Andy is right that customers have power in this market, if they keep pushing.
Since they are already doing that, but we see little movement, my intuition says that there must be forces pushing the other way.
I suspect they rationally see first mover disadvantage.
If you publish honest, comprehensive career outcomes for your students, then there is a risk that some horror will be found, and used against you.
So college X may find that (say) 6% of its students are unemployed two years after completing the course. That would be a mix of people who recently lost their job, never got one, and decided that this work was not for them.
That 6% may actually be much better than the average, it might even be about the best, but you can see how that might look against a school that stuck to the "99.9 % of our students get placed" line.
As it happens, the UCEM we have at P&D could do this objectively and consistently. But it would take 2-3 months to run, would give a table of the form:
School::Course::Graduation Date::Job Title::Employer::Employer Type
But even with our level resource there's a couple of obvious problems.
1) Persuading people to tell us their pay is non trivial so not in the above table.
2) We can't find all people who don't work in banking. They don't register on our radar, so if someone does an MFE, but goes to work in Starbucks, odds are that we never see him.
3) We don't see everyone, contrary to what some people seem to believe, the P&D database does not include all bankers. That means we could say that X% of grads from Y school work for GS at some point in their career, we couldn't give an actual count, and the number that is based upon may not be statistically sound.
4) There are aggregation problems, in that some universities and courses have multiple names that their alumni call it. This thread itself is an example where a course is called "MSOR", rather than it's full name. New York, London and Paris each have many universities and institutes, and although it's entirely feasible to untangle that with a sed script, it's a big boring effort.
5) Our data can't be trusted. As I've mentioned elsewhere, the founders of P&D are themselves part of the financial education industry, so would you really believe us saying we are really good ?
6) Representing it is an entertaining problem in visualisation.
Personally I think of these things as non deterministic finite state automata, where transitions have probabilities with memory and therefore are Markov chains, ie if you graduate from X, and your first job is at Y, the odds of your next job being in a hedge fund is Z.
But I don't see that as a mass market graphic. I actually did something like that once, and the art director of PC Magazine wrote that it was the worst image ever to have appeared on a computer screen.
