Sentences to ponder

Real earnings for young college grads have fallen by over 15% since 2000, or by about $10,000 in 2011 dollars

Michael Mandel’s tweet is here, and link to the underlying material is here.

Don’t be misled by claims of a “high” or “rising” college premium, that is indeed true relative to high school (or less), but many of those wages are down even more.  In absolute terms the return to college is not doing well.


Leftist choose to earn less.

We know they are leftist because they voted for the most radical leftist socialist in American history who rammed Romneycare down the throats of the American people.

Something happened on the way to Obamacare fixing the economy: Are you trying to claim that Romney is left because Obama is left or that Obama is to the right because Romney is to the right? Both? Whatever works? Maybe they are both just DA's. Whatever it is, a serious person might have spent some time getting some things shovel-ready instead of ramming through their barely/almost Constitutional pet boondoggle. On the other hand, if Obama is right that there is no such thing as shovel-ready, then he has officially put the nail in Keynes' coffin.

Indeed. I wonder if the graph excludes the occupiers.

I would like to see this compared to average hours worked by the respondants over the same time period. I would imagine that the change is even worse than this graph indicates.

I'd like to see this over a longer time frame (20-30 years) and with the trend line of average young high school graduates. Taken in isolation this graph doesn't tell me much. TLI! (too little information)

Almost certainly. Frankly, I don't believe that the mean earnings are as high as depicted on this chart.

Census data is available on and it has the variables to construct that out of. It's easy to work with - you want March CPS data, and you'll have to adjust for inflation. It shouldn't take long at all to pull together.

I thought this was interesting: average cost of compensation per hour worked: (~$20 in wages, ~$8 in benefits)

"In absolute terms the return to college is not doing well."

What does that even mean?

A person deciding whether to attend college has to look at hir two alternatives. The fact, if true, that both college graduates and less-than-college graduates face worse prospects than they did four years ago is not part of this decision [unless the shift in absolute equity drives the range of difference between the two groups into a range where it doesn't matter, which is unlikely at this time].


It's not just the marginal income that matters. It's also the cost of attending college, though lower wages for HS graduates reduce that cost by lowering the opportunity cost of going to college.

'“In absolute terms the return to college is not doing well.”'

'What does that even mean?'

Right. What that graph is really showing is the long-term stagnation/decline in real wages, as national income has been tilted ever more in favor of capital over labor. To call it a decline in "the return to college" is like saying that your bathtub drain is inadequate, because your bathroom is flooded -- while ignoring the fact that a tsunami just washed through your house.

The real story is not the decline in college graduates' wages. It's the decline in wages overall.

There is no decline in compensation overall

looks like "compensation of employees", gross, was down.

For the sake of clarity let us write Wc and Wh as the wage of a college graduate and of an "high school of less" worker. Let us also call Wa the average wage, a linear combination of the Wc and Wh with some positive coefficients summing to 1.

We have three empirical facts here:
1 - Wc is plumbing (this article)
2 - Wa is not declining (your post).
3 - Wc - Wh is increasing (this article as well)

Now, I think it is pretty clear the three facts are not compatible. From 1 and 3 you can say that Wh is declining too (faster then Wc). A linear combination with positive coefficients of declining stuffs is, not surprisingly, also declining. So 2 is not possible.
The three facts are not compatible. Which of the three facts are you prone to question?
At the very least, since you are stating a theoretically impossible result, can you post an empirical result in support of your thesis?

Yawn. Who really cares about the "average" level of anything? That is not the relevant measure. Let's focus on the margin ... What is the marginal increase (or decrease) of one extra year of education or one extra degree or one extra unit of whatever.

If you run a blog and have to go with the chart you have you might care.

Also, I'm thinking about how hard I worked in High School in order to get into college and hit the ground running. If the return to being American goes down, and the total return to being a top-educated American goes down, it may be a minor consideration, but I might go from graduating near the top of high school to just barely getting the degree.

I'm curious what the grad school premium is and what its trend has been during those years.

I have absolutely no data to back this up, but I believe it would look more or less the same. I believe nominal salaries for lawyers and MBAs have been stagnant to declining as well.

i have absolutely no data to back this up, but i believe the opposite

I have no data to back this up but I think it is a wash. Way too many liberal arts/education masters degrees out there to muddy up the gains made by MSEE's and petroleum engineers.

No. The engineering master degree's are a wash and the liberal arts master's shoot themselves in the foot.

I would suspect it would strongly depend on the field. A lot of public jobs, for instance, grant pretty automatic raises for more advanced credentials. So advanced, say, Education degrees probably show a strong correlation for increased pay.

I would not expect this to hold across all fields.

"Don’t be misled by claims of a “high” or “rising” college premium, that is indeed true relative to high school (or less), but many of those wages are down even more."

Tyler, that is what a "premium" means. A premium is relative.

I think we have discovered an example of mood-affiliation to one's own unpopular ideas, or "truthiness" as Stephen Colbert called it.

+1 Struck me as odd to but couldn't express it as well as you did. You also have to look at the starting points--if you are a millionaire, and lose $50k v. someone who earns $50 k and loses $50k.

I am with the crowd here: the income premium rather than the absolute post-college wage is what drives the computation of the present value of net benefit that a person should use to compare against the college tuition and opportunity cost of lost income. Of course changes in the real interest rate should also affect the decision calculus, but NO ONE ever mentions that in these going to college questions.

Wake up people, the wealth effect matters too! No one is denying the marginal return effect, not I certainly.

Wake up? Wealth effect???

You mean that education creates a wealth effect but not an income effect first? The usual way it works is that education creates income, which is saved, and later, as the wealth accumulates, there is a wealth effect from earnings on the saved assets.

The only immediate wealth effect I can see from graduating is your parents or grandparents giving you something in their will if you graduate. Or maybe a car?

Although I am in the top 1%, I will have to ask my college educated kids about this. Maybe their expecting something I didn't give them. Not the first time.

It's fine to consider the return to human capital and yes, for many people their ability to work (now and in the future) is their most valuable asset. (But as you might have noticed this is not a natural frame for non economists.) Before you assign human capital to the "they aren't as wealthy as they thought they were" heap, here's a few things to consider.

First the Census measures may be understating the trend in income. Here's a short summary of the issue by Gary Burtless of Brookings:

Second there's some cyclicality in the income series. Realized returns to education are important, but they don't necessarily define expected returns. Put another way, the weak returns over the recent past may not set the trend going forward. I would argue those expected returns drive your wealth effects from human capital.

OK, now I get it. You have the magical decoder ring and were able to decode Tyler to say that: "They got human capital, which will last their entire life." And, I would agree. But, do you see people arguing for public expenditures to increase human capital--or at least give everyone the chance to increase human capital--on this site?

It is true, though, that if you regard higher education as human capital--that is it is wealth--you arguably are correct: college educated have more options, and have a lower unemployment rate, so their risk adjusted rate of return to education is higher than the lesser education.

But, is that what Tyler meant?

I don't have a decoder ring.

Yes, Bill, wealth of knowledge! Only if I could pay off my student debt with my wealth of knowledge. Let's compare apples, Tyler.

I have a PhD in economics with macro and labor concentrations, so I am comfortable thinking about human capital. One part of my dissertation was whether households integrate their human capital with their financial wealth when choosing how much to hold in stocks. It won't surprise you, but few people seem to think about this like eonomists do.

I do think there is a lot of human capital investment discussion here...all the online education stuff is one example. Oh and I do not have a decoder ring for this blog, In fact, I could still be totally missing the point of this post.

TC and Byomtov are correct. The decision to go to college is not just relative to peers, it includes the cost of college which is going way way way up, by the way and total return which is going way down. If anything, Obama should have addressed that, but I guess he didn't have any shovel-ready ideas in that realm (despite it being ready-made for the low-capital solution of our time).

How much is driven by more mickey mouse colleges and major. Assume you re-weighted the earnings for the makeup of college graduates circa 2000. E.g. if Harvard graduated 800 students in 2000 and 800 in 2012 and University of Phoenix graduated 10,000 in 2000 and 100,000 in 2012, weight a current Phoenix grad by 0.1 and a Harvard by 1. Same by major (e.g. I don't think STEM majors are growing enrollment as fast as business majors).

My guess is that this is where most of the decreasing earnings are coming from.

A mickey mouse college may be better than no college at all, what is perhaps more relevant is whether (a) more people are going to college (b) what they would have done otherwise. So the supposed decline in wages may reflect a population coming into the data that earlier just went to school.

Mickey mouse colleges – yes. Mickey mouse majors, probably not. There have always been history and engilsh majors, and they have been very successful in business and commerce. Only recent grads have had a tough time.

Mickey mouse majors are more sensitive at mickey mouse schools. There's some minimum level of competency for most STEM majors. E.g. physics can only be made so easy.

The difference between an english major at Harvard and an english major at Greendale Community is far wider than the difference between a physics major at the respective institutions.

The difference between an english major at Harvard and an english major at Greendale Community is far wider than the difference between a physics major at the respective institutions.

I don't know about that - sure you can only make physics so easy but on the other hand there's no limit to how hard you can make it and there is a limit to how hard you can make English Literature.

Okay - but can we distinguish the fact that getting an English degree at Glendale Community probably puts you all in like what . $3,500? And it will probably impact the NPV of your future earnings materially. The problem is with students going to Glendale Institute of the For Profit Tech Academy of Arts and shelling out $16,000 a year for worthless degrees at "universities" that should be completely unaccredited and should not be receiving student aid loans were it not for their lobbyists and paid Congressmen.

Before all else, I say THOSE are by far the most worthless degrees, and the majority of the worthless, good for nothing degrees. They are taking advantage of students who don't know better. The hated $150K debt history student is more of a favorite anecdote than something that's a serious problem. Also, not everyone can become a plumber or a welder. I mean, you can, but your salary will max out at $85K if you are good. Its not the ideal career path for everyone.

Students who get English degrees at Public Universities and Private Universities will still do pretty well for themselves. I have plenty of peers in both who ended up in excellent careers with both. English is actually, in my opinion, quite a worthwhile major and I've always looked favorably on them when I'm hiring, primarily because communication is crucial in the service industry.

The mickey mouse colleges aren't graduating people at high rates. There's been a real shift in the proportion of college students who go to less selective schools, but the shift in the proportion of actual graduates coming from those schools is smaller.
The Census data sets wouldn't really let you test this hypothesis. The closest you could get is that ACS has a couple of years of major data, so you could compare college majors among older people w/ those among younger people. It doesn't say when anyone got their degree, though. But IPEDS has extensive data on this. By school, it has the number of degrees in each field granted per year, as well as some information on selectivity. I don't know how far back this goes. But that'd be where I'd start with.

I'd examine how many of the Mickey Mouse colleges are for-profit DeVrys.

The ROI on attending a Big Ten school, I'm guessing, its substantial. And I think, for some of our lesser educated citizens, the ROI of attending the Southern Illinois Universities and South Carolina State Universities are actually very high too, I think the highest when you take into account expenses and IQ. They put a lot of people into positions of finding meaningful employment where they might not have otherwise. Actually wasn't something South Carolina state school ranked the "best" university in one of those yearly rankings on this alone?

Just keep in mind the attending vs. graduating distinction. The graduation rates at those low-tier for-profit schools are very low. So they'd be interesting to look at for all sorts of reasons, but I think it's a stretch that the above results are just due to the changing composition of graduates.

I haven't seen what you're referring to w/ those rankings, but I'd be interested in you have a link. I'm imagining that they had to torture some data to get it to say that. Colleges don't have that much information on what their graduates do, and they certainly don't have the kind of information that would let them make statements about the counterfactual - what their graduates would be doing if they hadn't gone to college.

Looks like good news to me. A useful rule of thumb, I find, is: how does this stack up to Europe? My impression is that the baggage that accompanies hiring someone in many of these countries has priced many young people out of the labor market.

For reference, I was very happy to get a job for $22K out of college in 1987 ($43K in 2011 $s.)

"Mean earnings of full time workers 25-34 with only a Bachelor's" is not right out of college. Would you be very happy with 43k 9 years out of college?

+1 for catching that. I was thinking, wow $55K a year out of college, isn't that base pay at second tier consulting firm? Not bad. 9 years out of college 55K should be the value of either your bonus or stock options.

If I was average or didn't work in a STEM field, yeah.

If I was my brother the English major, I'd be tickled.

If I was a Chicago public school teacher (starting salary: $50K+) I'd be on strike. You know, for the children.

Also, where do you get 9 years from? Presumably, 31 year olds would be doing better than the average for 25-34.

"Presumably" a 30 year old would be doing about average among 25-34 year olds, assuming that compensation increases in a roughly linear fashion across the 25-34 age bracket.

Yes and I'm sure your tuition costs in 1987 were basically the same as they are today right? Just a slight adjustment for inflation?

If you're arguing that a fair number of people with crummy majors might be better off giving college a pass and learning a useful skill instead, well, Amen brutha.

A fair number of people with crummy majors and large loan debts would be better off if they had just stayed stoned in their parents basement for four years and then gotten a job as a clerk at minimum wage.

Funny, but sadly this is quite accurate.

Probably, but is there any evidence to indicate that this is happening at a material level? People like to pick on the English major who went to some Liberal Arts school and racked up $150K in debt . . but that's mostly fantasy and does not happen at material level

on the other hand, 25% of student debt being from students of for profit universities, that happens. Students of for-profit universities comprising 50% of defaulted student debt, this happens too.

The median debt being half the average debt. Only 10% of students at private universities take out more than $50K in debt.

But still we want to focus on the stereotyped hipster at Swarthfrankerson College because they are easy to attack than the for profit schools.

"People like to pick on the English major who went to some Liberal Arts school and racked up $150K in debt . . but that’s mostly fantasy and does not happen at material level"

Just saying something doesn't make it true. And where did the $150K figure come from? Average college debt in this country is around $25K. But graduating from a college with $25K in debt with a degree that's of marginal worth is not a good thing.

I was very happy to get $22k 2 years out of college in 2011.

how does it make sense to talk about "absolute returns" to college? I think what you mean is the absolute returns to being an American not being born into a wealthy privileged family are not doing well. That I can agree with.

via twitter:

Similar to Michael Mandel

tylercowen ‏ @tylercowen Follow Unfollow Cancel
J. Bradford DeLong ‏ @delong Follow Unfollow Cancel

just thought you'd be amused...

Just curious - the picture looks very similar, but the numbers I obtain when I go to the Current Population Survey and BLS sites show:
year mean income cpi real income
2002 39467 179.9 49347.79051
2003 40767 184 49837.43594
2004 40248 188.9 47926.65364
2005 41828 195.3 48175.87554
2006 44268 201.6 49392.85542
2007 44454 207.342 48226.78621
2008 45215 215.303 47238.62132
2009 43115 214.537 45205.46565
2010 42871 218.056 44224.23538
2011 43455 224.939 43455
This is for age 25-34 with BA only - all workers, mean income. Don't understand why my numbers are around $10,000 different from those posted. Mine shows a decrease of around 10% rather than 15%, but with the same profile.

Sorry, the table did not post the way it was formatted in the comment box. The series of numbers shows four numbers for each year: year, mean income (nominal), consumer price index, real income ($2011).

The chart above is just full-time workers.

Thanks, that makes sense. I guess since the real wages of full-time workers dropped more rapidly than the wages overall suggests some shift towards part-time work. It might also say something about the lack of wage stickiness among full-time work seekers.

I always thought the "college premium" was misleading. Certain majors earn much more than others, and certain majors (particularly engineering) tend to have higher earnings out of college before peaking earlier.

I would like to make a point that is a little off topic here. That is that it seems to me that at least our publicly funded universities should make it easier and more enjoyable to get a STEM degree and harder and less enjoyable to get a non-stem degree. It seems our schools are producing too many non non-STEM grads, so why subsidize the production of more?

I also think that those with the ability to do well in STEM degree programs are not always better than those who are not good enough at school to get a degree. For example in my field CS to good at school one must be able to understand complex algorithms, but accuracy at mundane programming tasks is less important. In the field one seldom needs to understand complex algorithms and one has time to learn them but accuracy at mundane programming tasks is often needed.

PS I would add degrees like accounting and nursing that are still in demand to the STEM category.

PS I would add degrees like accounting and nursing that are still in demand to the STEM category.

Well if you're adding nursing and accounting you can throw out S&M, take it from me the employment prospects of the S&M part of that acronym are about on par with your run of the mill sociology or psych major. But even Engineering isn't a sure thing it nowadays - seems like there's a lot of unemployed mechanical engineers out there so even in these "good" fields you have to be careful.

A few more years of this economy and no one will bother with the effort to get a STEM degree.

That's a sub-example of the point of this post. If college cost keeps going way up (maybe because you now have to struggle just to enter the employed class) while the total returns keep going down, your opportunity costs of college might start looking really good. Being a farmer keeps looking better and better to me these days.

If things keep heading this way and students abstain from attending university or getting STEM degrees, won't companies who benefit from the flow of educated worked intervene to restock their future talent pools? Or is not every company as far-sighted as Intel (one of the best at providing scholarships and the science talent search to foster future excellence)?

I can see corporate-funded scholarships to study certain fields and to buttress the rising cost of undergraduate education.


I don't think you want college administrators to put their thumbs on the scale to move people in a particular direction. But, you do raise a good point: perhaps the English and humanities professors have put their thumbs on the scale by giving out easy grades so as to attract more students to their class.

Perhaps we could ask them to be more boring, or grade on the curve.

I suspect that it is a fallacy that we today let/help people follow their dreams. On the other hand, you can make almost any major more quantitative and teach transferable skills in the context of the specific major. What they actually do is spend their time trying to differentiate themselves to create the need for more departments.

Thank you! This drives me crazy. The comparison between outcomes for college grads and high school grads is important, but so is the comparision between college grads of the past and those of today. Starting salaries are falling while debt loads are rising.

This isn't a reason not to go to college. But its a reason for people to start talking more about ways to force college tuition down. Everybody should still go to college, if they can, but they shouldn't be paying so much for it.

I get what you're saying Tyler, but I was making about 20K a year for a few years after high school, got laid off, and was homeless after my best friend from high school roommate got addicted to meth and had us kicked out of our apartment. Three years out of grad school (right now), I'm earning four times that much, own a home with a mortgage less than 10% of my income, and my roommate (wife) earns more than me and isn't a drug addict. The lifestyles and cultures alone are pretty damn radically different.

Yes, being addicted to drugs and not having skills and not having gainful employment is bad. College costs growing at double inflation while college graduate salaries fall 10-15% a decade is not somehow good if we compare it to something worse.

I believe that there are several forces at work here. I have no data to support this.

Companies have been working for years to reduce costs. Labor costs, training costs, benefit costs. They are using fewer people to do more work. They are offshoring work to reduce costs and to shift work. They are hiring more educated workers, assuming that they will require less training. They are hiring more part time and contract workers to reduce cost. All of this results in less full time openings and more people to fill each position. By supply and demand, wages go down. Its not a mystery.

I know fellow law graduates with over $200k in student loan debt. We have sold ourselves into indentured servitude chasing the Premium.

I laugh every time I see these numbers. I went to a good private liberal arts college. Graduated in 1995. BS with majors in physics and English lit. Had no idea what I wanted to do when I got out, just that physics grad school wasn't going to happen. Went to Japan for a couple of years mostly because I couldn't figure out how to start a career. Came back and started as a temp at my current firm in 1998 as a temp. Full time hire salary in 1998 at about $25k with nice benefits. I have no complaints now, but my returns came from learning the business I entered in spite of a low starting salary. At a Bachelor level, I know of about zero people who earned anything like the starting rates suggested. That's not to say I don't believe them, but I have the suspicion they are skewed by people directly entering finance or law or something.

FWIW, in 1999 my starting salary was $65k and my total cash compensation (including bonuses of various sorts) was $80k. But my major was engineering and that was a fairly good offer, though not anything out of the ordinary.

And my numbers are not adjusted for inflation, as the chart is.

My wife left college two years ago, got her CPA, and got a $45k starting salary from a company that pays more on the benefits side of the scale, and not anything like one of the accounting Big 4. She'll pass that average salary inside of two years employed.

OTOH, this was a job that really requires a college degree, not one where they'd simply only hire someone who had one.

Also, that chart is mean earnings, not median, which make it easier to believe.

"Don’t be misled by claims of a “high” or “rising” college premium, that is indeed true relative to high school (or less), but many of those wages are down even more. In absolute terms the return to college is not doing well."

In the absence of time machines, this seems irrelevant when discussing the cost/value/ROI of college education.

ryan, if the absolute return to college falls but the relative return holds steady (or even rises), there could be still important behavioral responses. This decline may not change who/how many go to college, but it could reduce the consumption/standard of living of college graduates. Just another cut of the TGS hypothesis. The absolute return to a college education certainly impacts the absolute benefit of a college education.

My God, it looks like around 2004 there was a bit of a boom that peaked around 2005-6 but then instead of going down and up again like the previous ups and downs it just kept going down in 2008. If I didn't know better I'd suspect maybe there was a recession or something that happened around then! I say we form a committee of our best minds to investigate!

Indeed. But if you structure your costs so that you require boom times to make it, you will have serious problems. Or better put, having a college degree doesn't make you recession proof. Fine all else being equal, but quite tough if you come out with substantial debt into a slow job market.

The question is lifetime income and the graph doesn't say that income is $0 during the valleys, just lower than the peaks. Clearly if a huge portion of workers have a college degree, then income from a college degree will vary with recessions/booms.

Maybe there are more and more jobs for college graduates in the lower-wage South / sun belt.

would love for this data to be segregated even more.

Does this data include kids who go to college but drop out?
How does this breakdown by major? By school?
Do wage premiums for college follow a power law distriution by school?

One other factor that'd be hard to include is the number of students who went to grad school as a way of postponing their job market woes.

Young college grads? Someone working full time aged 25-34?

These are "young professionals", not young college grads. A 34 year old will be out of college perhaps 13 years.

"fallen by over 15% since 2000"

Almost as depressing as the fact itself was realizing that an eminent economist and Harvard Ph.d. doesn't know the difference between over and more than.

There's a selection bias, in that more people get degrees now (and advanced degrees especially). So the pool of people with bachelor degree only has changed over that time. It now includes people that earlier would not have had a degree and has lost more people that now are getting advanced degrees.

Comments for this post are closed