How Hopeless A PhD?

Imagine that you have some estimate in your mind of the odds of becoming a professor, given that you start a Ph.D. program. Now imagine you see an article titled “The disposable academic: Why doing a PhD is often a waste of time.” How much do you expect that to change your estimate? Yeah, it should lower your estimate a bit.

Now consider actually reading the article. How much on average do you expect your estimate to change then? If your belief changes are rational, you should never expect your estimates to change – they might go up, might go down, but on average stay the same. OK, but do you so expect in this case?

Here is the article; test yourself. Quotes below the fold.

America produced more than 100,000 doctoral degrees between 2005 and 2009. In the same period there were just 16,000 new professorships. … In America only 57% of doctoral students will have a PhD ten years after their first date of enrollment. … British men with a bachelor’s degree earn 14% more than those who could have gone to university but chose not to. The earnings premium for a PhD is 26%. But the premium for a master’s degree … is almost as high, at 23%. … In one study of British PhD graduates, about a third admitted that they were doing their doctorate partly to go on being a student, or put off job hunting. Nearly half of engineering students admitted to this. (more; HT Omar Al-Obaydli)

So if you apply to a PhD program and get rejected, you haven’t lost much.  If you start a PhD program in the US, your odds are at least 57% of graduating, and pursing a PhD lets you enjoy a student lifestyle, if that appeals. If you quit early, you’ve only lost a few years and basically gotten a masters degree, which also has many uses.  Given a PhD your chance is about 16% of getting a prof job, for a total 9% chance given you start a PhD.  There are other nice non-prof jobs you can get with a PhD, your odds are better if you go to a better than average PhD program.  You could make sure to study subjects like engineering or economics with better job prospects.

Overall, my estimate of the chances of success went up after reading the article.

GD Star Rating
Tagged as: ,
Trackback URL:
  • fburnaby

    This had very little influence on my estimate of the odds. These are the types of data that I had used to calibrate my initial estimate, however.

  • michael vassar

    The article claims that America produces 64K PhDs annually and three paragraphs later, that America produced ‘more than 100,000’ doctoral degrees between 2005 and 2009. What’s going on here? Why can’t journalists fact-check their own articles against themselves, even in the economist?

    The article also failed to include info on the prospects of PhD students at the best schools, and most critically, info on the prospects of getting professorial jobs at top 5, top 15 and top 25 research universities and equivalent small liberal arts schools. Who cares if 16K professorial jobs opened up if 95% of them involve teaching 15 classes a week to IQ 115 undergrads who just want a degree and having your papers ignored by the people who set the direction of your academic field?

    • Doug S.

      “Between 2005 and 2009” is more than one year. 64,000 a year becomes “more than 100,000” in two years. I don’t see the contradiction…

      • In context, it’s weird to say “more than 100,000” instead of “more than 200,000” but not the sort of thing to go to war over.

        Michael, I think your second paragraph is a bit unfair. Its first sentence is actually pointing out a virtue of the article, although it makes it less interesting to high octane IQ performance/prestige nerd like you and me. Its second sentence seems unreasonable to me. One can argue that lots of shitty schools are unecessary deadweight, but I think it’s hard to argue that only the top 25 research universities (even if you just mean in the US) and equivalent liberal arts colleges are systemically useful. 1 research university per 10 million population size seems too low to me. I think it becomes debatable at 1 research university per 3 million (assuming a 100,000 person research university and support organizations workforce serving a 3 million client population) and my personal instinct leans towards 1 research university per 500,000 person population, since I think they’re about the most valuable things towards which to redistribute private income. and worth being about a 20% tax on our productivity.

    • Hamilton


      I do. The job you described sounds fine to me. Indeed, unlike yet another journal article that nobody reads, it sounds ideal.

    • Read more carefully before complaining. The second numbers come from a particular source and are reported as coming from that source.

    • michael vassar

      Hopefully Anonymous; I did just mean US universities, but I think we disagree about which schools are useful. Personally, it seems to me that for the most part, if you aren’t at a top 25 program your work will just be ignored however good it is, so I don’t see much point in other programs. I think that many people pursuing academic careers have the same opinion.

      For what its worth, the math SAT scores of 12-year-olds who go on to become professors in technical subjects at top 25 departments average about 3.75 sigma above the mean. For non-top-25 departments the average is more like 2.75 sigma above the mean and is the same for 25-50 as it is for 200-400.

      I’d be happy for federal funding for more universities if I thought that their productivity was simply lower on the margin, but I actually think, for reasons I won’t get into here, that their productivity is negative, that is, that they retard over-all scientific progress.

      • This rankings list seems pretty dense with influential research universities after #25, but I think it does start to get rather spotty after #56 (University of Maryland, College Park). So maybe we sustain great research universities at about a density of 1 per 5 million people?

      • didn’t include the link, but I’m using usnews.

      • I think it’s an important economic efficiency question, right up there with questions of healthcare, housing, justice market structures and expenditures, how to optimize the design of our education system.

        Too bad the discussion dies here -the essential problem in particular with higher education seems to me that we’re way overspending for reasons similar to healthcare: to show that we care, and to send status and prestige signals. The difference is that I think perhaps more money should go into elements of education (suchs as basic science, including social science, research) but that the American population should probably be educated very differently, on a middle similar to my understanding of how Germany works (tracking to trade school for the vast majority of people).

        I don’t know if Germany incorporates this, but I suspect we should also incorporate vocational retrains into normal career progressions (that people should expect to start at the lowest level about 7 different times during their working years, for example, with 6 month retrain job skill certificates each time.

        Quite a few arguably useless but tempting fields of study perhaps should have nationally limited student slots, or heavily taxed tuitions, or both.

  • Constant

    You also know whether you intend to become a professor. That greatly affects your odds. The 9% applies to the whole population.

    • A very good point. There are a number of professions which also need PHD to be taken seriously.

  • Jim

    It also varies greatly depending on the field. My field is biology/biotech. Chances to get an academic job in it are much lower than 15%. PhD glut is great because the demand for graduate student serfs is great AND because before an academic job can be obtained there is also an additional obligatory serf-like position called postdoctoral “training” (which used to be three years on average but now has four-five as more typical). The jobs in the industry are equally open to those with MS and PhD, so one is definitely better off not getting a PhD and having about four years head start.

    The whole system is a scam perpetuated by tenured professors for their profits.

  • Jim

    Just to add to my previous post. A concrete example. This was published awhile ago in Science:

    Those with access to Science can read it here:

    Some reference points:
    1. Yale’s Molecular Biophysics and Biochemistry is indeed a top notch program that generally attracts the best of the best who then end up working for mentors with the most influence, power and money (and thus provide most cronysm/networking opportunities).
    2. The situation described for the class of 1991 has only worsened significantly since. Those who say that there are many good opportunities outside of academy are either deluded or are outright liars. Industry doesn’t need and never needed that many PhDs either.

    The solution is simple: take away tenure. Then the scam crumbles and market has a chance to sort things out to benefit public rather than few winners in their academic sinecures.

    • Jeffrey Soreff

      Those who say that there are many good opportunities outside of academy are either deluded or are outright liars. Industry doesn’t need and never needed that many PhDs either.

      The solution is simple: take away tenure. Then the scam crumbles and market has a chance to sort things out to benefit public rather than few winners in their academic sinecures.

      Yes, industry won’t absorb the extra PhDs either (just look at the situation in chemistry!E.g. see Dr. Lowe’s blog).

      Taking away tenure won’t help the grad students. It will give the universities more control over the professors, but that probably just cranks up the competition of professors for grant dollars. It doesn’t do anything to make the employment prospects for students more transparent.

      Prof Hanson: Any suggestion on how prediction markets for grad student employability could help here? The time scales are moderately short, the data for validating bets should be accessible at modest costs. Is this a case that look viable?

  • cournot

    I don’t understand why this is a scam. Why can’t science Phds figure out the odds for themselves and then adjust accordingly? I think de Long once had an article about why there is more of a balance of supply and demand in Econ Phds and pointed to top departments doing “birth control” in the 1980s to discourage an explosion of grads. But humanities Phds, and science Phds with few chances of a tenured job should be able to figure out (by now!) that jobs are scarce. [Well, maybe not humanities Phds…

  • Sam Wilson

    Ah, that’s the article Prof. Rowley had referred to. Many of us were confused.

    I’m not so sure I’m completely on Bryan’s side when it comes to the usefulness of higher ed. I can’t much gainsay that the bulk is Tullockian, lost and forgotten in the aether of revelry and regret, but the general equilibrium solution is hazy at best. If my priors were consonant with what I know now, I don’t think I would have done anything radically different. My time at Mason has indeed changed my worldview, and I am not convinced that I would have had the impetus to do so on my own. Even our little leisurely gametime chats (of which there are far too few) have forced me to reconsider many of my formerly unassailable warrants.

    In that respect, grad school is quite something of a consumption good. As an investment, I’m willing to defer to Tyler’s voluminous commentary on the subject (links omitted for brevity) and state that if I’ve gotten anything especially useful from the courses I’ve taken (and will take), it’s been methodology and Russ Roberts’ elan in communicating economic thinking to non-economists.

    Robin Hanson is the economists’ economist. He’s the real hidden treasure at GMU. I could roll through the entire faculty, singing hosannas along the way (particularly to those I can confidently call friends), but I’m quite comfortable claiming that no living thinker has been quite the catalyst for updating my priors as Robin. That alone is worth the price of admission. And there lies my criterion for the value of a graduate program. If students can’t find their own Robin Hanson (or Thomas Schelling, or Gore Vidal, or Emil Fischer or whoever), why bother except for the pure signaling value of the sheepskin? And if that’s all you want, I assure you there are cheaper and more rewarding signals than a PhD.

    Also, statistics are bullshit. You can quote me on that.

  • The article is comprehensive, after a homo economicus style analysis it does express the idea that some people might be doing a PhD for other than income-maximizing reasons.

    In many of its ‘graphs the article reminds me of a Consumer Reports analysis of musclecars I read in about 1988, comparing Mustang GT, Camaro, Firebird, and similar cars. The article starts with the statement something like “First of all, we have no idea why anybody would want a car like these.” They then proceed to rate the cars against each other, downgrading the ones that are noisier, and downgrading some for trading off performance against comfort. Yup they were right the first time, they had no idea indeed.

    In my opinion, too much “whining” gets reported as though it were information. You want to overcome a bias this morning? How about overcoming the bias of young smart neocortex-alphas but aggression-testosterone-betas in reporting their own state of life. They (we, actually) whine! We whine at our low stipend even as we are flown around to fantastic meetings and spend many of our days interacting with the top few minds in the world, priveleged to wor on the cutting edges of various topics. We whine because we dreamed of rising to the top, of being 1 in 100 in the pointy end of the intellectual professions, and for many of us it didn’t pan out that way. You might as well go interviewing failed rock musicians, basketball players, and internet company starters to see what their whining sounds like. But this article sticks to a journalistic style: all it does is report the stuff as though it were information about the PhD rather than information about the whiner.

    Making things better for the PhD student on the basis of this article would be a mistake. It SHOULD be competitive. I would bet a careful analysis, alluded to in some of the final ‘graphs of the article, would show a value to society of even partial PhD training that is significant. The idea that PhDs make only a little more than Masters, and even less in some professions, well how about a little thinking? Maybe the populations that get a Masters do it to improve income while the populations that aim at PhDs are doing it for other reasons.

    I plausibly have less money for getting a PhD. But I have a thesis on my shelf, very likely the only “book” I will ever write. I have a group identification with people I admire. I have a recollection of a stimulating seven years that left me nearly spiraling in to depression when I realized what I was leaving behind upon becoming a professor at a 2nd tier school in the boondocks. I still do linear algebra and do amazing data analysis with fantastic graphical representations that all fly out of my brain onto a keyboard. But my reward now is more money and some patents, and it is not clear to me that this is better than wandering the halls at a scientific meeting in Cannes having graduate students from around the world seeking me out so they can have a chance to talk to me.

    Meanwhile, I never would have been one of these engineers that goes to work and gets a masters to make more money. It simply wasn’t my drive. But having gotten the PhD I had what I needed to get a job in industry, making more money than some with Masters EE, but less than a lot others. Yes I admit the statistics in the Economist article are interesting and should add to our thinking about the system. But the only “juice” they kept was the whining. PhDs are the college basketball players and top-tier professorships are the NBA. It isn’t mostly about the dollars.

  • vaniver

    My estimate stayed the same, but this issue has been of personal interest to me in the last few years so I had up to date data. My estimate was ~10%, which I expect isn’t significantly different from the article’s. A rough justification is that in a 20 year career a prof will graduate 20 graduate students, and that by the end the professor will need to be replaced and the field will have doubled- so 2 of those 20 get a professor job.

  • Well, seeing as I have my Ph.D. in hand, and seeing how I am writing this reply from my job as 3rd shift front desk clerk at a hotel . . . I have to say that nothing would have made me change my mind. For me the Ph.D. in and of itself was the dream. Now, there are a lot of things that I surely wish I had known about what I need to do to get an academic job (and I wish my younger self would have done them), but that’s a different story.

    Pop quiz:

    If you get a B.A. in recombinant gene technology (chemistry minor), do two years of grad work in biology, then get a M.A. in English, and a Ph.D. in the humanities, what kind of scholarly work are you like to be doing?

  • Rob

    I believe she only glanced upon one important aspect of PhD programs in America: that the immigration laws are set up to allow for cheap labor to enter into the country as long as they call themselves students – graduate students to be exact. Companies don’t have the freedom that universities have in hiring these well-educated foreigners … if they did, you can bet that a lot less foreigners would be pursuing PhDs and a lot more pursuing full-time work in America.

    Change the immigration laws and a lot of her issues will subside.

  • Answer: Austrian economics — particularly Hayekian spontaneous order work