Department of Defense

Re: US government accounting for R&D as share- ofeconomic growth

Submitted by Administrator on

Sent by: "Merrill, Steve" Steve:The Academies recently completed an extensive evaluation of the fiv=e largest SBIR programs -- DOD, DOE, NSF, NIH, and NASA -- as well as a cro=ss-agency review with recommendations. All are accessible at www.nap.edu.This work continues on selected issues of impact and administration.Steve MerrillDirector, STEP =20-----Original Message-----From: Fiore, Steve [mailto:sfiore@ist.ucf.edu]=20Sent: Tuesday, November 02, 2010 9:55 AMTo: Science of Science & Innovation PolicySubject: [scisip] Re: US government accounting for R&D as share ofeconomic =growthSciSIP:On a related note, does anyone know of any analyses of the federal governme=nt's SBIR/STTR programs?At a minimum, I'd be curious to see what percenta=ge has successfully transitioned to actual products (and by industry) but, =ideally, whether or not there have been any analyses of broader impacts fro=m this investment in small business innovation.Thanks,Steve Fiore________________________________________From: Regets, Mark [MRegrets@nas.edu]Sent: Tuesday, November 02, 2010 9:16 AMTo: Science of Science & Innovation PolicySubject: [scisip] Re: US government accounting for R&D as share ofeconomic =growthSciSIP:Thanks Carol,I think we are saying the same thing. In no way do I want to imply that t=he R&E satellite accounts are not a good thing to do.The macroeconomic =importance is huge. And R&D measurement and better understanding of what =is measured is obviously important to attempts to measure overall societal =returns to R&D.Mark________________________________________From: Robbins, Carol [Carol.Robbins@bea.gov]Sent: Tuesday, November 02, 2010 8:36 AMTo: Science of Science & Innovation PolicySubject: [scisip] Re: US government accounting for R&D as share of economic= growthSciSIP:SciSIP,Mark's point about the information needed about R&D for science policymakes sense.There is however, substantial value to accuratemeasurement of direct intangible investment in a national accountingframework.BEA's work will lead to measures of investment and R&D stockthat align with and impact internally consistent statistics of national,industry, and state level Gross Domestic Product. That said, the issue of spillovers does require a different kind ofanalysis, and the measure of depreciation that reflect these spilloversor externalities would indeed be different.BEA's accounts measure thedepreciation from the perspective of the owner of the R&D, thus itrepresents obsolescence and the loss of the owner's ability to capturethe economic benefits.In BEA's 2007 Survey of Current Business article on the SatelliteAccount, the topic was addressed this way:"The estimates provided in this release include only the direct impactof R&D investment, that is, the direct benefit realized by the investor.These estimates do not separately identify spillovers, the benefits ofR&D to firms that did not pay for the R&D. However, the Bureau of LaborStatistics (BLS) produces measures of the impact of technological changeon productivity as part of its estimates of multifactor productivity forthe business sector. These estimates measure spillovers directly. BLShas estimated that approximately one-fifth of the multifactorproductivity residual can be attributed to R&D in recent years. TheseBLS estimatesof the spillovers are broadly consistent with the BEA estimates of thedirect impact of R&D. For more information, see."A set of papers at BEA on R&D can be found here:http://www.bea.gov/national/rd.htmCarol RobbinsBureau of Economic Analysis202-606-9923-----Original Message-----From: Regets, Mark [mailto:MRegrets@nas.edu]Sent: Monday, November 01, 2010 2:23 PMTo: Science of Science & Innovation PolicyCc: scisip@lists.nsf.govSubject: [scisip] Re: US government accounting for R&D as share ofeconomic growthSciSIP:I have to agree with Gordon, but perhaps make it stronger.WhetherR&D is counted like investment in the National Accounts will make adifference to the measured growth rate of GDP, and hence tomacroeconomic policy. But it in no ways tell us anything useful aboutthe contribution of R&D to GDP except in the most trivial sense of theaccounting relationship (GDP=3D C + I + G + X).This is of interestto macroeconomic policy, but the main relevance of this to sciencepolicy may be that the added attention to R&D may lead tobetterestimates of it (a important matter to be sure).Mark Regetsmregets@nas.edupermanent email: markregets@scipolicy.com________________________________________From: Caroline Wagner [cswagner@mac.com]Sent: Monday, November 01, 2010 12:01 PMTo: Science of Science & Innovation PolicyCc: scisip@lists.nsf.govSubject: [scisip] Re: US government accounting for R&D as share ofeconomic growthSciSIP: Hi Gordon and colleagues,Thanks for these excellent points.I am attaching the original BEAreport, 2007, on how the Dept of Commerce is calculating R&Dcontributions to the GDP satellite account. This may have more specificinformation in it that addresses the issues you note.Caroline.On Nov 1, 2010, at 11:51 AM, Gordon Reikard wrote:The paper in the previous link is Dennis Fixler, Accounting for R&D inthe National Accounts, presented at the ASSA meetings in January 2009.Unfortunately, it does not do what it purports to do, and leaves thequestion of the contribution of R&D largely unresolved. Reading thispaper, it seems to wander ambiguously through some of the productionfunction literature, omitting many of the key references, withoutreaching any definite conclusions.There are really two issues in including for R&D in the national incomeaccounts. The first is whether to include private sector R&D in GDP,under business fixed investment. At the current time, private sectorR&D is treated as an intermediate input, and excluded. Paradoxically,Federal government spending is included in GDP, under governmentpurchases. If private sector R&D were included, business investmentwould be higher, and as a result, GDP would also be revised upwardhistorically.The second if of course the much more difficult issue of including R&Din the production function. This paper does not estimate a productionfunction, despite the fact that there are studies going back more than20 years that do. Instead, it merely reports the fact that BEA and BLShave come up with widely different estimates of the R&D stock.According to this paper, BEA estimates that in 2002 the total R&D stockwas only $931 billion, of which $581 billion was comprised by privateindustry. Presumably, this estimate is in nominal dollars. The BLSestimate for the private industry R&D stock for the same year was $1,295billion.Nevertheless, both estimates can be seen as implausibly low. BEAassumes very high rates of depreciation. This paper is rather vague onexactly how this is calculated, but other BEA documents indicate that 15percent geometric was used. This assumption is at best completelyarbitrary, and ignores much of the literature (see in particular BronwynHall's recent paper on measuring depreciation).BLS assumes nodepreciation of basic research, but 10 percent annual depreciation ofdevelopment and applied research. By comparison, if R&D is assumed notto depreciate, the corresponding values would be a stock of $2,649billion in industry-funded R&D in 2002, and $2,095 funded by othersources.The reason that these estimates are important of course is that the sizeof the R&D stock influences the elasticity of R&D in the productionfunction, and in turn, the contribution of R&D to output. Given howlow the BEA estimates of the R&D stock are, it is perhaps fortunate thatFixler's paper did not estimate a production function. The resultingestimates for the impact of R&D would be far below what has beenpublished in the academic literature.________________________________http://www.bea.gov/papers/pdf/Accounting%20for%20RD_02_09.pdfOn Oct 31, 2010, at 1:19 PM, Ervin, Dana Miller wrote:SciSIP:Can anyone refer me to a recent study (post Solow) that shows thecontribution of growth in knowledge to US output?ThanksDana Ervin---You are currently subscribed to scisip as: MRegets@nas.eduTo unsubscribe send a blank email to $subst('Email.Unsub')To send emails to the listserv, email scisip@lists.nsf.govTo subscribe to the listserv, send a blank email toSUBSCRIBE-SCISIP@LISTS.NSF.GOV---You are currently subscribed to scisip as: Carol.Robbins@bea.govTo unsubscribe send a blank email to $subst('Email.Unsub')To send emails to the listserv, email scisip@lists.nsf.govTo subscribe to the listserv, send a blank email toSUBSCRIBE-SCISIP@LISTS.NSF.GOV---You are currently subscribed to scisip as: MRegets@nas.eduTo unsubscribe send a blank email to $subst('Email.Unsub')To send emails to the listserv, email scisip@lists.nsf.govTo subscribe to the listserv, send a blank email to SUBSCRIBE-SCISIP@LISTS.=NSF.GOV---You are currently subscribed to scisip as: sfiore@ist.ucf.eduTo unsubscribe send a blank email to $subst('Email.Unsub')To send emails to the listserv, email scisip@lists.nsf.govTo subscribe to the listserv, send a blank email to SUBSCRIBE-SCISIP@LISTS.=NSF.GOV---You are currently subscribed to scisip as: SMerrill@nas.eduTo unsubscribe send a blank email to $subst('Email.Unsub')To send emails to the listserv, email scisip@lists.nsf.govTo subscribe to the listserv, send a blank email to SUBSCRIBE-SCISIP@LISTS.=NSF.GOV

Pages

Subscribe to RSS - Department of Defense