-taxable) salaries but the startup money needed for them start their research and start (hopefully) getting grants – and the established big names (Nobelists for example) Kahn describes are likely a better bet in terms of bringing those grants.

        Regarding your comment, I am appalled by the CEO pay I see reported in the news, and by the studies demonstrating its lack of correlation with CEO accomplishments (indeed, I’ve seen reports that CEOs often are richly rewarded with money and attention for making foolish acquisitions and mergers that later prove disastrous). I can’t really speak to your concerns about the rewards for achievements and more generally the pay structures in middle management.

  3. Well, the claim has been that one pays for one’s Social Security. That it’s not charity. And if you charge a basket baller who makes $3.5 million a year four per cent, and you only give him the same miserable pittance everybody else gets at pension time, you are making him do a charitable donation to those he has out-earned.

    You may choose to do that. But it’s different from what Roosevelt claimed it was to be.

    1. On the other side of that coin, it was brought into being as a “safety net,” not as a “retirement savings account.” That “safety net” terminology would seem to favor means testing for payment of benefits. Hah! Try sneaking THAT past AARP.

      It’s a thorny problem. Real life is more complex than any ideological oversimplifications.

    2. Not really. I think you would be suprised at the number of former professional athletes who are glad to receive social security when their careers are over. These people typically earn big money for only a few years, at most. They area often poor money managers, too. That’s the point of social security. No matter how poorly you manage your money, no matter how far you or the markets fall, you can’t lose your social security. It’s always going to be there for everybody, no matter what.

      1. Mitch, your “not really” might be misinterpreting means testing. As generally used, it denotes “means” at the instant of application, irrespective of how you reached that point.

    3. The (well, one) point of the OP is that the rich do NOT pay four percent. They pay four percent (or 6.8% or whatever assuming the payroll tax is restored to what it was before the financial crisis) on the first $100k. So your basketball player is actually paying 0.11% of his salary toward SS. Which counts as a highly regressive tax, and morally indefensible in my book.

  4. dave schutz’s explanation for having a cap is the one I’ve always heard; I also have heard that the argument for raising the cap is so it will more or less reflect the inflation that has occurred since it was last set.

    That said, the Social Security issue of the moment is not the level of the cap, or that SS is “about to go broke” (because it isn’t), it’s that benefits are about to be slashed in a very sneaky way through “chaining” them to the CPI. If this goes through, very many old and disabled people will be seriously hurt.

    To me that is more important than saving a employers a few bucks on their share of SS contributions.

    1. benefits are about to be slashed in a very sneaky way through “chaining” them to the CPI. If this goes through, very many old and disabled people will be seriously hurt.

      I’m counting on this being a trial balloon. I called my Senator and House Rep and the WH (didn’t get thru) to let them know I don’t appreciate this threat.

      1. Yes, the WH line was very busy yesterday. I forgot which one I ended up using (there are two phone numbers) but I folded a very big pile of laundry and still had lots of time to twiddle my thumbs before a volunteer answered. I am hoping both of our efforts help to prick this balloon.

    2. Eh. From what I can tell without anyone saying concretely what was in the proposal, it is more complicated, and more positive, than this. Yes, it involved using Chained CPI, which has the effect of producing slower growth in benefits and thus constitutes a cut. However, it was paired with a change in the formula which would have substantially increased the benefit calculation for those at the lower end of the benefit structure. For these individuals, the benefit increase from the formula would in almost all cases (it depends upon how long they live) have produced an increase in lifetime benefits. So the combination was subtle attempt at making the SS system more progressive.

      If that’s an accurate reporting of what the proposal actually was, count me as being in favor.

  5. Because the cap is applied per individual rather than per household, it also subsidizes one-career couples relative to two-career couples. Suppose Alice and Bob both work for $100K a year and at home share responsibility for housework and childcare. The responsibilities at home make each of the unavailable for some duties at work such as extra hours or travel. Now suppose Bob stays at home, doing all household and childcare work, while Alice devotes herself completely to he job. Thanks to her newfound availability for every high-profile task at the office, she gets promotions, raises, and bonuses to make $200K per year. The extra $100K made by one earner is above the cap, whereas the extra $100K made by a second earner is still taxed.

    1. Greta point. Plus, a survivor spouse who never worked gets all the benefits of the earning spouse after his death, whereas a survivor spouse who worked 30 years gets no increment even tho she contributed to the fund for three decades. It is a tax on working women, single and married, in favor of non-working married women. In other words, a huge reward to those who live out patriarchal gender norms.

  6. I surmise that removing the cap would not push down job creation or remuneration very much atall. Inthe case of high-alaryemployees, The ‘incidence’ is farmore likely to fall on the corporation, only reducing profits somewhat, therefore investors gains.

  7. Matt, the incidence of a payroll tax is completely different the incidence of the de jure payor of the tax.

    All the evidence suggests it is paid by the employees because the supply of labor is so inelastic and the university in your example largely passes it on in the form of lower wages.

    This neoclassical sort of argument, which can be nicely proved by use of graphs, does not apply in the short run in labor markets due to sticky wages preventing cuts, and opacity and lack of bargaining power by labor, which is generally unaware of how much their employers are paying in taxes on their behalf.

  8. A possible consequence of removing the cap on wages taxable by the payroll tax would be increasing the incentive for employers and employees to characterize the returns from their employment as something other than ordinary income. That already happens in some areas, the most notorious example being hedge fund managers who earn most of their yearly take in the form of dividends and capital gains (which are taxed at a much more favorable rate under our current tax code). Making that shift would be difficult for many employees, such as the Nobel Laureate in Matt’s example. But let’s change the example slightly. Let’s say a big law firm is deciding between hiring a rainmaker at $1 million per year or 5 junior associates at $200,000 per year. The same superstar subsidy that Matt describes would apply under the current system. If the Social Security cap were removed, the rainmaker (and the firm) would have the incentive to find ways to pay the rainmaker in some form other than “wages.” A big law firm would have a tax department that could be tasked with finding out how to effectuate this change, especially because it would not just benefit individual employees but the firm as a whole. There is, of course, a countervailing consideration that dampens the superstar subsidy effect: Someone (the assistant professors, the junior associates) have to do the actual work of the place (teach classes, do document review and depositions) that one superstar can’t cover.

  9. Are there proposals not to remove the maximum but to shift the income range upwards, raising the maximum by $20K, say, and at the same time exempting the first $20K (at least for employees if not employers), so low-wage workers get to keep more of their pay?

    1. Go back and read the history of how FICA came to be during the FDR administration. It is a core tenet that all receiving payroll pay into the tax. Exempting the lower earners while well meaning, would convert FICA to a “welfare” program.

    2. Most proposals I’ve seen do exactly that (sorry, no links this early in the morning); most of them go up to about $170k, which is where the cap would be if it had gone up with inflation/wage increases for the past 30 years or so.

      1. Sorry, none of the most common proposals address your second point (the low earner exemption) for the reason RickG points to.

  10. Don’t have much time this morning but wanted to stop by and see how this thread is going. Looks like dave shultz and RickG know their SS history and understand why it is set up to be a *universal* program for everyone and not a “welfare” program for just some, and why it is important to keep it that way.

    Also, yes, Betsy, the way things are set up is sexist. But I repeat, there is only one SS issue in front of us now that is worthy of our immediate attention, and it is not the math problem that Matthew Kahn lays out, it is the threat of SS being chained to the CPI.

    Here is a 3.5 minute video of Congressman DeFazio explaining that the excuse that SS recipients undergoing a cut in their benefits can merely substitute one good/service for another is completely bogus, and also how chaining SS to CPI will have the additional effect of raising taxes on lower-income people while giving higher-income people bit of a tax break (maybe that helps explain why Republicans like this idea).

    Finally, here is part of the letter I recieved from the ARC the other day. You will note that this is indeed, a CUT, not a “slowing of the rate.”:

    “Additionally, while cuts from the chained CPI start small, they get bigger every year. Remember how compound interest is supposed to make our 401(k)s grow into a secure retirement? The chained CPI is just like that, but in reverse. For a person receiving the average 2012 Social Security Disability Insurance benefit of about $1,100 per month, adopting the chained CPI would mean a benefit cut of about $347 per year (2.6 percent) after 10 years, $720 (5.4 percent) after 20 years, and $1,084 per year (8.13 percent) after 30 years.”

    (Don’t snicker at that “30 years” either. Most people might not live for 30 years after retiring, but this is about DISABILITY benefits, and yes, a disabled person starting to recieve benefits upon reaching adulthood may very well live into old age).

  11. Matthews microeconomics are impeccable. This comment thread just goes to show that microeconomics is a very limited basis for policy formation.

    1. Matthew’s microeconomics assume that human beings are rational utility maximizers with good information about alternatives.

      Every bit of that assumption is false.
      Human beings are not very rational, their actions are not primarily chosen on the basis of utility maximization, and they don’t in general have good information.

  12. This is an important area in employment economics in general, though. In the US, pretty much every part of the safety net that’s routed through employer contributions gets charged on a per-worker basis, with the obviously predictable effect of discouraging the employment of additional workers (at least workers who report enough hours to be part of the system) and encouraging employers to squeeze every last drop of labor out of the workers it employs (especially “exempt” ones). In other parts of the world, much of the safety flows through general revenues (either corporate or government) so that the effects on employment are lessened.

    The last time this kind of structure was seriously examined, iirc, was during the westernization of the former east bloc, when some economists asked questions like “how do we design economic institutions to encourage employment of the workforce we have, rather than devil-take-the-hindmost and then relying on the dole to support those who can’t get a job?” (In accounting terms, after all, you pretty much end up at the same place.) But then all of that pretty much turned into a corporate bidding war where one side had access to cash and the other side didn’t…

  13. But … but … but … everyone *knows* that Matt, as a freshwater economist, must be merely a tool of the plutocracy. So when he makes a tax proposal so wildly progressive that it’s a complete political non-starter, there must be some deeper force at work. Perhaps Paul Krugman has placed him under Vegan mind control.

    Or perhaps Matt is willing to follow his logic where it leads.

    And by the way, there’s no particular reason why a microeconomist working mostly on environmental issues should have any professional familiarity with retirement-income policy.

  14. The table that Warren Terra referred to above has the data on how the cap has changed over time. Initially flat, it increased beyond exponentially from the 1950s through the 1980s, probably in an attempt to catch up to wage increases from 1937 on. Growth was approximately exponential thereafter until the past few years, when it slowed to zero last year.

    I’d still be interested in finding out how an increase in the cap would affect SS funding and viability. Of course, as Peter Appel pointed out, those with higher incomes would doubtless game the system.

    I hadn’t realized the implicit sexism in the SS survivor policy, noted by Kevin Long and Betsy. Of course, SS was initiated when dual-income families were a rarity, and it’s surprising that so few people (myself included) realized it.

    In any event, thanks, Matt, for taking my suggestion and opening up the topic for SameFacts commenters to join in. I always learn a lot from this blog.

  15. If we raised the cap – I mean eliminated trhe cap- we could exempt the first $20K pf income. Exempt the income status, not the person. Everyone would get a raise. Make this about the people, not the individual income unit. An earlier poster also pointed out that centewring on individuals makes having individuals a burden.
    We need more commonality and community in all ways. It will even help with random shootings.

    1. I would hate to be the people responsible for writing the code that exempted the first $20K of income. It’s not uncommon for people who do casual labor and things like it to have multiple W-2 or 1099 statements. (Ages ago, when my mother was apprenticing as an accountant, I think the most she saw was 16.) You’d either need some kind of realtime tracking system or else just take all the money out and rely on a miracle to return it properly at the end of the tax season.

      1. It might not be so bad. You could add a section -a few lines, maybe – to the IRS forms, and it gets integrated into your refund/liability for the year. The SS fund reimburses the general fund for its share of your refund, boom, it’s done.

        The withholding algorithms for regular tax are already kind of messy, and especially so for people who work multiple jobs and/or have big swings in pay from week to week. (Anecdotal opinion from my experience dealing with payroll in a low-pay industry.)

Comments are closed.