11 March 2010

Salary Data Double Header

Future Mrs SB7 passed along another one of those surveys about which universities and which majors have the highest initial and mid-career salaries. What I'd really like to see is a report that merges both of those two data sets, so we could see what the median salary was at different schools holding major constant.

A lot of the variability between schools on these things has got to be because their graduating classes have very different mixes of majors.  Notre Dame tends to do well on these surveys, but ND also has higher proportion of it's students in Engineering, Science, Business than most of it's peers.  Eight of the top ten lucrative majors in this report are in the College of Engineering at Notre Dame, another is in Science, and ND has the top undergrad business program in the country.  Together these make up 58% of the undergraduates, which I was made to believe was abnormally high, especially for a school whose self-identity is still very wrapped up in Liberal Arts.

~ ~ ~ ~ ~

On a similar vein, I've criticized comparisons of government and private-sector salaries before because they don't hold the types of jobs constant. Most menial government jobs are contracted out, which skews the mix of jobs upwards. I'm no supporter of increasing the government payroll, but let's criticize the leviathan with good evidence in hand.

Here's a recent examination which tried to correct for this problem of the labor mix by breaking down compensation by occupation. Of course no two sets of duties are exactly the same even with the same job title, but this gets us much closer to a sound comparison.

Unfortunately I don't like what I see. Holding the mix of occupations constance, base salary is 20% higher for federal employees. Including benefits, federal compensation is 55% higher. Add to that an order of magnitude more job security for federal workers, and you're looking at a real mismatch.

PS Peter Orszag has been making some noises about this study still failing to account for age and education.  As a young person (and, I would like to think, a rational person of any age) I am not sure why I'm supposed to be reassured that the federal government is employing lots of old people and paying them more than their younger private-sector counterparts with comparable duties.  The only reason I can think of is that these government workers have more experience and are therefore more productive, but does "government workers have higher productivity" pass your smell test?

I'll hand off to Bryan Caplan about the matter of education.  He's convinced it explains the gap in salary, but not the gap in total compensation.  He also points out that since federal workers have higher job security, their compensation shouldn't merely be the same as private sector workers, it should be noticeably lower.  This, of course, is widely believed to be the true state of affairs, though evidence contradicts this particular bit of received wisdom.


  1. but does "government workers have higher productivity" pass your smell test?

    Not me - but then I've worked for the FDIC.

    Actually, out of a staff of a dozen government IT workers that I worked with, I'd say 2 were good technicians, 1 was a wizard, 1 a good manager.

    Of the remainder, 1 was well-meaning but clueless, 1 an actual menace and the rest were just bodies, neither bad nor especially good at their jobs.

  2. I'd guess that education plays a large role in the gap in salaries for engineers and scientists. There's a big difference between an engineer with a BS and one with a PhD, but they both fall under the "engineer" heading.

    Given the DoD's intense research operations, the feds as a whole probably employ a more-educated engineering/scientist workforce than the average private sector firm. Many private firms have no need of anyone with more than a BS - on the job training is far more useful to them in many cases.

    It doesn't explain the whole gap, but it is something to keep in mind. On the other hand, there's no way a federal crane tower operator is worth 20% more wages than his private counterpart, much less the massive benefit disparity.

  3. Yeah, education probably plays a roll. I have no idea how much though. I don't have a lot of confidence in my memory on this one, but last I heard the federal government employs about 20% of people with science doctorates.

    I don't have any numbers for this at all, but I think you might be underestimating how much of research the Pentagon farms out to contractors and universities. The NSA on the other hand... I think they're still the world's largest employer of mathematicians.

  4. I know the DoD farms out tons of research to Universities (being a grad student whose advisors have several military contracts). But I'm betting 90% of those universities are state-run (and thus non-private) institutions. I know little of the private research firms, but I'm betting you've got a point there.

    On the flip side, I was trying to be very careful not to underestimate the number of tiny firms who employ only a handful of engineers to design their widgets, software, or whatever (or as contracting firms). They exist, and are often ignored.

    For example, I worked for a private engineering firm for a summer that did nothing but contract work for the rail industry. They employed something like 8-10 engineers, but nobody had a higher degree than a BS except the owner, who had an MBA. I didn't even know companies like that existed until I ended up working for one between undergrad and grad school, and I'm betting there are tons of them out there doing the little stuff for big companies that don't want to bother with hiring the staff to do it themselves. Nobody outside their respective industries know they exist, and so it skews the public view of the job mix.

    Note: I'm not trying to be confrontational; I think we're largely in agreement.

  5. I don't know why I discounted state university faculty as being public employees. I should know better: I'm a grad student at a state school. I'm not on my A game right now.

    But yeah, I think we mostly agree. No worries.