THE FUTURE OF THE SNAS ASSET BOUNDARY BY
THE FUTURE OF THE SNA’S ASSET BOUNDARY BY BRENT MOULTON AND NICOLE MAYERHAUSER TECHNOLOGY ACCOUNTS IN THE NATIONAL ACCOUNTS BY WENDY LI Discussion by Paul Schreyer (STD) OECD IARIW Conference 16 April 2015
Overview • Brent Moulton and Nicole Mayerhausen deal with possible extensions of SNA asset boundary – – Knowledge assets Environmental assets and natural resources Consumer durables Human capital • Under Knowledge assets I will bring in Wendy Li’s paper on technology accounts • Sprinkle comments into presentation of key points in the papers 2
Moulton-Mayerhauser: Knowledge Assets • SNA 2008 recognises intellectual property products: R&D, mineral exploration, software and databases, entertainment, literary and artistic orginals • Economics literature (e. g. Hulten, Sichel, Corrado 2009) suggests in addition brand equity, firm-specific resources, advertising – these are factors that shape competitiveness • Should they be part of the SNA? M-M point to: – Conceptual issues: e. g. brand equity affects demand but not production function – Empirical issues: valuation of stocks, measurement of investment, deflation and depreciation are all very difficult • M-M still judge that at least some additional knowledge assets should have a good prospect of making it inside the SNA asset boundary (but don’t say exactly which ones) 3
Knowledge Assets: comments • Relevance of knowledge assets for analysis • So usefulness of developing measures is clear • But measurement difficulties significant – so probably too early for core accounts • But also issue of how to use information: are services from knowledge assets different from other assets? – Affect workings, organisation of a firm as a whole: shape of rather than movement along a production frontier – Affect the demand function (including the capacity to reap monopoly rents) – No wear and tear, so depreciation = technical obsolescence, change in consumer preferences and loss of monopoly rights – Huang and Diewert: Can. JEconomics 2011 • Consequence: treat knowledge assets as determinants of overall MFP rather than as particular input 4
Good point to introduce Wendy Li’s paper on Technology Accounts in the National Accounts • Actually less a paper that sets out a set of technology accounts than a proposal how to measure technical progress • It is well known that MFP is not a pure measure of technical change • Li proposes the rate of R&D depreciation to act as a measure of the pace of technological change • Starting point (similar to Huang&Diewert): – R&D depreciate because they contribute less to a firm’s profits – Reasons • technical progress that renders knowledge obsolete • increasing competition • Li’s hypothesis: if an industry in country A has a technological advantage over the same industry in country B, R&D depreciation should be lower in country A than in country B 5
…but how are R&D depreciation rates obtained? • Optimal R&D investment model: • Today’s marginal expenditure on R&D = discounted future increases in revenues from R&D • Estimation requires: – Forecasts of sales – Measure of maximum service life of R&D – Measure of maximum future mark-up ratio on sales due to R&D over time – Discount rate • Then, optimality conditions imply that the only unknown parameter is depreciation rate and Wendy evaluates it numerically with non-linear procedure • Results for 4 industries and 4 countries 6
Example: The Pharmaceutical Industry: Large differences in δRD – usually disquieting But to be expected if Wendy’s approach holds Country δRD Ranking Forbes’ Ranking United States 10% 1 1 Japan 13% 2 2 Germany 23% 3 3 China 46% 4 4 South Korea 76% 5 5
More comments • Every difference in technology may imply difference in R&D depreciation but does every difference in R&D depreciation also imply difference in technology? • Or: how to control for market structures and institutions? • Some assumptions need further clarification, e. g. , why can IΩ be set to equal r? • Compare with Huang-Diewert approach • Robustness checks? • Embed in accounting framework (or change title of paper) • Overall, good to have work and additional evidence in this area as empirical basis is weak 8
Back to Moulton-Mayerhauser: Environmental Assets and Natural Resources Quick reminder: SNA Asset classification Produced, non-financial assets Fixed assets (e. g. machinery, equipment) Inventories Valuables Non-produced, non-financial assets -Natural resources Land incl. surface water Mineral and energy reserves Non-cultivated biological resources (e. g. natural timber) Other natural resources (radio spectra) -Contracts, leases and licences, goodwill and marketing assets Within SNA asset boundary but often not measured (exceptions eg France, Australia, Canada, Korea) 9
Back to Moulton-Mayerhauser: Environmental Assets and Natural Resources This is acutally not very different from asset classification in SEEA central framework, except: • Broader in scope (e. g. proven but not economically viable mineral resources, land without ownership right) • Separation land soil • Groundwater resources included What most people think about when it comes to the environment is ecosystems and their services • Ecosystem = dynamic interplay of individual environmental assets • Examples: atmosphere, oceans, forests • SEEA recognises them but outside central framework Brent and Nicole do not propose including ecosystem measures into the national accounts But they appear even sceptical about measuring natural resources (within the SNA boundaries? ) 10
Environmental Assets and Natural Resources: Comments • I agree: leave ecosystem services outside the accounts • But account for SNA natural resources: e. g. , land, subsoil assets, biological resources: stocks, discoveries and depletion • Obtain depletion-adjusted income • Brent and Nicole : in this case, symmetry requires that discoveries are treated as output and as GFCF • My own preference: keep discoveries as other volume changes: • Hard to explain treatment as output: no immediate income flow is generated, only wealth effect • Not too worried about asymmetry • Brent and Nicole note significant challenges left in the valuation of natural resources: I agree and consider these important areas of work for implementation of the current SNA 11
Moulton-Mayerhauser: Consumer Durables • SNA does not recognise consumer durables as assets • SNA reason: capital services of CD outside production boundary (§ 3. 47) • But is this reason enough? – Asymmetry in treatment of the same good depending on purchase by HH or corporation – Importantly: downward bias in HH net worth, indebtedness and savings – Loan for car = HH liability but car is not asset! • Brent and Nicole suggest bringing CD into asset boundary • They work through the accounting implications and provide three options, their preferred one being a treatment similar to OOH • My full support! 12
Moulton-Mayerhauser: Human Capital • Key concept in research and analysis • Recognition, including by SNA that education has character of capital: increase of future production possiblities (§ 1. 95) • Efforts to develop HC measures by researchers – May require extended production boundary (HH combine education output, time etc. to produce additions to human capital) – Cost or income valuation: can yield very different results (Abraham 2010) – Best-known (and most comprehensive ) income approach: Jorgenson and Fraumeni (1989) – Scope: labour income only or non-labour income as well? – Whatever precise estimates, they tend to be large and can swamp the value of other assets – HC cannot be transacted so interpretation as part of balance sheets has to be different from other assets • M-M conclude that the core SNA is not ready for HC and satellite accounts are proper venue for monitoring stocks and flows of HC 13
Human Capital: comments • I concur with this overall conclusions by M-M • But HC estimates as such are useful and policy-relevant (e. g. , managing immigration) • Also – Jorgenson-Fraumeni demonstrated that additions to HC due to education is a valid measure of the output of the education industry – Break down of labour input and Co. E by skill/educational attainment and construction of labour quality index = capturing evolution of HC in labour input • Thus, HC can be an empirical tool for the NA within current boundaries 14
Conclusion • WL paper: new take on measuring and interpreting R&D depreciation –useful in an area where empirical evidence is still scarce • M-M paper: very thoughtful and thorough discussion of asset boundaries in the SNA and implications from changing theme • They support including consumer durables and some knowledge assets into the SNA – I’d definitely add natural resources (but not ecosystem services) to the list • Overall, 2 interesting reads – thanks! 15
- Slides: 15