Federal Support for Networking and IT R&D: Much Less than Meets the Eye

Good news:  According to the federal budget,  about $4 billion per year goes for networking and information technology (NIT) R&D. (see here)

Bad news: “The Nation is actually investing far less in NIT R&D than is shown in the Federal budget.

That’s according to an authoritative new report, Designing a Digital Future: Federally-Funded Research and Development in Networking and Information Technology, from the President’s Council of Advisors on Science and Technology. Here’s what they have to say: 

The aggregate NITRD crosscut budget significantly overstates the actual Federal investment in NIT R&D. By leading policymakers to believe that we are spending much more on such activities than is actually the case, this discrepancy contributes to a substantial, systematic underinvestment in an area that is critical to our national and economic security.

And then they go on to say.

Most obviously, a large portion of the “High End Computing Infrastructure and Applications” budget category, which accounts for roughly $1.5 billion of the $4.3 billion NITRD total, is attributable to computational infrastructure used to conduct R&D in other fields, and not to NIT R&D or to infrastructure for NIT R&D.

Beyond this, however, various agencies include in their reports for other NITRD budget categories investments in NIT that support R&D in non-NIT fields. The laudable transparency of NIH’s NITRD grant reporting allowed an expert in NIT at the Science and Technology Policy Institute (STPI) to review the abstracts of the top 100 awards (by award size) in NIH’s 2009 NITRD crosscut for actual NIT R&D content. STPI developed a 14-part coding scheme to categorize the 100 projects. The analysis showed that of the 95 projects with available abstracts, only 4 appear to be making a contribution to actual NIT R&D, such as novel computer science methods, novel simulation methods, or novel system design and computing methods. Another 14 projects could be considered “borderline” although they seem to focus on NIT development alone. The remaining 77 of the 95 projects with available abstracts appear to have no NIT R&D content. (For example, they may involve NIT infrastructure to support biomedical or biochemical R&D, but with no novel NIT R&D contribution.)

<short snip>

In short, STPI concluded that 86% of these awards, by dollar value, have no NIT R&D content; 3% could not be assessed, 9% were judged borderline, and 2% were judged to be making a contribution to actual NIT R&D. Although other agencies do not report NIT R&D spending in sufficient detail to make the same analysis possible, it seems likely that in many cases a similar confusion in classification of NITRD investment occurs.

These findings highlight both the increasing ubiquity of NIT infrastructure for conducting R&D in many fields and the difference between this infrastructure and actual NIT R&D – work that makes a novel contribution to NIT.

I draw two conclusions from this.  First, the U.S. has made and continues to make an enormous bet on biosciences as the  driving force for future economic growth.  That’s shown by our pattern of federal research support.

Second,  if we want to maintain our lead in networking and information technology research, we need to put more federal dollars there. Now.


  1. On a related note, it has been convincingly claimed that many tech companies charge product deployment and product support activities to R&D. OTOH when there is no formal distinction between design and implementation, and (e.g. in the case of software) the software is still hacked while being deployed in “Beta”, and there is no clear separation between development and support work which are all done by domain experts in “multiple hat” roles, it’s difficult to draw the line. But the causation may be the other way around – keeping things fuzzy so you have more leeway in accounting.

    And as for (N)IT support in other fields – every field has special requirements and its special idiosyncrasies, and you cannot use off the shelf systems and cookie cutter approaches. Surely the accounting and other corporate IT systems in biotech will be the same as everywhere, but not the lab tech. You will be able to use stock hardware in your biotech equipment, but the software will be custom, and a lot of the hardware too, as you will be solving fundamentally different problems than in other fields and under different constraints and operating environments.


  1. […] This post was mentioned on Twitter by Toby Elwin and Daniel Altman. Daniel Altman said: RT @MichaelMandel: Federal support for IT research less than expected http://wp.me/pIGEy-gQ […]

  2. […] Federal Support for Networking and IT R&D: Much Less than Meets the Eye (innovationandgrowth.wordpress.com) […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s


%d bloggers like this: