Jonathan Haskel Imperial College Business School Imperial College London and CEPR UKIRC jhaskelicacuk National Academy of Sciences Washington DC 11 th July 2011 Slides plus spares links to papers etc in spares section ID: 318594
Download Presentation The PPT/PDF document "The fork in the road for innovation meas..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
The fork in the road for innovation measurement: which way should we go?
Jonathan HaskelImperial College Business School, Imperial College London and CEPR, UKIRC j.haskel@ic.ac.ukNational Academy of Sciences, Washington DC, 11th July 2011Slides plus spares, links to papers etc. in spares section
1Slide2
The exam questions(to be answered in final slide…)
The objective of this session is to identify recent developments in measuring STI and what is currently planned for the future. Discussion should reveal what has been successfully and unsuccessfully measured. What are critical bottlenecks and perceived opportunities? What global STI metrics and indicators should NCSES develop in the near and medium term (the next 5-10 years)?2Slide3
The fork in the road for innovation measurement
Approach 1: EU-type innovation surveysTried for 15+ years in EU, Applied in latest US workKey questions“have you process/product innovated?”Spending on innovation e.g. software, marketing, trainingPotpourri of other questionsInformation for innovation (e.g. importance of information from universities high/med/low)What stopped you from innovating?Approach 2: extended R&D/spending type surveysFollows Griliches/Jorgenson/CHS/Nat Accounts to measure knowledge capitalImplementation: software and (forthcoming) R&D spendCHS agenda: need coinvestments with R&D e.g. design, training, marketing, business processCurrent method: potpourri of surveysUK: intangibles investment survey, similar style to R&D surveyKauffman
3Slide4
What theory do we have to guide measurement?
Innovation as TFP i.e. output net of tangible K and labourLogically consistent, suggests measurement (e.g. depreciation, investment)Appropriate if all knowledge from free spilloversCorrado, Hulten, Sichel Recognises innovation is routinised in many firms e.g. organised spending on marketing, training , design etc. Helps understand innovation in non-R&D industries e.g. financeStrength: CHS categories are matched to functional bodies in (large) firms e.g. HR dept for training , marketing dept for brand etc. Still scope for knowledge spillovers Measurement implicationsSpending, investment, depreciation, asset and rental pricesSlide5
European Innovation survey questionnaires
UK example: product innovation 5Process innovation Slide6
6
US data: Boroush (2010), Other: OECD Measuring innovation (2010), p.26Michael Mandel: “You can’t be an innovative economy if only 9% of your companies are innovating”What do innovation surveys find?Slide7
What does the “have you innovated” question measure?
Other findingsthe total is procyclical: rises strongly in IT boom end of 1990s Item non-response can be very high e.g. 70% to intangible spending questionsLarge firms show poor response ratesSo what does the question measure? Crespi , Criscuolo, Haskel (2007, Appendix 1), UK CIS3 (98-00)45% of firms reported technological product/processthey were asked for text describing their innovationThey are given instructions on what to include and exclude e.g. include “robotised welding”, exclude “packaging” Analysis of 874 firms showed7Slide8
The “have you process innovated “ question measures new capital equipment
Summary: of firms who provided the information, reported process innovation is mostly new capital spendingThis is why measured “innovation” is highest in IT boomConjecture: we would get same/better(?) information from a standard investment survey (+ good deflators)8Slide9
An alternative approach guided by CHS
What does this framework suggest you need to know?Knowledge services within the firm = knowledge capital stock and rental ratesInnovation spending on broad range of knowledge assets: software, design etc. Convert to investmentConvert to real stock: need depreciation and deflatorsImplication: need a spending survey. Can piggy back onto R&D survey or new one (or modify innovation survey)Knowledge from outside the firmInformation flowsPatents is one source: but service sector?CIS information flowsSlide10
Possible model: Investment in Intangible Assets Survey
(www.ceriba.org.uk/bin/view/CERIBA/InvestIntangAssetsSurvey))Extension of UK R&D survey, Autumn 2009, through ONS: so linked to national accounts data and R&D and CIS surveySurvey follows (for large firms) typical business functions10Slide11
Each section has a filter question which defines the asset with examples
11Slide12
Then asks purchased and own-account
12Slide13
Finally life lengths
13Slide14
14
Life lengthsAll asset life lengths are greater than one year and range between 2.75 years for training and reputation to around 4.5 years for R&D.Slide15
My answers to the exam questions
The objective of this session is to identify recent developments in measuring STI US innovation survey raises puzzles for survey design. Intangible assets framework gaining policy traction. Some surveys moving in this direction. what is currently planned for the futureReview by OECD of innovation surveys. More UK intangible surveys. Kauffman survey to continue?Discussion should reveal what has been successfully and unsuccessfully measured. yes/no innovation questions have not worked. IIntangible spending can work if asked to business functions, own-account explained.Business process has not worked in UK survey What are critical bottlenecks and perceived opportunities?Shortages of funding for more surveys. Opportunity (a) in EU, convert Innovation surveys to intangible spending surveys, (b) in US to convert supplement to BERDIS to intangible spending survey.General lesson: take best practice from each survey What global STI metrics and indicators should NCSES develop in the near and medium term (the next 5-10 years)?Intangible spending, life lengths. Integrate with National Accounts work on deflators, service sector output.
15Slide16
Spares
16Slide17
What do innovation surveys find?
17US data: Boroush (2010), Other: OECD Measuring innovation (2010), p.26Slide18
Crespi et al, 2008. What do firms write down as their most significant process innovation?
Replies 53% are other, mostly including capital equipment (e.g. steel wire machine, spectrophotometer..)23% are hardware related (computer, CAD, PC) and software related (website, email, software, internet etc.)Summary: of firms who provided the information, reported process innovation it new capital spendingNot surprising that measured “innovation” is highest in IT boomWe likely get equivalent information from a standard investment survey (and spend resources on better deflators)18Slide19
The future
Take best practice from elements of different surveysInnovation surveysDon’t ask % new, % innovationDo ask information sourcesR&D/IASDo ask for spending Keep structure of questionaire allied with business function: ICT dept for software, HR dept for training, Marketing dept for brandingComplement withWhat do firms do? Innovation data at project level ?What do young firms do? Likely very intangible intensive. Kauffman survey. What do consultants do, especially intangible consultants?What do financial institutions do esp young firms, they are trying to value innovation?Measuring business process huge challengeChance for NAS/OECD to lead as OECD did on software