[kaosjester]

Don't forget the part where those graduate students do some 80% of the work and make <20% of the income of their their supervisor (and, in many cases, <10% or none at all in several), and the university hosting the lab absorbs some 50% of all grants "off the top." A half-million dollar grant pays a single graduate student less than 20k/year for maybe 5 years, which, in the context of other companies that get these grants, is absurd. Independent research firms working from the same grants pay their employees competitive wages and still accomplish research. If these companies could issue Ph.D.s, academia would shrivel and die: imagine working at a company for 4 years, getting a salary 3-4x what a graduate school would offer, doing real work in a professional setting, doing enough research to write a dissertation, and receiving a degree. The entire incentive to attend a university would melt away.

[jamesblonde]

That's the Anglo-Saxon model. In Sweden/Norway/Switzerland/Netherlands, a Phd student costs at least 100k dollars/year to a lab, with full-pension payments. It's a (relatively well-paid) job here.

[jostmey]

I don't think many people (even graduate students and postdocs) realize how much money universities suck in with grants. For every RO1 a professor gets, the university gets to add a substantial (typically greater than 50%) indirect costs. The PIs don't care because their budget is the same, but what it means is that there are fewer grants being handed out. Then the PIs somtimes pay for the "training" that graduate students and postdocs get using their own funds. For a top university, this can be 50K/year.

[greeneggs]

That's not exactly how it works. If an agency wants to give a $1 million grant, for example, and the PI's university has a 60% overhead rate, then the PI will draw up a $625,000 budget (as 625*1.6=1000). So the overhead definitely matters to the PI---except there is little the PI can do about it after joining a university. Most big universities have similarly high indirect cost rates, which they steadily increase over time.

UC Berkeley (2016): 57%

MIT (2018): 59%

Harvard (2018): 59%

Stanford (2018): 57%

This story, from 2013, gives some of the context and history, as well as averages for universities and other research insitutions (which can have much higher overheads):

http://www.nature.com/news/indirect-costs-keeping-the-lights...

[killjoywashere]

And all that overhead money is going to 3 things: increasing healthcare costs for the burgeoning retiree population of staff, increasing the number of administrators subservient to the king (I mean, provost), and new construction. The political class sees these additional jobs from the last two as the core justificatiin for the government increasing student loans for students and overhead for grants. They don't comprehend the science or frankly give a shit about it. Maybe when they get cancer they'll have a vague sense of interest in their organ system of choice, but that's about it.

[Fomite]

Indirect costs cover things that we aren’t allowed to write into grants. And at my institution at least, it’s questionable as to whether or not that indirect rate actually fully covers the cost of research.

[pcrh]

In the UK at least, independent research at pharmaceutical companies or Contract Research Organizations typically costs 1.5- to 2-fold what it costs at a University. Mostly because they pay higher wages than PhD students get, but also because they have heavier overheads. They're also typically less nimble as their research programmes are larger with more layers of management.

[chillydawg]

Uni research is probably minimally taxed, too.

[matt4077]

How, exactly, is commercial research taxed? Only profits are taxed, and research is all expenses, and can be deducted from profits for tax purposes in the year they occur.

[PoachedSausage]

In the UK a commercial profit making company, even if it is doing research would still be subject to a form of local property tax called business rates[0]. Whereas a university would likely get relief from such taxes.

[0] https://en.wikipedia.org/wiki/Business_rates_in_England

[cyht]

They have this in some European countries like Denmark where it's called an "industrial PhD." However, they also compensate graduate students more fairly and seem to have more systematic and closer collaborations between academia and industry.

[mcguire]

Such an independent research firm could grant Ph.D.s. Getting a Ph.D. in US universities like where I went is primarily a matter of satisfying a committee of existing Ph.D.s that the student has done suitably creditable work.

A firm with good researchers, whose newly minted students also did good work, would soon have the suitable reputation. And those letters after your name are only as valuable as the reputation of the institution.

[PeterisP]

At least my university does offer such an option, it's not often used, but sometimes is - if you've done (and published!) an appropriate amount of research in the industry, then you can apply to the committee to defend a thesis and get the degree without doing a PhD program. You still need to write the thesis, though, and it's much easier to do it while being paid as a PhD student/researcher.

[gaius]

A PhD from say Microsoft Research ought to be as good as any. Or back in the day, Bell Labs.