Will the new tax plan destroy American graduate education?

Yesterday, the House passed their version of tax reform, the creatively-titled Tax Cuts and Jobs Act, 227-205.  The higher education universe has been (rightfully) freaking out for the past week, as the bill would eliminate the tax credit on waived tuition, raising the tax burden on paid graduate students by anywhere from ~$3,000-$15,000 annually.  The New York Times said this change would “bankrupt” these students, and Forbes went so far as to say it will “destroy graduate education” entirely.  The reactions of my grad school friends echoed the media panic: some see a major threat to their savings, others are contemplating leaving school early to get a job.  Unsettled, I ran the numbers on my graduate financials.

At Vanderbilt, my stipend was approximately $28,000/year, which incurred a federal tax bill of about $3,700.  Vanderbilt’s tuition was around $34,000 annually, but it was entirely waived (essentially, Vanderbilt paid itself that money when I was a teaching assistant).  The new tax plan would require tax payment on the whole $62,000, bringing my tax bill to almost $11,000.  If nothing else were to change, my net income would fall from $24,000 to $17,000…coincidentally equal to my average living expenses for my two years in Nashville (granted, I live a low-frills lifestyle).  I can imagine how this would be a major deterrent for potential graduate students, and how only those from affluent backgrounds could attend under this system.

But something else would have to change, based on the current economics of graduate education.  Research institutions make money and build prestige according to the number of productive labs/departments on campus, and these departments depend on graduate student research/enrollment to thrive.  I don’t believe for a second that universities would sit back and watch as people stop applying for graduate programs.  For starters, they could lower graduate tuition to ease the tax burden: under the House proposal, the rate increase from 12% to 25% would occur at $45,000 (if Vanderbilt halved tuition, for instance, taxes would also be halved, topping out at $5,400).  These universities already benefit far more from skimming a large percentage off the top of any government and industry research grants awarded to institutional labs.  Alternatively, the universities could raise the stipend to offset the taxes, deflecting a portion of the costs to the external grants that support some graduate researchers.  This would increase the university’s per student financial burden (by about $7,000 in my Vanderbilt example, up 11% from $62,000 to $69,000), but department employment could be maintained at ~90% of current levels if total funding for graduate students is constant.  With the job market for Ph.D.s as tight as it is, a slightly more competitive admissions process may not be the worst outcome.

Fortunately, this is not law yet; the bill must first pass the Senate.  The current Senate version mercifully keeps the tuition waiver tax credit, instead targeting interest on endowment investments (still fundraising from universities, but by a much more palatable means).  More than likely, the House’s provision to eliminate the tuition credit will be dropped from the bill, as most contentious provisions are.  While I believe there are many ways the academic research system should be improved, this bill is not an answer.  In fact, it’s not fully clear what the purpose of this proposed change is, whether it’s a sneaky bid to undermine higher education or just a misguided attempt to close a loophole.  Anyway, hopefully the Senate gets it right.  Hey, I’m not in grad school anymore, I am allowed to be hopeful!

Seismic Signatures of Tornadoes

Two earth scientists from the University of California at Santa Barbara recently published an article describing the seismograph recorded near the Joplin tornado in 2011.  The seismic data, freely available for research, was recorded by the USArray Transportable Array of over 400 seismometers nationwide. Of the 32 stations recording within the state of Missouri, only the station nearest Joplin picked up the tornado’s seismic signature as the deadly vortex passed within 2 km of it. This station recorded a distinct wave signature that correlated with its strength, from a monstrous EF4/EF5 to a dying EF1/EF2. So why are we just reading about it now, a whole 6 years after the storm?

The history regarding seismic recording of tornadoes is porous at best. The classic “freight train rumble” has been observed since pioneer days, an observation that has sparked research into seismic, sonic, or infrasound measurements around tornadoes. Although hurricanes and other ocean storms were seismically tracked as early as 1947, the first proposition to track tornadoes using seismographs is in this 1994 article. The article itself is fascinating: as seismic data was unavailable to the researchers, they chronicled several eyewitness accounts as evidence of ground tremors. The rest of the article comprises an instruction manual for the seismic study of tornadoes, a method that several research teams adopted. However, initial collections struggled to differentiate important features like strength and location of these tornadoes, and the focus quickly shifted to infrasound detection. I can only suspect that these past uncertainties led to extra delay and scrutiny for the recent Joplin tornado study, though I believe its publication restores relevance to the idea that seismology could potentially help identify at least the strongest tornadoes.

How cool would it be to spot a tornado based on a real-time seismograph reading?  Rather than waiting for a trained spotter to visually confirm a tornado on the ground, it should be rather easy to identify a tornado touchdown by its sharp pressure perturbation.  It would also be cool to have a realistic estimate of the surface wind speeds, a strong possibility if the proposed analysis proves repeatable. Moreover, if seismic detectors are placed in a finer grid, it could be possible to triangulate the exact position of a tornado, and significantly faster than radar to boot.  All in all, this could be huge (or should I say…earth-shaking!) for the field of tornado tracking, and I’m excited to see the research that follows.

Figure 1: Source strength amplitude derived from the raw seismic data during the Joplin tornado. The red shading indicates the duration of radar/damage-indicated EF4-EF5 intensity, and the blue shading corresponds to EF0-EF2 strength…the amplitudes do seem to scale with tornado strength! Source: Valovcin and Tanimoto, 2017.