More CS Ph.D.s than ever, what about research jobs?

According to the latest CRA Bulletin, the number of new science Ph.D.s is increasing steadily in the USA. It is now at its highest level since 1997, with 30,000 new science Ph.D.s a year. Meanwhile, the number non-science Ph.D.s is holding steady at 15,000 a year.

What is fascinating is their Figure 2, where it is clear that the number of new Computer Science Ph.D.s is growing exponentially since 2004. Up until around 2002, there were less than 1,000 new CS Ph.D.s. In 4 years, the number has reached 1,500 new CS Ph.D.s a year.

I have no number about the number of new tenure-track positions in Computer Science, but my guess is that, at best, it is stable. After all, Computer Science departments lost at least half their students since 2000 and things are not better, though we did stop part of the bleeding. I have good friends who have been looking for academic jobs and they are having a tough time. When we last opened a position, I was amazed at the quality of the c.v.s we got: some very good people have been hunting for a professorship for years…

If I am right, this means that we will have a serious problem on our hands. The overwhelming majority of Ph.D. students plan to get a professorship. Meanwhile, we are flooding the market with new Ph.D.s because training Ph.D.s is a great way to spend grant money, and to publish more papers, so that, in turn, we get more grant money… So the number of new Ph.D.s is growing exponentially. It helps also that hardly any students who plan on getting a Ph.D. reads my blog, and among those who do, few listen to my advice. How long can we sustain this system?

Thankfully, the industry market looks better than ever. At least, everyone seems to be getting an offer from Google. But again, I do not have any number. The trouble with industry jobs is that if you are in an area like Montreal, there are just not very many sensible research jobs in industry unless you are in the pharmaceutical industry. Government jobs are better, but not everyone is happy being a civil servant. But even if the industry and the government eat up this flood of new CS Ph.D.s, the value of the Ph.D. degree is going to go down. Elementary economics! I have said it again and again. Ph.D.s are part of an ecosystem. They can reproduce. But reproducing too fast, without any predator lurking, is going to lead to trouble.

My solution is quite simple. We must inform incoming students better about jobs. I am always amazed how Ph.D. students do not even believe the numbers I quote to them. Very often, they just go along, believing that they will inherit a professorship if only they can get this thesis finished. It does not work like that! We need to tell them! This simple step, called telling the truth, will go a long way toward making the ecosystem saner.

Hint to students: Consider a Ph.D. in accounting, law, or marketing if you must get a Ph.D. I hear that academic departments cannot fill these positions and end up betting for the few fishes who were crazy enough to get a Ph.D. in accounting.

Source: Thanks to Owen Kaser for pointing out the numbers to me.

14 thoughts on “More CS Ph.D.s than ever, what about research jobs?”

  1. “the number of new Computer Science Ph.D.s is growing exponentially since 2004”

    I wonder what is causing this growth? I spent 9 years in university (4 for Bachelors, 1 for Masters, 4 for PhD). Let’s assume this is typical. Let’s assume the majority of the CS Ph.D.s also have a bachelors in CS. This suggests something happened in 1995 that inspired many people to head in a direction that resulted in them getting a CS Ph.D in 2004. What happened in 1995? That was the year the World Wide Web hit the mainstream. Coincidence? I think not.

  2. I’m perhaps an outlier since I just received my CS PhD and will be moving up to Montreal from the states to start a job in industry.

  3. Sadly, one motivation for students to continue with a Ph.D. is that the job market for non-Ph.Ds in CS is not very exciting. Web development / embedded software / assembly line development in large enterprises …. I know a lot of people who just drop out of the field after 5-10 years and turn to more interesting things like Swimming Coach, Yoga Instructor or Pranic Healing. I know one who even became a (very good) professor of Education.

    Most of industry doesn’t know what to do with Ph.Ds in CS. – except Google, perhaps. At Nortel, I saw a very bright Ph.D. in Neural Networks end up on the assembly line coding Pascal for a telephone switch. It was pretty depressing actually.

  4. “The overwhelming majority of Ph.D. students plan to get a professorship. ”

    From my perspective at a large CS department, this isn’t necessarily the case. I would say the ratio is closer to 50/50 at least in my area (language technologies & information retrieval). This is one area where the industry opportunities are particularly good (google, yahoo, microsoft & numerous startups doing everything from ‘smart’ email systems to blog filtering).

  5. Like John, I’m not sure I agree that most students walk into a program assuming they will walk out professors. In fact, I’d guess that the percentage of students believing this is probably a function of the esteem of the school, the field of study, and the citizenship of the student.

    The same link Daniel posted has a figure suggesting that most of the growth in PhDs is accounted for by non-US citizens. It would be interesting to see this broken down further by field and school ranking. There may just be an increase in industry-grade PhDs.

  6. Perhaps I am a bit of an idealist, but I’m doing my PhD not for where it will let me go, but for my own interests and self-gratification. I do it because it’s a challenge, and I fully expect the step after completing it to be a challenge.

    The scarcity of professorships and the doldrums of assembly line programming — I’ve already been told I’m “overqualified” for some jobs — won’t dissuade me from finishing and it certainly won’t stop me from trying to make the interesting. I have ideas I want to explore and whether I do this in a tenure-track position, within an existing company or making my own company doesn’t matter to me. What matters is that I explore ideas that I think are worthwhile.

    I totally agree with you Daniel: Students going into CS PhD programs should be told up front about their prospects. But perhaps they should be encouraged to look for ways to explore ideas without having to rely on academia or existing research facilities.

  7. I am a software engineer at a successful search engine, but I plan on entering a PhD program next fall to study the intersection of information retrieval and natural language processing.

    Why am I giving up a successful career as a programmer to work on a PhD? Is it worth it? From a purely monetary perspective, it probably isn’t. My reasons are similar to those of Geoff, I think there are interesting and challenging problems being studied in academia and I want to a part of it. Although, judging from the recruiting at conferences, the job prospects for PhDs with my specialty is currently quite healthy.

    As far as what I hope to do with my degree, I’m keeping options open, but I’d say I am more inclined to go into industrial research at GYM or a startup rather than looking for a professorship.

    Taking a broader perspective, I like the field of CS, but basic software development jobs are being outsourced to countries with cheaper labor. I need an advanced degree in the field in order to stay ahead and have long-term career advancement opportunities. I also need to be flexible adapting my knowledge to benefit research in other disciplines.

    One pattern I see is that there is less opportunity in “pure” C.S. and more opportunity for those with advanced interdisciplinary degrees in areas like computational biology. My advice for those considering a PhD in CS would be to also consider a PhD in one of these emerging fields.

  8. 1995? It doesn’t take 12 years to go from a bachelor to a PHD. A lot of these folks started college during the dot com boom. By the time they graduated, the jobs were in India, so they just stayed in school. The number of CS majors has dropped by 70% since 2001. I expect the number of PHDs to drop by a similar amount or more.

  9. I doubt the “overwhelming majority” of CS PhDs plan on an academic career. I am a prof. at a top-20 department, and estimate that fewer than 1/3 of our graduates look for academic jobs after graduation. I also suspect that the fraction drops as the dept. rank drops, and remember that most CS PhD are not coming from top-10 schools.

  10. I’m a CS graduate working as a Solution Architect in industry. I have always been apalled at how little of what gets taught in CS program gets used out here where it needs to be used. For the past three years I’ve been on the Semantic train, with Semantic Integration, Description Logics and Reasoning being my primary interest. I’ve considered going back for a PhD for my own growth and interests but I am seiously doubting the worth of that degree. If I did not have a family or money where not an object I would definitely do it. As it stands now it would be fiscally irresponsible for me to do it.

  11. Another Area that is extremely mentally fruitful is pure mathemathics. I was led down this road when researching ontology matching. I bumped up against metric geometry, topology and similarity searching. All fascinating stuff. Wish I did not have to be a slave to money, but by the same token 13 years of a career in industry has given me a very comfortable living.

  12. I guess what I’m saying in all this babble is that I think we need more PhD’s but the market will not support them, because their work is not being used.

  13. Hi everyone !!
    I am a CS graduate student. I am about to complete my MS in CS . I am in a bizarre to go for PhD in CS or not. Can anybody tell me is it worth to do PhD in CS for getting a good position in the computer industry in US.

  14. PhD students in New York, Chicago, London and some other cities should take a look at the financial/capital markets area. The main attraction: there are a lot of bright people, including top notch software developers and PhDs from a variety of fields. There are frequently physics PhDs who get shunted this way. As usual whether it’s worthwhile depends on which firm. As a Goldman alum, I highly recommend them, though the interview process is a bit, umm, intense. It’s also not a good place to be if you’re not ambitious. Yes, its a bit of bloodbath at some firms right now, but there are still growth areas. But students in those cities probably already know all this …

Leave a Reply

Your email address will not be published. Required fields are marked *