This morning, an article in the Guardian is making the rounds on the academic twitter feeds. The article is written by an anonymous graduate student, and it argues that graduate students are underpaid and that their salaries should be doubled. When reading the article, I couldn’t help but feel that it lacked the careful analysis, logical reasoning, and deep thought befitting of a graduate student. The author is either not aware of or willfully ignores how graduate students are paid and why. I am willing to give the author the benefit of the doubt, though. I recently had a conversation with a senior faculty member at my institution and she also didn’t really understand how graduate students are paid. So here’s the quick summary: per hour, graduate students are compensated better than most postdocs and lecturers. After tenure-track faculty members, graduate students probably get the best deal in academia.
A few caveats up front: I’m speaking only for natural sciences. Also, I’m speaking primarily for the US system. However, I got my PhD in Germany, and most of what I’m saying here applies to that system as well, as far as I know. I am aware that in absolute terms, graduate student salary is low. I think most professors are aware of this. In fact, our department is currently going through a strategic planning process, and the number one issue identified by our faculty was low graduate student salaries. However, I suspect that most people are thinking about a ~10% increase in graduate student salaries, not a doubling. Finally, I completely agree with this blogger that everybody in academia is underpaid. In fact, when I meet with high-level executives from industry and tell them how much I make, they just stare at me in disbelief. So keep that in mind. Nobody is in academia for the money. With that, let’s have a look at graduate student compensation, let’s consider for what work they actually receive it, and let’s see how it stacks up compared to postdoc and lecturer compensation.
Graduate students are enrolled in a university degree program. As part of that program, they are expected to take classes and carry out independent study. The independent study component is basically their research program, and the final demonstration that they have completed their independent study is the PhD thesis. Note that none of this effort is compensated in any way. It is a fundamental principle of university education, not just in the US but in most places, that you don’t get paid to participate in a degree program. In fact, in most places you have to pay tuition to do so. In the US, tuition is quite substantial, and ultimately the graduate students are responsible to pay it in full. Now, in most reasonable graduate programs in the natural sciences, graduate students don’t actually pay tuition, and they do receive a stipend. How does this work? One possibility is that students can receive fellowships, but I don’t want to address fellowships here. The more common scenario is that graduate students are being paid as either Teaching Assistants (TAs) or Graduate Research Assistants (GRAs) or a combination of the two. In this scenario, the stipend graduate students receive is for their TA or GRA work. Generally, TA or GRA appointments also come with funds to cover tuition. So on top of their stipend, graduate students receive substantial additional compensation, on the order of $10,000 a year at my institution, and amounting to much higher levels at other schools. Because graduate students are paid for TA or GRA work, not for their studies, TA and GRA appointments are generally half time, 20h per week. I remember, when I did my PhD in Germany, that my GRA appointment contracts always contained a clause to the effect that my supervisor was required to leave me adequate time outside my GRA duties to carry out my own, independent research.
So how much do graduate students actually make? Typical salaries are around $2000 per month. Lucky graduate students make a little more, maybe $2200-$2300, while the unlucky ones may make only around $1900. Remember, though, that these amounts are for a 50% time position. If you extrapolate to an annual rate at 100% time, you end up somewhere in the range of $50,000, give or take. Almost no postdoc and very few lecturers make that kind of money. (Postdoc and lecturer salaries are closer to $40,000, and often below that.) And remember, that’s without tuition remission, which goes on top of the salary.
Thus, when graduate students are employed to do research (as GRAs), they are payed more per hour than postdocs are. And when they are employed to teach (as TAs), they are payed more than lecturers are, for a comparable workload. In particular, when a graduate student works as TA for a lecturer, the student makes more money than the lecturer does for that course, even though the lecturer has more responsibilities. (At my institution, one course pays for 33% of the time of a lecturer but for 50% of the time of a TA, and the TA has the higher base rate!)
Graduate students that work in well-funded labs generally get employed as GRAs rather than TAs, at least a substantial fraction of the time. Strictly speaking, the work they do as GRAs is separate from what they are working on for their degree, but most professors will try to give students work assignments that align with their own research plans. So if you’re a graduate student, are employed as GRA, and have no assignments beyond working on something that will count for your thesis, you’re getting a really good deal. You’re basically given the option to double-dip, to use the same work for both compensation and education.
Now, as a thought experiment, what would happen if we actually paid graduate students at 100%? It’s simple. Most GRA positions would disappear. Already, if you do the math, you realize that a graduate student and a postdoc are comparable expenses. One year of graduate student salary plus tuition is comparable to one year of postdoc salary, in particular at schools where tuition is high. (As faculty member, if I want to employ a student as GRA, I have to pay the student’s tuition from my grant.) I know many colleagues who say they’d rather hire a postdoc than a graduate student because they get a better deal for their money. If you now increased the annual cost of a graduate student by another $22,000, nobody in their right mind would ever hire another graduate student. You’d have to double the salaries of everybody, graduate students, postdocs, lecturers, and professors, to keep some sort of balance among the various compensation levels, and to prevent professors from hiring only postdocs. I’m personally not opposed to doubling the salaries of everybody, but realistically the money is just not there.
To summarize: Of all the things we can (and should) worry about in academia, high undergraduate tuition, low funding rates, poor job prospects for postdocs, tenured professors who’ve turned into dead wood, and so on, paying graduate students at 50% effort is the least of my worries.
 Incidentally, in computer science it’s the other way round. Because most PhDs in computer science can go to industry and make a lot of money, postdocs can command a much higher level of compensation in that field, and hence professors are more likely to hire graduate students than postdocs.