How many computer programmers does society actually need?

[Foucaulf]

I'm mostly going to discuss the "talking point" mentioned in the OP.

Any professional wage can be looked at both in terms of marginal productivity for an hour of labour and in terms of supply relative to demand.

Marginal productivity: If more output can be created from an hour of programming than using other specialized labour, then the socially optimal outcome is for people to retrain themselves and have a more productive economy in the process.

Automation is obviously what I'm referring to here, but the premium for programmers is already appearing now, before the robots all take over. Complementarities already exist between programming labour and capital investment. Any firm facing a shortage in labour are going to substitute towards capital, which is a demand shock for programmers in the short run and could increase programming productivity further in the long run.

Supply relative to demand: Obviously, the share of programmers in the U.S. remains small. More realistically it takes time to train people to know programming, so this would cause a lag between the wage of programmers observed now and the number of people actually training for it. In reality there are also good programmers and bad programmers, so the two are separated by signalling mechanisms, i.e. college, coding bootcamps and accreditation.

Already you should start thinking about splitting potential labour supply of programmers into two groups: those who have no training and need training to reach competency, and those who have the ability to attain competency, but are really investing in their signal so they get the best programming job.

When the OP's talking points get raised, it's almost always considered by the latter category - college majors or adult professionals who, to keep up with their income expectations, can't just learn how to code but need to do it well. But, even if this ends up being most of the potential programmer supply, it's still a small chunk of the US labor force.

The irony though is that plenty of people think they're really aiming the talking point at the former category, those who need training to reach competency.

Let's get real - you can talk about "accessibility" of programming all you want, but for someone who couldn't get past Algebra II more than a decade ago, on the margin programming training is not a good choice. And it's a scar on the U.S. that there are plenty of people like who I described there.



If I were actually trying to give good job advice to people in the former category, I would say very little which they or the market doesn't already know - the fastest growing industry in the U.S. if not the developed world is nursing.

Personal Care Assistants alone account for more employees than all programmers and software developers in the US combined, according to the BLS. This one group excludes all the other nurses and caretakers employed in hospitals, jails, clinics, etc.

[Foucaulf]

I am glad this thread was reopened, since I was confused why my post was interpreted as "not thinking outside the confines of neoliberalism," but let me state my case more simply:

1) As an economist I don't really care what SV/coastal talking heads say either. That's persuasion, not a market mechanism.

2) If you look at the actual market for higher education, you would see that people looking to be retrained from the bottom up don't listen to any of the persuasion. Here are two figures from the National Center for Education Statistics:

1: Bar graph of associates' degrees conferred by field of study.


2: Bar graph of graduate degrees conferred by field of study.


What this indicates, if anything, is that more people want to get a Masters in Education than all the aspiring engineers and programmers combined, despite the attack on teachers' unions and the average-below average hourly wage including overtime.

To be fair, here is also a graph of four-year undergraduate degrees conferred by field of study:



Computer-related majors are off this graph but are still at 59k or so in the most recent year, so I think MEd's still trump all the computer-related degrees.

I also repeat my claim in the previous post that everybody has learned through market signals that nursing is the highest-growing industry, and are training appropriately.

3) The point here is that the market for higher education adjusts far more quickly than the discourse surrounding higher education. If anything, the question of "making honest choices about what society must orient around" seems better left to the market than to academia or punditry, both of which are rigidly hierarchical.

4) That doesn't mean the current market for higher education is perfect by any means. What I'm saying is just that, of the problems facing higher education, whether the system is churning out enough programmers is not a major concern in my opinion. A much better question would be: "if we're making honest choices about which industries should grow in the U.S., should we be allowing all these new realtors?"

[Foucaulf]

Thanks for reiterating/clarifying, Foucaulf. This does actually answer my question as a question, although partly it was a rhetorical question anyway, inspired by frustration over your point 1.

I should clarify that I don't actually think academia or punditry as they currently exist would really be any better as socioeconomic trendsetters than market forces are and I'm sorry if I implied otherwise.

No problem, sorry for being facetious as well (I mean, my argument actually is a good example of not thinking beyond the confines of neoliberalism)

It's a curious fact that I'm quite more right-wing on higher education than the rest of my profile would indicate (contrast with all the European right-wingers who are defensive about their public education).

I think the reason why comes down my points 3 and 4 in the last post. To expand on that, programmer fetishism isn't a new feature of US education policy: I would trace the tradition of politicians throwing money to make technological education go the way they want to all the way back to Sputnik. Instead of trying to achieve education goals by lobbying and flattering these politicians' sensibilities, you should let philanthropy keep a few private schools alive or create a regulated private student loan market.

It is easy to counteract my point by bringing up the for-profits (or, in the case of programming, "boot camps") and their exploitative business model. But elaborating further is both going off-topic and out of my pay grade.

[Foucaulf]

Going to refashion this thread once more in light of an article that hit up the econ blogosphere: "Why do so few people major in computer science?"

You must be logged in to read this quote.

You must be logged in to read this quote.

I stand by what I wrote earlier in this thread, except for what I wrote about "persuasion." It is remarkable how ineffective efforts to persuade people into getting computer science degrees can be, to the point where the sheer intensity of continued rhetoric tips off how desperate the persuaders are.

Let me also admit I may be barking up the wrong tree: people could be taking minors or just a few courses in computer science, become software developers and this wouldn't be reflected as well in the data. In absence of better data, though, I still think there are two puzzles here:

1) Why has the enrollment gap in computer science majors and other STEM majors not widened, given the job openings gap in those two fields have widened?

2) Why has the enrollment gap in computer science majors and Masters in Education not diminished, given that the latter leads to increasingly devalued positions and vice versa for the former? (I think it makes more sense for now to think about teachers switching over to becoming programmers instead of nurses, cops, etc.)

I'm hoping for some responses before I write down what I mean to say.