Wednesday, August 20, 2014

Science and engineering research infrastructure - quo vadis?

I've returned from the NSF's workshop regarding the successor program to the NNIN.  While there, I learned a few interesting things, and I want to point out a serious issue facing science and engineering education and research (at least in the US).
  • The NNIN has been (since 2010) essentially level-funded at $16M/yr for the whole program, and there are no indications that this will change in the foreseeable future.  (Inflation erodes the value of that sum as well over time.)  The NNIN serves approximately 6000 users per year (with turnover of about 2200 users/yr).  For perspective, a truly cutting edge transmission electron microscope, one instrument, costs about $8M.  The idea that the NNIN program can directly create bleeding edge shared research hardware across the nation is misguided.
  • For comparison, the US DOE has five nano centers.  The typical budget for each one is about $20M/yr.  Each nano center can handle around 450 users/yr.  Note that these nano centers are very different things than NNIN sites - they do not charge user fees, and they are co-located with some truly unique characterization facilities (synchrotrons, neutron sources).  Still, the DOE is spending seventeen times as much per user per year in their program as the NNIN.
  • Even the DOE, with their much larger investment, doesn't really know how to handle "recapitalization".  That is, there was money available to buy cutting edge tools to set up their centers initially, but there is no clear, sustainable financial path to be able to replace aging instrumentation.  This is exactly the same problem faced by essentially every research university in the US.  Welcome to the party.  
  • Along those lines:  As far as I can tell (and please correct me if I'm wrong about this!), every US federal granting program intended to have a component associated with increasing shared research infrastructure at universities (this includes the NSF MRI program, MRSEC, STC, ERC, CCI; DOE instrumentation grants, DOE centers like EFRCs, DOD equipment programs like DURIPs) is either level-funded or facing declining funding levels.  Programs like these often favor acquisition of new, unusual tools over standard "bread-and-butter" as well.  Universities are going to have to rely increasingly on internal investment to acquire/replace instrumentation.  Given that there is already considerable resentment/concern about perceived stratification of research universities into "haves" and "have-nots", it's hard to see how this is going to get much better any time soon.
  • To potential donors who are really interested in the problem of graduate (and advanced undergrad) science and engineering hands-on education:  PLEASE consider this situation.  A consortium of donors who raised, say, $300M in an endowment could support the equivalent of the NNIN on the investment returns for decades to come.  This can have an impact on thousands of students/postdocs per year, for years at a time.  The idea that this is something of a return to the medieval system of rich patrons supporting the sciences is distressing.  However, given the constraints of government finances and the enormous sums of money out there in the hands of some brilliant, tech-savvy people who appreciate the importance of an educated workforce, I hope someone will take this possibility seriously.  To put this in further perspective:  I heard on the radio yesterday that the college athletics complex being built at Texas A&M University costs $400M.  Think about that.  A university athletic booster organization was able to raise that kind of money for something as narrowly focused (sorry, Aggies, but you know it's true). 

7 comments:

Anonymous said...

In your experience, has the cost of equipment/instrumentation been rising faster than inflation? If so, why do you think that is the case?

Douglas Natelson said...

I think one issue is that what constitutes cutting edge instrumentation has been getting progressively more expensive, as peoples' expectations rise. In general, I think instrumentation costs have gone up only a bit faster than inflation. (Faster, and service agreements even more so, because skilled technical labor costs are more of a driver than actual hardware.)

Anonymous said...

Service contract costs are exorbitant and getting worse - we were asked to hike a contract on one machine by 150%! You can guess what we said. The current rumor is that manufacturers are selling equipment with very slim profit margins in order to secure the contract and then tie you into six-figure (per annum) service contracts that are needed to keep the instruments running. Sustainable, it ain't.

Anonymous said...

I wonder if foundations and billionaires are the answer. These organizations are usually guided by a few individuals and follow narrow agendas. Like American sports, it is great for the "chosen ones", but not so much for the field. The vitality of the CM field (unlike particle physics expt or even AMO these days) is at least partly because nice work/discoveries could come out of anywhere, not just from a few big groups.

A side story to this is the quality of faculty in US universities. If the current rate of decline continues, and other countries continue to invest in science, the university faculty will decline (perhaps this is already happening in some fields). This has an interesting (but perhaps small) effect on the overall debate about higher education in America.

Anonymous said...

You mention that there is no good mechanism to deal with "recapitalization". Would allowing universities to charge for depreciation (either through overhead or by charging directly) be one way to deal with this? I'd imagine this would at least work for smaller/cheaper equipment that costs less than $300k or so (vector network analyzers, semiconductor parameters, etc).

Douglas Natelson said...

Anon@10:04 - good question, and somewhat complicated. In general, universities can charge depreciation. (For those that don't know, this means if the university buys a new gadget for $Y, over the next n years - typically 7 - they can try to collect, through user fees, $Y/n dollars per year, which would go into a "depreciation fund". The idea is that at the end of the n years, you have built back up enough money to buy a new gadget.) Where this gets tricky is, if the gadget was purchased with US federal funds (e.g., through a government grant or in the old days Congressional pork), then user fees paid from federal research funds cannot pay into depreciation of that gadget. In other words, if the feds bought you the original, they don't want to also be buying you the replacement through research funds. If the university bought some gadget with its own resources, then user fees from federal grants are allowed to pay into depreciation. The philosophical issue is then, to what extent should universities be financing equipment replacement through the research grants of their faculty, rather than through direct internal investment? (I must be getting old. I just wrote a whole comment about accounting rules.)

Unknown said...
This comment has been removed by the author.