Bryan Caplan  

The Federal Pay Scale: Inflated Yet Compressed

Practical Monetary Policy and ... Contra Garett Jones: Interest ...
Back in 2010, USA Today and other popular media loudly declared that federal employees were grossly overpaid.  Famed economist Peter Orszag swiftly denied this allegation on behalf of the Obama administration.  I found his response extremely suspicious, but was too lazy to confirm my doubts.  Recently, though, I discovered that the Congressional Budget Office did all the necessary legwork for me in early 2012.

Here's what the CBO found:

1. After adjusting for education, occupation, work experience, and other observable characteristics, federal salaries are only 2% higher than in the private sector.

2. HOWEVER, federal workers' fringe benefits are 46% higher than in the private sector.  As a result, total compensation (salary + benefits) is 16% higher for federal workers than comparable private sector workers.

3. Overcompensation is highest for the least-educated federal workers - +36% if you've got a high school diploma or less.  As education rises, the federal worker premium falls.  Federal workers with professional or doctoral degrees actually earn 18% less than private sector counterparts.  Full results:


Prettier version:

Two observations:

1. People occasionally claim that useless credentials only pay because the government hires on the basis of useless credentials.  If you take a careful look at the table, however, both the education premium and the marginal dollar payoff of education are higher in the private sector!  Yes, the federal government pays more for B.A.s than the private sector does.  The reason, though, is not that the federal government overpays for credentials.  The reason, rather, is that the federal government overpays for breathing.  If someone who never set foot on a college campus manages to land a federal job, his marginal reward for acquiring further credentials is exceptionally low.

2. Despite many virtues, the CBO report still ignores unmonetized fringe benefits, including the Big Kahuna: Job security.  During the last five years, how much would the typical private sector worker have paid to enjoy the same level of job security that federal workers take for granted?  We won't have an apples-to-apple comparison until we estimate that dollar figure and add it to the final tally.

As far as I can tell, the White House never responded to the CBO's report.  Am I wrong?

P.S. Also check out Bewerunge and Rosen's working paper on this topic.

Comments and Sharing

COMMENTS (25 to date)
BLM4L writes:

I bet that professional degrees in the Federal Government get paid much more if measured on a per hour worked basis rather than a salary basis.

Example: New York white collar defense associates receive probably around ~ 200K/year in total comp. but are expected to bill ~ 2200 hrs/year (i.e. actually be on the job at least 3000 hours). But SEC lawyers are expected to be on the job 2000 hrs or so a year but need not bill anyone for time at all (i.e. they can slack off).

Tom West writes:

Is the idea that Federal workers are overpaid at the lower echelons a product of the fact that wages for less-skilled workers have plummeted, leaving a government job as one of the few places where the uncredentialed can make a (lower) middle-class living?

It's not hard to believe that in the absence of government jobs, the non-college-educated, non-trades middle class might be near non-existent.

Data junkie writes:

This result is driven entirely by which benefits one counts and the counts in the studies is always biased to make it look like the federal government pays more than it does.

Studies are limited by what they can measure and for benefits that means tax-advantaged benefits, mainly health care and pensions. What they don't count are benefits such as per-diems for travel, meals paid for by bosses, office parties, car service, corporate memberships, gym memberships, etc. These don't exist in the federal government but are prevalent in many professional firms. Consider that federal employees have to bring their own coffee machines to the office and make their own coffee. Pay studies should always be looked at skeptically because the results depend crucially on what measure is used. This is one reason to not include "benefits" as they are measured with so much noise and in such a biased way.

Doug writes:

"Studies are limited by what they can measure and for benefits that means tax-advantaged benefits, mainly health care and pensions. What they don't count are benefits such as per-diems for travel, meals paid for by bosses, office parties, car service, corporate memberships, gym memberships, etc... Consider that federal employees have to bring their own coffee machines to the office and make their own coffee. "

Oh well yes of course. Everyone knows that $25 coffee machines and an annual christmas party are far more costly than lifetime healthcare and a 90% pension after 30 years. It's not like healthcare costs rise far faster than inflation, long term interest rates are at zero, or that life expectancy for a 20 year old is double the 30 year old pension time.

This is of course why you hear of large corporations around the country falling into bankruptcy because of their unfunded and unsustainable employee gym memberships and travel per diems:

Peter writes:

Other points to remember are:

1: Education is a bad measurement though I understand it's easy. You would be better to compare GS levels and job series to their private sector counterpart. For example my regional director who directs four thousand employee's across 13 sites with major national public security responsibilities in six nations with a $50 million budget is only a GS-15, i.e. makes a whopping $129K a year. I highly doubt his private sector counterpart is making anywhere near that. Basically the rule of thumb (though of course anecdotal) is GS-10 on down make off like bandits, GS-11 and 12 parity, GS13+ screwed. Also education has nothing to do with it outside the initial hire level, once you are in you can progress based on one year past experience at the next lower grade.

2: You are absolutely correct on you have no incentive to further your education once hired BUT you seem to frame it as a bad thing instead of realizing most jobs, private and public, are over credentialed. I have seen Federal job advertisements in the science field which require a PhD to watch a screen all day and if an alert sounds once every two or three years, push a red button .. seriously you DON'T NEED A PHD FOR THAT! Also, speaking from personal experience, I would that a high school graduate working in the Federal system for twenty years now who has absolutely zero incentive to get a degree (I'm a GS-13, my promotion potential is GS-14 .. sure I could get a 15 or SES if I got a masters by the money isn't worth it the cost). There are plenty of folk in my shoes.

3: A big one folk forget is plenty of government jobs are make work programs for the disadvantaged. We have requirements to hire military spouses, disabled vets, American Indians, people pushing mental retardation, etc under the theory (my guess) that giving these folk meaningful work (in their own head) is better than them sitting on welfare and probably cheaper given they pay into social security, Medicare, and they are ZMP. For example, I know we have a mentally disabled gentleman who has been with us for twenty-five years, he’s been in public housing the entire time, is a GS-2 (i.e. works for basically free), and his whole job is to drop envelopes in the mail (for an office of ten people who do next to no business via mail). Sure he only does productive work maybe ten minutes a week but that’s ten minutes I don’t have to waste and the public probably wins out on GS-2 (including benefits and retirement) v. a lifetime of welfare and disability.

4; I know everybody likes to pick on the Fed’s but try comparing private sector to state / municipal/ tenured university employee’s. I’m pretty sure the Fed’s come off the worse even counting benefits in that comparison.

So yes the low level guys make out but that it intentional and the high levels guy get screwed on pay but the government compensates by making them nearly unfirable.

Andy writes:

I'm a federal government employee and I think in this economy we are "overpaid" relative to the private workforce but I think that's mainly because the private sector is in such bad shape, especially with regard to employment. Things were quite a bit different in the late 1990's and then again after the 2002 or so. Since 2001, this happened in defense and intelligence (which is my area of expertise). I know many people who left public service for what appeared to be lucrative contracting jobs. The shoe is now on the other food though since contractors are getting cut first and many are desperately looking to get back into the public sector. It's going to get a lot worse for them after 2014. Did they come out ahead? Hard to say as individual circumstances are vastly different.

Another issue is the federal pay scale which is relatively fixed. There is rarely premium pay for talent, so when opportunities arise, many of the most talented will jump to the private sector since compensation is much more fixed in government.

PrometheeFeu writes:

One should ask if some of that salary difference could be explained by what kind of people are hired by government jobs. For instance, let's say you were to take high technology jobs and control for education, you might end up with a higher salary than average because engineers make more than English majors. So if the government is biasing its hiring towards people with highly-valued degrees and skills, that could account for the salary difference.

MG writes:

I have found Andrew Biggs (and partners) research on the issue of Federal Employee Compensation and Teachers Compensation comprehensive, rigorous and convincing. Here are links to (as I recall) the most recent reports:

As I recall, the teachers' study (if not both studies) even included a measure of "intelligence/aptitude" as a control/explanatory variable. (Needless to say, this proved to be highly controversial in the way that "truth hurts" kind of way.) They also addressed pretty much all the issues mentioned in comments, from the value of guaranteed pensions, retirement medical insurance, vacations days, and even job security.

I have seen these papers and other related studies presented at forums with "committed" critics, and I think Biggs deal with the critiques very convincingly.

Jeff writes:
Studies are limited by what they can measure and for benefits that means tax-advantaged benefits, mainly health care and pensions. What they don't count are benefits such as per-diems for travel, meals paid for by bosses, office parties, car service, corporate memberships, gym memberships, etc.

The company I work for does a fair amount of work for the federal government, and we get more generous per diems when we travel on federal gov't business than we do when we travel on internal company business, (ie, for training and department meetings). I can only assume those same per diems are available to actual employees of the federal government. I doubt they would pay for a contractor's meals but not their own employee's when they're on the road.

Beyond that, how many private employers actually offer fringe benefits like the one's you mentioned? Not that many. And the value is pretty small. How much is a gym membership worth? a few hundred bucks a year, most places.

The one form of compensation that (some) private sector workers do have that is hard to measure is stock options and warrants. These are basically out of the money call options which can have a very large impact on compensation if a small company really takes off. Or they might turn out to be worthless. But financial service firms manage to put a value on stock options using certain financial models, so difficult to value does not mean impossible.

"People occasionally claim that useless credentials only pay because the government hires on the basis of useless credentials."

I am partial to this claim. I'd like to see the data on the earnings of degree-holders in the private sector and in government broken down by major. It is no surprise whatsoever that a something-hyphen-studies PhD hired by the government for diversity consulting earns less than a physics PhD hired as a quant on Wall Street. What is a surprise is that the something-hyphen-studies PhD was hired by anyone at all.

Mike W writes:

According to the CBO study only 20% of the 2.3 million (non-military) federal employees are in the high school or less category; compared to 41% in the private sector. But, “In addition, about 800,000 people work for government enterprises that typically pay for their employees’ compensation through the sale of services rather than through tax revenue. (By far the largest government enterprise in terms of employment is the Postal Service.)”

So the significant labor cost (528,000 career employees and 101,000 non-career employees) of the Postal Service, most of which is in the category of education of high school or less, is not included in the CBO analysis. This cost is moved “off the books” by creating the fiction* of “government enterprises that typically pay for their employees’ compensation through the sale of services rather than through tax revenue”.

*Fiction because the USPS lost $16 billion last fiscal year, most of which (about $15 billion) was accruals for future liabilities for employee benefits. About $1 billion was a current year cash loss from operations. And how was that cash deficit covered…by selling debt or equity to the public? No, it was covered by borrowing from the US Treasury. So how are these folks not actually federal government employees?

Mike W writes:

@ Peter: So yes the low level guys make out but that it intentional and the high levels guy get screwed on pay but the government compensates by making them nearly unfirable.

So then how are they "screwed on pay" if their job security is an element of their compensation? I often hear this argument from upper-educated public employees (including those in education) that they not being paid what they are *worth* compared to similarly educated private sector workers. So why do they stay in the public sector?

Jeff P writes:

Also Federal Civil Servant here.

Mike W - He was refering to "screwed on pay" meaning that a private sector firm is going to pay a manager with similar responsbilities considerably more than the government would pay. Heck, as a GS the most you can make is 155,000 per year, until the next time we get any kind of pay increase. Now, if you are an SES (Senior Executive Service - think civilian General/Admiral - I work in DoD) the money goes up, but still no higher than $200K for certain. Show me a private 20,000+ employee company deputy CEO who makes less than $200,000 a year.

As far as 90% pensions go, that is completely not true. In 1987 we transitioned from CSRS to FERS. In other words, if you started working for the govt before 1987 you could get a 50% pension for working 30 years. Under FERS, you can get a 30% pension for working 30 years and you get 5% matching in a 401K-like account. Normal government health care options expire once the employee is eligible for medicare. You can pay for subsidized additional coverage, but that is pretty costly and doesn't cover much.

Last, our per diem for travel and mileage reimbursement are managed by GSA. They are the source of that and it is standard across the government.

Anyone seeking more information will find all that they need and more on or . Everything you ever needed to know about your government civilian compensation is available for your perusal. Also, at you can also find military salaries. No secrets here, unlike the private sector, everyhting is open source.

johnson85 writes:

When you look at the higher pay scale, you also have to take into account the benefits they are getting from the ability to go through the revolving door. They may look underpaid at the moment, but if two years in their current job gets them a much more lucrative job lobbying or otherwise greasing the wheels for somebody in the private sphere, that's pretty valuable deferred compensation.

Tom West writes:

So why do they stay in the public sector?

Following that line of reasoning, it's tautologically impossible to *ever* be 'screwed on pay', in either the public or private sector.

Emily writes:

@Garrett M. Petersen: ACS data now does include majors. I don't think it's big enough to validly compare wages across majors in the private sector vs. the federal government (unless you aggregate majors a lot), but you could test the theory that college graduates working for the federal government tend to come from softer/less quantitative majors quite easily.

@Tom: If federal government employees left their jobs voluntarily at higher rates than private sector employees, this would be a point of evidence in favor of underpay, because we would see that they were not sticking around. In fact, it's the opposite.

Mike Rulle writes:

The amount public or private sector workers get paid in absolute dollars should not determine if public sector gets overpaid. We should also compare productivity.

Oops! I forgot. We have no method to measure government productivity. Can government productivity even be defined? Probably. But I have not seen an attempt.

A simple method could be the size of deficits as a first order of magnitude----minus all those multipliers those deficits produce.

Multipliers!--Oops again---another immeasurable.

Peter writes:

@Mike: Because perceived job security is far, not near, and benefits are as fickle as politics. Whereas sure a company can fire you,they can't retroactively take your pay for the last twenty years where as the Government can, and does, retroactively kill benefits nor do we have a union contract to stop them (senior level government employee's are union exempt for the most part). For example (and this is just a made up example) if I join the Federal service and accept less pay now because of future benefits of job security, a defined pension plan based on high three, and 1% retirement tax and fifteen years now the road it's now low five, 25% tax, and performance (i.e private) based employment they are basically retroactively renigging on the deal. In the private sector you take those benefits up front in increased pay and should they fire you / change the deal fifteen years in you still pocketed those increased benefits.

As for why, in many cases (especially at higher levels) the jobs simply don't exist in the private sector. A prime example are the two tsunami warning centers. If you are a PhD geophysicist and want to be involved in real time tsunami research, the government is your only real option. Nor is there really a private sector market for guys who sit on beaches all day and count turtles. What many people forget is large numbers of government jobs simply aren't transferable. Now you could argue why do we have them but the traditional answer is they exist because of a publicly recognized need and market failure, i.e. we want people to provide emergency warnings about black swan events (Tsunami's) or incoming ICMB's.

Also I explicitly stated "pay" and not benefits in my initial comment. Regardless of whether or not folk make more overall, people are people and they don't think rationally. They perceive themselves with crappy benefits when they look at their neighbors and know in dollar terms they make less hence it bothers them. For example my wife can't understand why I make less (in absolute dollar terms) than the Gucci store manager in the condo next door though my work has exponentially more responsibility. You can argue all day about overall benefits but what she sees is "He has a Mercedes, we have a Volkswagen".

@Tom: Not really because of a couple reasons. One many jobs simply aren't transferable. For fields with a high rate of cross over (IT, HR, etc) you do see people leaving their jobs for the private sector at higher rates, especially those under 40 and/or under ten years of Federal service. Many folk in these fields bounce in and out of the public/private sectors routinely else you lose your competitive edge. The other big one is sunk cost (or the fallacy of if you wish). By the time you reach the higher levels where you are behind your private sector peers (GS13+) you often have fifteen / twenty years of service and at that point it's ridiculous throwing away real benefits you will earn in five to ten years for a 10% increase in pay. Lastly, believe it or not, some people simply like working for the government and see it as a duty / like their employer in the same reason some folk like working for (just an example) Google even though Google treats them worse than a competitor might.

@johnson: That is a media misperception. While I won't deny you see it at the SES/flag/politico level the vast majority of the civil service isn't in that camp. Most of them are retiring around 65 after 30 or 40 years of service and retiring just like everybody else. The one exception are military members who often retire at 38 who then revolve back either as a contractor or a civil service (to double dip) but that is a fault of the military retirement system which is a whole 'nother conversation and one nobody seems to want to have at the political level.

Lastly I'm not trying to defend the Federal system here but sitting as a GS-13 in highly technical niche field with high private sector crossover I find it extremely challenging to get quality employee's and the primary reason is I simply can't compete with the private sector. A twenty-eight year old technical SME isn't going to take a GS-12 Step 1 position with the promise that forty years from now, if the government doesn't change any laws (which we know they will), you will get an OK retirement and not have to live on catfood. So instead I hire guys who can't get a job in the private sector and have to deal with them as something is better than nothing and equally have to hear about it from the public "why do we get such crappy service AND overpay them for it". You get crappy service because I have to hire crappy people and they get a premium because the law define what I have to pay him.

PS: Sorry on lack of conciseness and/or ramble. Bit of a rush and I'm not a huge fan of typing on a smartphone. Also no hard data, just a bit of me opining :)

David Youngberg writes:

As BLM4L wrote, and don't forget the hours are much better in the government. Not only can you slack off during the normal working day, you're never expected to come in on weekends or stay past five.

Seth Weissman writes:

Disclaimer: I'm a federal worker.

Keep the historical context in mind. The GS scale was created in 1949, when the level of income inequality was lower than it is today. And since then, the GS scale has increased pay at each level and step proportionately, preserving the inequality level of the post-ww2 period, in contrast with hat has been happening elsewhere in the economy.

So one way to read these pay disparity at low and high income levels between the GS system and the private sector is as a baseline against which the rise in inequality in the private sector can be compared.

Tom West writes:

I remember when the right-leaning provincial government of Ontario decided to publish the salaries of every civil servant earning over $100K.

It initially generated the desired newspaper stories about those over-paid bureaucrats. Less expected, it also generated a feeding frenzy among HR firms who realized that high-level government managers were earning a *fraction* of what they would in the private sector.

A while (and a whole lot of personal losses of the government's best managers) later, and salaries for the top end went way, way up...

MingoV writes:

I directed a lab at a VA Medical Center. From my viewpoint, the claim that comparable workers in the federal government receive only 2% more salary is nonsense. The study used education and years of employment as the main criteria for comparing employees. The comparison omitted the most important factor: worker productivity.

Productivity of VA medical center employees is much lower than similar workers at private hospitals and clinics. The VA medical center had approximately 1.5 times as many administrative employees as same-sized private hospitals. The VA medical center nurses had the highest salaries in the region, and the lowest productivity. When I came to that VA medical center, the lab techs (also the highest paid in the region) spent 2-4 hours per shift reading magazines or doing crossword puzzles because of over-staffing. Excessive staffing was common throughout the medical center.

My estimate is that correcting salaries for productivity will show that federal government employees receive at least 25% more than private sector employees.

Glen Smith writes:

Based on my experience and a small number of my closest friends. Low-level workers in government jobs were had workers (often unproductive but hard workers). Low-level workers in private industry almost always did the minimums. My guess is that the low-level government people I worked with were working for more than just the pay while the low-level workers I worked with were just working for just the pay.

Dan writes:

Disclaimer, I am a Fed.

The cost of living in the DC area is very much higher than in most places in the country. A proper look at Fed salaries should be in relation to people who work in the same expensive areas.

An older house in a neighborhood with decent schools starts at half a mil around here. Therefore the average Fed with a college degree starting a family can't even get that.

My agency did not have the pick of the litter in most years because the local job market is pricey.

Bryan Caplan should know all this because he lives in Northern Virginia. Why did he ignore the obvious?

Eric H writes:

Hard to be surprised at the number of federal employees defending this on statistically improbable grounds that were already acknowledged in the OP. What percentage of civil servants are SES? Managers? Living in DC?

After working for a series of Army contractors for many years, I observed many people jumping from our ranks to the government, but only one who went the other direction (not counting retiree double-dippers). He eventually went back. Our technicians (all with at least AS degrees) were routinely held to a higher standard that even the government technicians (mostly HS diplomas) acknowledged. They could break stuff and laugh about it, we couldn't. I got chewed out one time for turning on the lights in a building I had just entered because every single CS technician in the facility was -- I swear I am not making this up -- sleeping on his workbench at about 1:30 on a Friday afternoon. One technician had not been to his workstation more than twice in about 7 years; they finally determined that he had been hanging out at the golf course (!) and allowed him to retire. I know a government supervisor who got jumped by one of his technicians in a Hooters bathroom on a day off with no consequences. I could go on like this for hours. Job security (being unfireable) is a very valuable thing.

As for the difference in responsibility, I would say that in many but not all cases, the government manager had a contract counterpart. The government manager had paper authority, but it was frequently though not always the contract manager who actually ran things and provided the government manager with the information to report upwards. In part, this had less to do with incompetence than it did with the promotion process which didn't start until after the former manager retired, and also favored a number of intangible factors like seniority (regardless of *relevant* experience), whereas the contracts had continuity in senior positions. I knew of one R&D group where the contract engineers would write RFPs for their gov't engineer, who would then send them back through proper channels for them to respond to; they also wrote the performance reviews for him. So while on paper it may look like these poor, undercompensated government managers are overworked, I wonder about how many of them are just rubber stamp org chart occupants.

Every year, we read about the COLA that they got, while in some years ours were 0 (and the broader economy was seeing layoffs) - that is the main reason for the differential. For those that went into the federal equivalent of 401(k), the matching percentages were much higher. The time (1995?) there was a shutdown while I was there, they were eventually compensated and we weren't, so it was just a paid vacation (worth remembering every time there is talk of how the poor federal employees may get sent home without pay over budget issues). There was time off given for Nixon's funeral (yes, really) - everyone else had to work as if it was just a normal day. The government workers either got the time off without having to use their PTO, or they had to work it and they were later compensated with additional PTO "to make it fair". There are dozens of these perqs, and it has nothing to do with preserving the egalitarianism of the 40s or additional responsibility. There are no measures of productivity, they get to make the rules, and nobody is watching the watchers to the level of detail needed to prevent abuse.

Comments for this entry have been closed
Return to top