The President’s budget request for FY 2007 has just been released and we’ll be dissecting it and providing our analysis as we get through it. But I wanted to post a quick snapshot of the Networking and Information Technology Research and Development program, the federal government’s multi-agency IT research and development effort, because 1) it’s the number of greatest concern to the computing research community and 2) it highlights the near-fruitlessness of trying to track the federal investment in IT R&D on a year-to-year basis.
For FY 07, the President is requesting $3.09 billion in aggregate funding for NITRD, an increase of $930 million over the FY 06 budget request — a huge increase. However, in that peculiar DC way, that’s only a 2 percent increase over FY 06 actual. That’s because the baseline budget has changed significantly since the Administration last calculated its IT R&D expenditures. The Department of Defense apparently discovered it was funding a lot more IT R&D than it previously thought — $851 million more in FY06 than it spent in FY05 an increase of more than 400 percent.
How can this happen? Well, each agency is responsible for determining what its own contribution to the NITRD program actually is. If the criteria that agency uses to determine whether a particular expense is IT R&D related or not changes, the department’s contribution can change dramatically. Does it mean that there actually will be $851 million more available to researchers in FY06 than there was in FY05? Not likely, but I’d sure like to get my hands on the spreadsheet used to produce that number. Perhaps we’ll get a better look when the NITRD coordinating office releases its annual “blue book” report for FY07.
Anyway, the good news is the NITRD program is slated for continued growth in FY07 (despite the widely fluctuating baseline numbers). Overall, the program will increase 2 percent, higher than any of the other government-wide “crosscutting” research programs (Nanotechnology will actually see a 2 percent decline, though that’s subject to some of the same odd DOD accounting changes; and the Climate Change program is flat). NSF would see a 12 percent increase in its NITRD funding, and DOE would see an increase of 23 percent.
Update: (2/7/06 9:45pm) – I really should just trash this entry and start over, but it seems somehow more appropriate to leave the big strikethrough section for posterity.
After consulting with Simon Szykman, who heads the National Coordinating Office for NITRD, I’ve got a little better information on what is actually going on with the widely fluctuating budget numbers in the NITRD cross-cut. I can’t say I completely understand all the reasons, but I at least have some sense of what’s going on. Apologies to Simon if I screw this explanation up. This is likely uninteresting to all but the most hard-core federal funding geeks, but to me, it’s a great lesson in how tricky it is to rely on aggregate funding totals for any insight into federal policy.
In the early ’90s the decision was made — for reasons I don’t yet know — to exclude a number of programs in DOD from being counted as part of what would become the NITRD “cross-cut.” In particular, IT R&D investments at the DOD service labs — Air Force Research Lab, Army Research Lab, Naval Research Lab — weren’t included in the “Defense” line and weren’t calculated as part of the overall NITRD program.
For the FY 07 budget, the White House Office of Management and Budget (the gatekeepers for the budget process in the executive branch) reviewed the program accounting and decided that the legacy way of reporting the NITRD cross-cut was no longer accurate. To describe the full breadth of the federal government’s NITRD investments, R&D spending in the DOD service labs had to be included. So OMB produced this chart — which ran in the original version of this post — and included it in the Analytical Perspectives (pdf) supplement to the FY 2007 Budget Request. (Though I added the first column, “FY 05 (est),” just for comparison’s sake.)

Obviously, we’ll have much more as we get a little more time to dive into the budget. Stay tuned…
Now, as we’ve figured out, this spread of numbers isn’t very useful for year-to-year comparisons. The White House Office of Science and Technology Policy realized this, too, so they now keep a second set of numbers which uses — roughly — the same set of agencies and programs that had been the norm until FY 07. Here’s the OSTP version:

You’ll note that when the military services are pulled from the numbers, DOD actually appears to receive an increase in the request versus FY06, which seems to indicate that the service labs don’t fare particularly well in the President’s budget. Szykman indicated that the DOD numbers for FY06 and FY07 in this chart also include, for the first time, funding from the DOD High Performance Computing Modernization Office. This isn’t new money.
Finally, you’ll note that some of the non-DOD numbers have changed in the second chart. According to Szykman, that’s likely due to further refinement as the chart worked its way through OMB to final release. Apparently the OSTP version is the “newer” version, and therefore its numbers are likely to be more accurate.
Presumably, we’ll have the final word when the NITRD NCO releases its FY07 Budget Supplement (the “Blue book,” which is now — of course — red) sometime in the next few weeks.
So what’s the take-away from all of this? I don’t really know, honestly. OSTP indicates that NITRD is up 9.4 percent in the President’s request over last year, but that includes additional funding in the calculation for FY07 that isn’t really new money. The OMB numbers indicate it’s more like 2.0 percent, but those numbers include a whole bunch of funding that’s apparently never been considered before.
Update: (2/8/06 8:39 am) – Ok, final update to this post. After some additional clarification from Szykman, it does appear that the OSTP-indicated increase of 9.4 percent is an accurate estimate of the status of the NITRD “legacy” programs OSTP is tracking. We’ll have further details in future posts about what exactly that 9.4 percent increase includes. But for now, maybe what’s most important for computing researchers is the knowledge that the traditional three big supporters for fundamental computing research — NSF, DOE and DOD/DARPA — all would see increases in the coming year under the President’s plan.
From OSTP:

High-end computing (HEC) continues to be a major focus of NITRD. DoE’s Office os Science (DoE SC), NSF and NASA are all engaged in developing and/or operating leadership class computing systems as recommended in the 2004 Federal Plan for High-End Computing, with the goal of deploying petascale computing systems by the year 2010. The DoE SC 2007 investment of $103M in leadership class computing, coupled with NSF’s investment of $50M in their Office of Cyber Infrastructure, will ensure that U.S. scientists and researchers have access to the most powerful computational resources in the world. Similarly, NASA continues to emphasize high-end computing within its NITRD portfolio through the operation of the Project Columbia supercomputer. All three agencies have pledged to make a portion of their leadership class computing systems available to other Federal users and the larger research community.
A 9% increase in support for advanced networking research in 2007, primarily by NSF, DARPA and DoE SC, will ensure that large-scale networking technologies will keep pace with the rapid development of petascale computing systems, so that the results of petascale computations are immediately accessible for analysis.
The 2007 Budget also includes significant increases in long-term fundamental research in cyber security and information assurance, as recommended by the President’s Information Technology Advisory Committee. Budget increases in cyber security and information assurance for NSF (+28%), DHS (+43%) and the National Institute of Standards and Technology (+11%) will support substantial new research activities to help secure the Nation’s information infrastructure, including fundamental research, and support for large-scale cyber security test beds and data sets.

We’ll have more on some of the agency-by-agency specifics as we get a chance to pour through the budget documents a little closer. As this episode points out, even a close reading of the documents isn’t always enough. And to think, if it’s this difficult to figure out the dollar amounts involved in this cross-agency program, imagine how difficult it is to coordinate research priorities and research activities….
Stay tuned.