1 toward a standard benefit-cost methodology for publicly-funded s&t programs crossing borders,...
TRANSCRIPT
1
Toward a standard benefit-cost Toward a standard benefit-cost methodology for publicly-funded S&T methodology for publicly-funded S&T
programsprograms
Crossing Borders, Crossing Crossing Borders, Crossing BoundariesBoundaries
2005 Joint CES/AEA Conference2005 Joint CES/AEA ConferenceToronto, CanadaToronto, Canada
October 26-29, 2005October 26-29, 2005
Jeanne PowellJeanne PowellEconomic Assessment Office
Advanced Technology ProgramNational Institute of Standards and Technology
2National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
OutlineOutline
BackgroundAEA 2005: Analysis of Differences and
Considering an Approach to Comparability
Moving Forward
3National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
BackgroundBackground
ATP Mission and Project Selection Criteria
Evaluation FrameworkRole and Types of Benefit-Cost AnalysisPerformance MetricsBenefit-Cost Studies To DateStudy Similarities
4National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
ATP Mission …ATP Mission …To accelerate the development of
innovative technologies for broad
national benefit through
partnerships with the private sector.
ChemistryChemistry
BiotechnologyBiotechnology
ElectronicsElectronics
ManufacturingManufacturing
AdvancedAdvancedMaterialsMaterials
PhotonicsPhotonics
InformationInformationTechnologyTechnology
5National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
Scientific and Technological Merit (50%)– Technical innovation– High technical risk with evidence of
feasibility– Detailed technical plan
Potential for Broad-Based Economic Benefits (50%)– National economic benefits– Need for ATP funding– Pathway to economic benefits
ATP Selection CriteriaATP Selection Criteria
6National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
What We Measure WhenWhat We Measure WhenE
CO
NO
MIC
IM
PA
CT
SE
CO
NO
MIC
IM
PA
CT
S
Short-TermShort-Term Mid-TermMid-Term Longer-TermLonger-Term
1010or Moreor MoreYearsYears
-1 0 1 2 3 4 5 6 7 8 -1 0 1 2 3 4 5 6 7 8 99AnnounceAnnounce
Compe-Compe-titiontition
AnnounceAnnounceAwardAward
Complete Project
Post-Project Period
TotalEconomic Benefits
Benefits toAwardees
(Inputs/Outputs) Award/participant
characteristics Project funding R&D partnering Acceleration of R&D Innovative technology
development– Patents– Publications– Competitive advantage– Prototype products &
processes
(Outputs/Outcomes) Commercial activity
– New products– New processes– Licensing
Attraction of capital Strategic alliances Company growth
(Impacts) Broad national economic
benefits– Return on investment
– Public– Private– Social
– Inter-industry diffusion– Increased GDP/ tax base– Societal impacts
7National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
Types of B-C Case Studies in ATPTypes of B-C Case Studies in ATP
Traditional cash-flow based, return on investment measures– Benefit-to-cost ratios– Internal rates of return– Net present value
Hedonic price index models Macroeconomic models
8National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
Benefit-cost Analysis:Benefit-cost Analysis:Return on Investment MeasuresReturn on Investment Measures
Private return– return to ATP project participants on their
own investment
Public return– return to nation on ATP investment
Social return– return to ATP project participants and
nation on total ATP and private
investment
9National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
Performance MetricsPerformance Metrics
Return on Investment MetricsBenefit-to-Cost ratio is computed by
dividing the present value of benefits by the present value of investments
Net present value (NPV) is the present value of benefits minus the present value of investments
Internal rate of return is calculated by iterative solution for a rate at which the discounted value of investments equals the discounted value of benefits
10National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
ATP’s B-C Studies To DateATP’s B-C Studies To Date
12 studies published, including quantitative case studies of 28 projects
2 studies underway, including quantitative case studies of 4 projects
All studies performed by independent contractors
11National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
ATP’s B-C Studies To DateATP’s B-C Studies To DateShort Title (# of Projects)
Number Contractor Pub Date
Study Cost-$
Photonics Cluster (2) GCR 05-879
Delta Research
2005 228,000
Composites Cluster (2) GCR 04-863
Delta Research
2004 200,000
2mm Project—Retrospective (1)
GCR 03-856
MIT 2004 357,000
HDTV Joint Venture (1) GCR 03-859
RTI International
2004 75,000
A-Si Detectors for Digital Mammography (1)
GCR 03-844
Delta Research
2003 136,000
Component-Based Software (8)
GCR 02-834
RTI International
2002 154,000
Closed-Cycle Refrigeration (1)
GCR 01-819
Delta Research
2001 77,000
Digital Data Storage (2) GCR 00-790
RFF 2000 138,000
Flow-Control Machining in Auto Industry (1)
NISTIR 6373
NIST-Building & Fire Research
1999 90,000
Tissue Engineering (7) GCR 97-737
RTI International
1998 122,000
2mm—Early Assessment (1)
GCR 97-709
CONSAD 1997 25,000
Printed Wiring Board (1) GCR 97-722
Albert N. Link
1997 22,500
12National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
Study SimilaritiesStudy Similarities
Address ATP Mission, emphasizing public benefits to industry customers and end users and need for ATP funding to achieve these benefits
Cash-flow based approaches use:– Good practices consistent with public
finance literature; NPV, B:C, IRR– OMB-mandated 7% real discount rate (all
but 1 study) and constant $ estimates– Year of ATP project start for base year– Relatively short study periods
Results are not presented as “representative”, but rather as portfolio minimums
13National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
AEA 2005AEA 2005
Analyzing Differences and Considering an Approach to Study Comparability
14National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
AEA 2005: Analyzing Differences and AEA 2005: Analyzing Differences and Considering an Approach to Study Considering an Approach to Study
Comparability Comparability
Analyzing four major differences :
1) Study timing relative to ATP funding and uncertainties about future outcomes
2) Identifying the counterfactual and attribution of benefits
3) Which metrics? Public versus social return on investment
4) Adjusting for timing differences across studies and projects
15National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
AEA 2005: Analyzing Differences and AEA 2005: Analyzing Differences and Considering an Approach to StudyConsidering an Approach to Study
Comparability Comparability
Considering a standard approach based on three criteria:1) Meeting evaluation objective
2) Quality and accuracy
3) Comparability
16National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
1. Study Timing and Uncertainties 1. Study Timing and Uncertainties About Future OutcomesAbout Future OutcomesIssues:
Uncertainties concerning technology benefits are reduced over course of innovation process– At time of ATP funding, risks of technical failure (or
of meeting only limited technical objectives) are very high.
– Market/financial/business risks remain high during and after ATP funding phases of technology development and into commercialization phases
– Innovation life-cycle/timing varies by technology area Earliest studies conducted early in innovation
process – No projects had matured– Direct experience with range of possible outcomes
(and empirical data about innovation process) was limited
17National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
1. Study Timing and Uncertainties About 1. Study Timing and Uncertainties About Future Outcomes: Contrasting Studies Future Outcomes: Contrasting Studies
Study
Timing Relative to
ATP Funding
Technology Status
Treatment of Uncertainty
Tissue Engineering(7 cases)
Early in ATP for 3 cases; soon after ATP for 4 cases
Technologies not developed and/or not ready for commercialization(Biotech projects have very long timeline)
Barriers to meeting technical and economic goals not assessed. Sensitivity analysis performed but showed results not very sensitive to input variables tested. Considered only one application per case in order to be conservative
Component-Based Software(8 cases)
Soon after ATP
Technologies complete; in commercialization(IT projects have short timeline)
Selection emphasized projects with revenues to date and near-term prospects.Benefit analysis quantified only products actually on the market and assumed short product lives.
18National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
1. Study Timing and Uncertainties About 1. Study Timing and Uncertainties About Future Outcomes: Contrasting Studies Future Outcomes: Contrasting Studies
Notes: Tissue Engineering: 3 technologies made
considerable technical progress and commercialization continues to evolve but much more slowly than anticipated; 3 technologies failed to develop and companies dissolved; 1 is in transition to new company, future as yet is uncertain
Component-Based Software: focus on existing products and short-term product life reduced the amount of research required per project and enabled study of more projects
19National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
1. Study Timing and Uncertainties 1. Study Timing and Uncertainties About Future Outcomes: ConclusionsAbout Future Outcomes: Conclusions
Under conditions of considerable technical/business uncertainty about future outcomes:
Rigorous, complex economic models may or may not generate useful performance metrics.
Probability-based outcome measures that consider the broad technical, market, and business risks, as well as technology benefits, are likely needed
Metrics are more reliable/capable of precision as projects overcome technical, market, and financial barriers
20National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
2. Identifying the Counterfactual and 2. Identifying the Counterfactual and Attribution of BenefitsAttribution of Benefits
Issue:Where does project being analyzed start and
end? And what is the counterfactual, or baseline?– In theory, investment costs and benefits
included in analysis need to measure increment over the counterfactual (what would have happened without the project being measured).
– In practice, benefits (e.g., profit from sale of a product embodying new technology) derive from multiple public and private investments and from one or two of many product generations.
21National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
2. Identifying the Counterfactual and2. Identifying the Counterfactual and Attribution of Benefits—Contrasting Studies Attribution of Benefits—Contrasting Studies
Study Approach
Tissue Engineering
ATP project benefits are defined in terms of acceleration of benefits and increased probability of technical success relative to a hypothetical project without ATP. Public plus private benefits are compared to public plus private investments to compute social return on social investment.
Component-Based Software
Identified “project” that resulted in identifiable product sales and compared public/private benefits with all costs associated with that project to compute social returns
A-Si Detectors for Digital Mammography and Composites
Company interviews indicated “no project without ATP”; public benefits from product sales are compared to ATP investment to compute public return on ATP investment
Photonics Technologies
50% attribution to ATP where two different federal programs providing similar levels of funding were deemed jointly responsible for realization of any benefits.
22National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
2. Identifying the Counterfactual and 2. Identifying the Counterfactual and Attribution of Benefits—Conclusions Attribution of Benefits—Conclusions
Researchers use a variety of mechanisms to model benefits relative to a counterfactual (e.g., increased probability of achieving benefits, acceleration, allocation in accordance with amount or importance of funding sources)
ATP-funded technologies are often embodied in multiple products, each of which resulted from a number of different public/private funding sources and “projects”. Researchers measure benefits of products with strongest ATP linkages and share attribution where needed
Social rate of return has weaker assumption about “causation” than public return on ATP $ does.
23National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
3. Which metrics? 3. Which metrics?
Social return metrics– Economics literature compares social rate of return to
private rate of return (to see economic spillover) and then further compares private rate of return to private hurdle rate to identify market failure that justifies public funding
– Cost-shared projects imply comparison of benefits to both public and private recipients with investments from all sources
AND/ORReturn on ATP investment
– Directly addresses: What is the ATP program impact?– Involves substantially less historical/company proprietary
information than social return metrics
24National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
3. Which metrics? 3. Which metrics? Contrasting StudiesContrasting Studies
Issue:Some studies/contractors emphasize
social return metrics; others emphasize public return
25National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
3. Which metrics? 3. Which metrics? Contrasting StudiesContrasting Studies
Social return metrics Public return metrics(NC = Not computed)
NPV ($M)
B:C IRR (%)NPV ($M)
B:C IRR(%)
Component-based Software 1 -1.2 .37 N/M NC NC NC
2 789 39 363 NC NC NC
3 -1.22 .63 N/M NC NC NC
4 29.6 9.6 103 NC NC NC
5 2.06 1.8 31 NC NC NC
6 21 7.6 51 NC NC NC
7 52 18 136 NC NC NC
8 1.2 1.2 13 NC NC NC
26National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
3. Which metrics? 3. Which metrics? Contrasting StudiesContrasting Studies
Social return metrics (NC = Not Computed)
Public return metrics(NC = Not Computed)
NPV ($M)
B:C IRR (%) NPV ($M)
B:C IRR
Digital Video (Midpoint)
165.9 4.24 28.6 NC NC NC
A-Si Detector(base case)
NC NC NC 219 125 69
Composites(base cases)1
NC NC NC
552 221 57
2 NC NC NC 510 187 58
Photonics(base cases)1
NC NC 43 184 75 49
2 NC NC 51 201 244 350
27National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
3. Which metrics? Social versus 3. Which metrics? Social versus Public Returns Public Returns——ConclusionsConclusions
Both have merit for different purposes, and both have advantages and disadvantages– Social return metrics reflect cost-sharing
aspect of ATP and other public-private partnership projects; however, they require company-proprietary data and historical data about multiple funding sources (generally only rough estimate is feasible) .
– Public return on ATP investment provides measure of direct program impact; however, it requires judgments about of “causation”
28National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
4. Adjusting for Timing Differences4. Adjusting for Timing Differences Across Studies and Projects: Across Studies and Projects:
Constant $ and Base Year Constant $ and Base Year Issue:Different case studies based cash flow
estimates on constant dollars of different years
Although nearly all studies used the same discount rate (7%-real), they discounted cash flow estimates to different points in time (start of each project). (This practice evolved in financial world’s prospective analysis of potential investments and is used in Excel.)
Thus, reported NPV metrics are not comparable across projects.
29National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
4. Adjusting for Timing Differences 4. Adjusting for Timing Differences Across Studies and Projects: Constant $ Across Studies and Projects: Constant $
and Base Year and Base Year EExample 1: Ranking does not changexample 1: Ranking does not change
Component-Based
Software 8 Cases
Reported NPV ($M)
Constant $ Year
Base Year for
Discounting
Adj. NPV--Consta
nt 2005$ ($M)
Adj. NPV--
2005 as Base Year ($M)
Aesthetic Solutions
-1.2 (7)
2000 1995 -1.3 -2.6 (7)
Commerce One JV
789 (1)
2000 1997 880 1,513 (1)
Extempo -1.22 (8)
2000 1995 1.36 -2.68 (8)
Intermetrics 29.6 (3)
2000 1998 33.0 53.0 (3)
Real-Time Innovations
2.06 (5)
2000 1996 2.30 4.23 (5)
SciComp 21 (4) 2000 1995 23 46 (4)
Tom Sawyer 52 (2) 2000 1996 58 107 (2)
Xerox PARC 1.2 (6) 2000 1995 1.3 2.6 (6)
30National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
4. Adjusting for Timing Differences 4. Adjusting for Timing Differences Across Studies and Projects: Constant $ Across Studies and Projects: Constant $
and Base Year and Base Year Example 2 : Ranking ChangesExample 2 : Ranking Changes
Other Cash-Flow –Based
Cases
Reported NPV ($M)
Constant $ Year
Base Year for
Discounting
Adj. NPV--
Constant 2005$
($M)
Adj. NPV--
2005 as Base Year ($M)
Digital Video 165.9 (6)
2002 1995 177.7 349.5 (5)
A-Si Detector 219 (3) 2002 1998 236 374 (4)
Composites:Applied SciencesLincoln Composites
552 (1)510 (2)
20032003
19971998
579535
996 (1)
860 (2)
Photonics:X-Ray OpticsIon Optics
184 (5)201 (4)
20042004
19942000
188206
396 (3)
288 (6)
31National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
Adjusting to Constant 2005$Adjusting to Constant 2005$
NPV in Constant 2005$ =
Where: IPDGDP is Implicit Price Deflator for GDPand IPDGDPstudy is IPDGDP for Constant $ year used in study
1 + Q2 2005 IPDGDP – IPDGDPstudy
IDPGDPstudyx Reported NPV
32National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
Adjusting Base Year to 2005Adjusting Base Year to 2005
Adjusted NPV for Base Year 2005 =NPV in Constant 2005$ x 1.07n
Where: n = 2005 minus base year used in study
33National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
4. Adjusting for Timing Differences 4. Adjusting for Timing Differences Across Studies and Projects: Across Studies and Projects:
Constant $ and Base Year ExampleConstant $ and Base Year Example
Note:Rankings “before and after”
adjustment are for illustrative purposes only.– NPVs indicate net benefit magnitudes– NPVs do NOT indicate investment
efficiency
34National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
4. Adjusting for Timing Differences 4. Adjusting for Timing Differences Across Studies and Projects: Constant $ Across Studies and Projects: Constant $
and Base Yearand Base YearConclusions: Both computed value of NPV and rank order can
change by adjustment to constant $ and common base year
NPV adjustments to constant $ and common base year are easy to make—directly from reported NPV
Comparability and aggregation of NPV results indicate need for such adjustments although practitioners may differ on best approach
IRR and B:C do not require adjustment; however, they cannot be aggregated
35National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
Moving ForwardMoving Forward
Work with contractors/researchers toward greater understanding of need for comparability of B-C results while seeking improved methodologies
Recognize and build on strengths of individual contractors/researchers ; help them overcome inconsistencies with comparability goals
Improve understanding of B-C methodologies among S&T evaluation community
Work with NIST experts in both B-C analysis and standards development
36National Institute of Standards and Technology • Technology Administration • U.S. Department of Commerce
Call Jeanne Powell: 301-975-4196 Send e-mail to: [email protected]
Visit our website: www.atp.nist.gov
View our publications: www.atp.nist.gov/eao/eao_pubs.htm
For further information:For further information: