Being a graduate student is in some sense in the middle of two extremes: being a student primarily benefiting oneself and being a paid worker benefiting society. Before graduate studies, one is an undergraduate where nobody would expect to be paid to be an undergraduate. After graduate studies, one is (hopefully) going to be paid a paid a sum commensurate with the skill the knowledge accrued during graduate studies and be doing work that, by and large, can be said to "benefit society". During graduate studies, however it is somewhere in between. Graduate students are both benefiting themselves by increasing their future potential and also doing teaching and research that is benefiting society. How much, then, ought they be paid?
How graduate studies typically work:
A graduate student typically has two major roles. The smaller of the two roles is what makes up a higher portion of their funding package: teaching work. They are being teaching assistants that run tutorials, mark midterms, hold office hours, and the like. In latter years they are also course instructors which run full courses much like a professor. This is unquestionably real work. If you blind yourself to anything but this specific work, graduate students are typically well compensated for this specific work, over $40 and hour at the University of Toronto. That is very comparable to a moderately high skill hourly wage in the private sector. I have no problem here. The issue is that the number of hours is fairly small, something like 200 hours a year, so about 10% of what "full time" work is.
The larger role is doing research to produce a thesis. This sucks up the vast majority of the time spent in the latter years after courses and comprehensive exams and the like are completed, and for this portion of work graduate students are paid very little, far less than what research work in the private sector might be. Together the compensation for these two components is typically a less than poverty line amount to take home.
Graduate students are usually paid an amount in between these two extremes of undergraduate and full time work, something like the 15k minimum funding package offered by the University of Toronto. It varies from school to school. This is an amount far larger than what undergraduates are paid (nothing, minus tuition), and an amount far smaller than what is typically paid for full time work after one graduates. It doesn't seem immediately wrong to me that this in between place in life gets paid an in between amount of money, the question will be one more of where, precisely, to put the number.
No graduate student becomes a graduate student for the money earned while in graduate studies. One is almost certainly going to be paid far less than if one tried to find a job after completing their undergraduate degree. However, the main motivation for being willing to spend 4-6+ years severely underpaid relative to the undergraduate degree job market is because of a combination of increasing future earnings (a PhD being far more competitive than a BSc) and/or a genuine passion and interest for the subject of one's studies. Part of the calculus that goes into the decision to go to graduate school is the assumption of a (relative) loss now.
Two types of arguments:
As with many questions of policy, there are typically two rather different types of arguments. Firstly, we have what I will call "economic" arguments. These claim some sort of larger economic or perhaps soceioeconomic benefit from the policy. For instance, when talking about TFSAs, one can argue that a society with more middle class savings has smaller recessions which is good. This is at its root an economic argument.
Secondly, there are what I will call "social" arguments, that presents some form of normative benefit like equality or freedom that will benefit from the policy. I support gay marriage not because I think society will be better off economically, but because of these ideals of freedom and equality.
The best policies are often ones where there are strong arguments in favour considered both ways. For instance, during a recession, spending money on employment insurance is one of the best stimulative policies possible (as the money so quickly enters the local economy) so there are strong economic arguments for it but there are also strong social arguments about reducing inequality and suffering.
Returning to graduate funding, I'm going to find that sweeping macroeconomic arguments for increased funding are somewhat lacking, but that there nonetheless remains strong social arguments. The economics first:
In between two extremes:
Let's take this to the salary extremes to see the effect on graduate schools. Firstly, let's act as if graduate studies is comparable to undergraduate studies, as in it is primarily about advancing one's own knowledge and ability to compete in the market later in life. Let's assume we pay zero dollars for graduate studies, and charge tuition to boot. The same motivation mentioned above (that people would earn more in the future with an advanced degree) still applies, but the threshold changes. A half decade of going into debt requires a huge future return on investment that for many people won't be justified. The obvious effect is going to be a substantial decline in the number of people applying to graduate schools, resulting in less of the best BSc graduates ending up going to get graduate degrees. There are strong arguments that a highly educated society has many positive socioeconomic consequences, and so far less competent people wanting to go into graduate schools would be a loss. Of course, there would be the gain that provinces would have to pay less for universities and could thus tax less or spending elsewhere.
There is also something of a tipping point effect. Four years of unpaid undergraduate work is a serious strain on either family finances or building up of student debt for many people. It is doable, just, but hard. A decade of it (undergrad and graduate together) is simply too much to be sustainable for many people, no matter how positive the future return on investment. So this extreme seems like it would have a big cost in the terms of having a high quality, highly educated society with marginal benefits on tax bills.
Now take the other side, that graduate studies are paid similar to private sector jobs. Say, $40k per year, just to pick a number. This is going to make graduate school even more desirable since your get the future earning potential as well as being paid reasonably well immediately. If we hold the number of graduate positions constant (this will already be a substantial burden on governments to finance, increasing the number of positions would be yet more again), what this increased incentive will do is mean that the composition of graduate students will improve. Very bright people who might not have done graduate school will now do it and so we will have, on average, a better cohort of graduate students (and a worse cohort of people who go to the jobs market after BSc). The problem here is that we likely come upon a resistance point. The best and brightest typically (not always, but typically) already go to graduate school, because the incentive structure already is enough for them. And for those that don't - bright people who leave to the job market leaving less bright people to get accepted to graduate school - it isn't clear to me that keeping these people in graduate school is actually that big of an advantage for society. I also want very bright people to found startups in the private sector and the like. So this extreme seems like a big cost on provincial coffers for little socioeconomic gain.
In between the two extremes thus seems like a reasonable place to be. Not $0, not $40k. Can we argue for the difference between $25k and $15k (the UofT minimum now)? Clearly this is going to require more detailed arguments than the more sweeping big pictures arguments thus far. Indeed, from these sort of bigger picture arguments there doesn't seem to be any clear reason why one is better, it really depends on the sensitivity of these smaller changes in price to the incentive structure.
Unaffordable for some:
For me personally, the level of funding is sufficient. While obviously I would like more, the advantages that graduate school provides meant that the benefit relative to working after my undergraduate were not that sensitive to how much I was going to be paid. However, I have various advantages (family support during undergraduate, having savings, cost sharing with my wife, etc) that meant that living on graduate funding was a very doable position. I have a reasonable quality of life despite living in Toronto and extra money at this point would only be saved for future spending.
For many, especially those without my financial advantages, this is not the case. Without financial support such as being able to live at home during their undergraduates, one will often leave with substantial debt and spending a decade running a deficit is simply untenable for many people, particularly those with children. Graduate school funding minimums are substantially below the poverty line, and while this can be often tolerated for amount of time (during ones undergraduate, for instance), it is hard to maintain for so long.
Economically, the consequence of graduate funding being at this level is that it pushes some people out of graduate school and into the general market based on its level of unaffordibility. As such, the calibre of the graduate school cohort declines as otherwise competent people don't enter.
Higher packages probably don't cost society much:
If we tax higher incomes slightly more (as in the incomes graduate students will often make) and pay graduate students a little bit more, what is effectively happening is a forward shift of future money to the present. The income curve with respect to time will be less steep with more money made in the first decade and less in latter ones. Even if we act as if there are precisely no other socioeconomic consequences (this move very likely isn't zero sum), it is probably a preferable switch for most people who will get graduate degrees. It would, of course, be a negative for people who end up earning high incomes but didn't do graduate school, but these are people already privileged by ending up with high incomes and being paid more during the years graduate students are students.
The net economic argument:
Putting all this together on the economic front, I think it is something of a wash. The majority of the cohort is not sensitive to very large increased wages in the sense that increasing pay isn't going to make more of the best students go into graduate studies versus entering the job market after undergraduate as the advantages of graduate school are sufficient to dominate already. That said, there is still nonetheless a portion of the potential cohort where the financial challenges do push them out of the possibility of grad school. Large increases simply are not necessary to increase the quality of the graduate school cohort, but small increases will pull those who simply can't continue at the current levels (due to debt, kids, lack of family support, etc) to stay in. Taxing more on higher wages to pay for higher funding now largely transfer the time distribution of money from two similar cohorts.
What precisely the "slightly higher" level of funding is debatable, and changes substantially depending on the local cost of living. One convenient metric is the poverty line which already takes into account many of the financial challenges facing people. Changing the minimum funding package to being the poverty line will pull up those currently unable to do graduate school based on financial challenges.
The social argument:
Regardless of the above economic considerations, I think there are strong social reasons to consider moving up to poverty line levels of funding. At its core, this is an argument about equality. There is a portion of the potential graduate school cohort who can't go to graduate school due to financial challenges, whether this is the need to help out family, whether it is kids of one's own, whether it is due to past debt (for instance from an unassisted undergraduate). Disproportionately these people are going to come from lower income families.
I've always believed that a key role for for governance in society is to create a broad floor beneath which we can't fall. Graduate school is not like access to healthcare in that it is meritocratic and can't be done by everybody. There are plenty of ways where demographic considerations change one's potential to be able to get accepted for graduate school and nothing in this post is going to address this. But at a minimum, I think we can accept the idea that people should not be unable to attend graduate school based on reasonable financial limitations. There are people in that category. We should change that.