CONCLUSION

Strategy& reports:²⁵

By the year 2020, an entire generation, Generation C (for ‘connected’), will have grown up in a primarily digital world. Computers, the Internet, mobile phones, texting, social networking - all are second nature to members of this group.

Some observers assume such ‘transformations’ automatically translate to major, positive impacts on economic growth and productivity. But do they?

Before we reprise our answer to this question - which is essentially the same question as the subject of this chapter, the impact of the Internet on productivity - we need to underscore that we have discussed and analyzed ways in which measured living standards, as reflected by increases in labor productivity, are impacted by the direct and indirect channels of Internet-related economic activity. We did not discuss impacts on consumer surplus or what Greenstein and Nagle called ‘digital dark matter’, economic activity that cannot be seen because it is beyond the production boundary set by national accountants when they tally up GDP. This would take us to non-market uses of household time, where changes due to connectivity may indeed be dramatic but for which economic implications are unclear.

Household and business connectivity has increased dramatically since the late 1990s, however, suggesting that the economic impacts of the diffusion of digitization are creep­ing into the statistics. By analyzing the productivity experience of the last dozen or so years and discounting the extraordinary productivity experience of the late 1990s (as Gordon would have us do), we are able to rely on trends established during the period described by some as the Internet second wave.

We sorted through this terrain in two steps. First we relied on an analysis of ICT price change that revealed a central finding, namely that prices of products that determine the capacity of Internet and wireless networks suggest that advances in communica­tions technology are not slowing down. The analysis also revealed slowdowns in price declines for (1) computer and software investments and (2) Internet access services after 2007 that were difficult to explain. Altogether these factors are rather consequential for assessing the size and significance of recent tech-driven contributions to productivity growth.²⁶ Contributions from ICT capital deepening and TFP in ICT-producing indus­tries accounted for a noteworthy fraction (two-thirds) of the growth in labor productivity in Europe and the USA from 2001 to 2013. If trends in the components underlying ICT continue at post-2001 average (adjusted) rates, ICT will contribute between 0.66 (Europe) and 1 (US) percentage points per year to labor productivity growth going forward.²⁷

How much of a stretch is it to assume that other major sources of growth - increases in labor composition, non-ICT capital deepening, and productivity growth in non-ICT producing industries - will contribute another 0.66-1.0 percentage point, thereby raising labor productivity growth to between 1.5 and 2 percent per year?²⁸ Put differently, con­sider whether the Internet and/or ICT use contributes an extra punch to productivity growth via externalities, spillovers, or complementarities and ICT thereby boosts eco­nomic growth even further.

After setting out how network effects might be expected to leave their footprints in an SOG analysis, we reframed the story of the 1990s and 2000s to center on the Internet and the demand for connectivity as the impetus for change, not on the microprocessor as an independent supply-side driver of change. Rather, following Metcalfe’s Law, the ICT contributions that were so evident in the 1990s were just part of a story that played itself out via network effects (TFP gains) in the early and mid-2000s (prior to the financial crisis). The available evidence suggests that TFP in non-ICT-producing industries was boosted via network effects by 0.25 percentage points per year in the market sector of eight major EU economics from 2002 to 2007 and by 0.5 percentage points in the private business sector of the USA from 2001 to 2005. As noted in our discussion, not all of these gains are sustainable; indeed, some might argue against building them into trends to the extent they were part of an already played out process of computerization that began in the 1990s. The impetus to innovation stemming from connectivity and the availability of high-speed communication platforms has not played out, however, and this is part of the just-quoted estimates. To answer the question originally posed at the conclusion of section 6.2.1, then, we believe that Internet and digital technology remain a supply-side force for future TFP change, and that the potential impact is relatively strong: if such technology contributed about 0.1 to 0.2 percentage points per year to market sector TFP change during the mid-2000s via innovative adaptations of networks, then we would expect this force to be at least as strong, if not stronger, going forward.²⁹

The network effects framework we offered in Table 6.2, itself drawn from Corrado (2011), Corrado and Jager (2014), and based in part on Bresnahan and Trajtenberg (1995), sets out a way of thinking about production and innovation due to the Internet.

All told, many subtleties are involved in how the Internet and investments in networks impact productivity growth, and one of the chapter’s main messages was to underscore that, if the Internet and wireless networks are the highways of the modern age (on which traffic is growing at explosive rates worldwide), we need to use models and data that are up to the task of analyzing their macroeconomic impacts. We believe the source-of- growth framework works well in this regard, provided one agrees with how we ‘fit’ the Internet and networks into it. As to the data required for this analysis, we will not reprise specific comments on ways that existing data fall short for an analysis of ICT develop­ments since the early 2000s, but another conclusion of this chapter is that filling those gaps is an important subject of future research.

NOTES

* We thank Abdul Erumban and Kirsten Jager for the growth accounting estimates reported in this chapter.

1. TFP thus includes the impacts of implementing knowledge and technology ‘borrowed’ from elsewhere. In noting this, we are underscoring the important distinction between scientific invention and innovation, where the latter includes the economic value created by all activities associated with the commercialization of knowledge and technology.

2. EU-27 excludes Croatia, which joined the European Union as its 28th member on 1 July 2013.

3. IoT devices exclude PCs, tablets and smartphones. The sources for these forecasts are Boston Consulting Group (accessed 3 July 2015 at http://www.marketwired.com/press-release/g-20s-internet-economy-is- set-reach-42-trillion-2016-up-from-23-trillion-2010-as-nearly-1611718.htm), Cisco’s Global Cloud Index (2013-18) (accessed 3 July 2015 at http://www.cisco.com/c/en/us/solutions/service-provider/global- cloud-index-gci/index.html) and Gartner (accessed 3 July 2015 at http://www.gartner.com/newsroom/ id/2636073).

4. In this regard, Byrne and Corrado’s assessment of price change for communications equipment is in sync with Hilbert and Lopez’s compilation of figures on the world’s capacity to compute, store, and com­municate information (Hilbert and Lopez, 2011). Hilbert and Lopez cover fewer years, 1986 to 2007, but also consider broadcasting; Byrne and Corrado (2015a) cover telecommunications only, but from 1963 to 2013. See Byrne and Corrado (2015b) for further discussion.

5. See, for example, the keynote by Seizo Onoe, CTO and EVP and Member of Board of Directors, NTT DOCOMO, ‘5G technology’ at the workshop ‘Can Mobile Broadband Realize its Full Potential?’, spon­sored by the Wireless Technology Association (WTA), Mobile Computing Promotion Consortium of Japan, and Georgetown University, Washington DC, 29 October 2014.

6. Even though the remainder of this section mostly discusses the US Bureau of Economic Analysis (BEA) price deflators, the BEA computer price index is used to harmonize ICT capital measures in EUKLEMS, and price measures for all ICT components for non-EUKLEMS countries are harmonized to US prices in The Conference Board’s Total Economy Database, the source for the growth accounting results reported in Table 6.1.

7. The communications equipment price index plotted is a research series originally developed by Byrne and Corrado in 2007 to include wireless equipment and to update earlier work by Doms (2005). The original Byrne-Corrado indexes continue to be maintained and updated by Federal Reserve Board staff. BEA incorporated some of this work into the National Income and Product Accounts (NIPA) in 2011. The overall ICT index plotted in the figure is constructed by aggregating the Byrne-Corrado research price index for communication equipment for all years with BEA’s official prices for computers and software, thereby providing a consistent historical series for total ICT investment prices from 1959 to present. Real prices are constructed using the price index for business output.

8. The story as reported in Byrne and Corrado (2016) goes like this: In the early 2000s computer chip makers found that pushing the performance envelope by increasing clock speed got harder due to power and heat considerations. Following standard practice of pulling circuit design innovations from the high-performance computing (HPC) world onto personal workstations, the computer industry turned to parallel processing via a multicore architecture. Note that the move toward parallel processing thus shifted the computing challenge from creating faster computer microprocessors to designing computer systems and software that utilized large numbers of processors efficiently - a problem then on the cutting edge of academic supercomputing R&D.

9. Note that the estimated impact on labor productivity is the same as private investment because, after adding in other components of final demand and subtracting imports, the average GDP weight of computer final sales is essentially the same as the GDP weight of computer investment (2007 to 2013). See US Bureau of Economic Analysis (BEA) National Income and Product Accounts (NIPA) Table 9.2U.

10. The producer price index (PPI) for application software has two published components from June 2006 on, desktop and portable device application software and other application software. The first component is flat and the second increases 10 percent between mid-2006 and mid-2014. According to Census Bureau data for 2010 to 2013, enterprise or network software is 73 percent of application software revenues excluding PC software, suggesting roughly 1.0 percentage point per year increases in application software price change stem from prices of enterprise and network software.

11. The measure used is from OOKLA, a company that tests and benchmarks ISP speed and performance, and which also compiles global metrics on broadband performance. The OOKLA net index of US metrics is available since January 2008. It is based on ISP-reported upload and download speeds, actual measures of connection stability, and survey-based estimates of median broadband costs in megabits per second and median ratio of actual download speed to promised/advertised speed. Note further that a characteristic price is a form of hedonic approach implying a simple, linear function through the origin of the hedonic price-single characteristic plane (Triplett, 2004). The OOKLA characteristic price measure reflects both commercial and residential customers (we have no way to separate the two segments) and will not be strictly comparable to the CPI to the extent mix shifts or differential hedonic surfaces prevail across segments.

12. See note 6 for why we apply the results of an analysis of US ICT prices to Europe.

13. Figure 6.5 does not show a particularly large drop off in national accounts value added in ICT manu­facturing, in part, because a sharp pullback in purchased services and increases in R&D and software investment on own-account offset a 22 percent drop in the value of factory production in computer and electronic manufacturing (NAICS 334) in the United States between 2000 and 2007.

14. See Karamti (2007) for results of a hedonic model of telecom pricing in France.

15. Metcalfe’s Law is named for a researcher once at Xerox’s famed Palo Alto Research Center. See ‘Beyond the ether’ in the Economist magazine’s Technology Quarterly, 12 December 2009, p. 23 for more information on Metcalfe and Metcalfe’s Law.

16. See Brynjolfsson and Kremerer (1996); Mun and Nadiri (2002); Roller and Waverman (2001), respectively, for examples of these types of studies.

17. To be clear, we do not take Metcalfe’s Law literally; it is based on certain assumptions, such as equal value of all connections, that may overstate the benefits from networks.

18. To the extent increases in consumer welfare occur via network effects (i.e., via the number of users on a network), then economic theory suggests a cost of living index should take these welfare-enhancing effects into account. But the number of users does not play into the computation of consumer price indexes; indeed no externalities do. For example, as noted in Fixler et al. (2001), the CPI for auto alarms is not adjusted according to the rate of car thefts deterred, and so on.

19. That impetus behind the late 1990s investment boom was the demand for Internet access, not the microprocessor, is argued more fully in Corrado (2011).

20. As a theoretical matter, Berndt and Fuss (1986) showed that to the extent capital productivity (which varies directly with capital utilization) is proportional to the marginal product of capital, capital utilization is absorbed in capital income and capital services as conventionally calculated. For further discussion and caveats see Hulten (2009).

21. Some key studies that looked for externalities from ICT capital in growth accounting datasets but were unable to find them are Stiroh (2002), who used industry-level data for the United States, and Inklaar et al. (2008), who used industry-level data for ten EU countries and the United States. For recent reviews, see Biagi (2013) and Cardona et al. (2013). Note that the IT spillover estimated in the Corrado and Jager (2014) study mentioned in the previous section and attributed to network effects applied only to the period after 2002; testing for the same effects in the same dataset using observations before 2002 failed to uncover a statistically significant spillover coefficient for IT. In view of this finding, and the earlier literature that also could not detect spillovers to IT capital investments, Corrado and Jager argued that what they found was related to network effects.

22. In addition to developments at national statistical offices, this is owed to work by researchers too numer­ous to mention and to the support given to them by, among others, the European Commission, NESTA, Organisation for Economic Co-operation and Development (OECD), Research Institute of Economy, Trade and Industry (RIETI), The Conference Board, UK Intellectual Property Office, US National Science Foundation, and World Intellectual Property Office.

23. The latter is a cross-country econometric study of ten European countries using intangible capital aug­mented growth accounts data from 1995 to 2007. The study included the possibility of productivity spillo­vers from investments in ICT capital but none were found. In both sets of findings, computer software is included in ICT, that is, ICT is the conventional definition in the economics literature, and non-R&D intangibles refers to all intangibles (Corrado et al., 2005, 2009) except software and scientific R&D, that is, it includes (1) other new product development expenditures, for example, on artistic and entertain­ment originals and industrial design, and (2) investments in economic competencies, for example, brands, organization structure, and firm-specific training.

24. The findings on complementarities by asset type refer to the Corrado et al. (2015) study.

25. Accessed 3 July 2015 at http://www.strategyand.pwc.com/global/home/what-we-think/digitization/ megatrend.

26. As previously noted, the price and tech analysis is drawn from other works (Byrne and Corrado, 2015a, b, 2016).

27. Some observers may be surprised to see such similar figures for Europe and the United States, but the dismal EU productivity performance of late does not have its roots in the quantity of its ICT investments or productivity of its ICT producers. That said, policies directed at ICT are a path forward, as argued in Van Ark (2014a).

28. The average rate of growth in output per hour in the United States for the past 63 years is about 2 percent.

29. The estimate of 0.1 is the point estimate from the EU study, and the estimate of 0.2 is obtained by applying the same model to parse the US study’s estimate of total network externalities.

REFERENCES

Bartelsman, E.J., P.A. Gautier and J. de Wind (2010), ‘Employment protection, technology choice, and worker allocation’, De Nederlandsche Bank Working Paper No. 295, accessed 3 July 2015 at http://ssrn.com/ abstract=1951839.

Basu, S., J.G. Fernald, N. Oulton and S. Srinivasan (2004), ‘The case of the missing productivity growth: Or, does information technology explain why productivity accelerated in the United States but not in the United Kingdom?’, in M. Gertler and K. Rogoff (eds), NBER Macroeconomics Annual 2003, Cambridge, MA: The MIT Press, pp. 9-82.

Berndt, E.R. and M.A. Fuss (1986), ‘Productivity measurement with adjustments for variations in capacity utilization and other forms of temporary equilibrium’, Journal of Econometrics, 33 (1), 7-29.

Biagi, F. (2013), ‘ICT and productivity: A review of the literature’, JRC Technical Reports Digital Economy Working Paper No. 2013/09, Institute for Prospective Technological Studies, Seville, Spain, accessed 4 July 2015 at http://publications.jrc.ec.europa.eu/repository/handle/JRC84470.

Black, S.E. and L.M. Lynch (2001), ‘How to compete: The impact of workplace practices and information technology on productivity’, Review of Economics and Statistics, 83 (3), 434-45.

Bloom, N. and J. van Reenen (2007), ‘Measuring and explaining management practices across firms and countries’, Quarterly Journal of Economics, 122 (4), 1351-408.

Bloom, N. and J. van Reenen (2010), ‘Why do management practices differ across firms and countries?’, Journal of Economic Perspectives, 24 (1), 203-24.

Bresnahan, T.F and M. Trajtenberg (1995), ‘General purpose technologies: “Engines of growth?”’, Journal of Econometrics, 65 (1), 83-108.

Brynjolfsson, E. and L.M. Hitt (2000), ‘Beyond computation: Information technology, organizational transfor­mation and business performance’, Journal of Economic Perspectives, 14 (4), 23-8.

Brynjolfsson, E. and C.F Kremerer (1996), ‘Network externalities in microcomputer software: An econometric analysis of the spreadsheet market’, Management Science, 42 (12), 1627-47.

Brynjolfsson, E. and A. McAfee (2014), The Second Machine Age, New York: Norton.

Brynjolfsson, E., L.M. Hitt and S. Yang (2002), ‘Intangible assets: Computers and organizational capital’, Brookings Papers on Economic Activity, 1, 137-98.

Brynjolfsson, E., S.T. Kim and J. Oh (2014), ‘User-generated capital and firm value: Theory and evidence from Internet firms’, paper presented at the 2014 NBER Summer Institute Digitization Workshop, Cambridge, MA.

Byrne, D.M. and C.A. Corrado (2015a), ‘Prices for communications equipment: Rewriting the record’, FEDS Working Paper 2015-069 (September), Federal Reserve Board, Washington, D.C.

Byrne, D.M. and C.A. Corrado (2015b), ‘Recent trends in communications equipment prices’, FEDS Notes (29 September), Federal Reserve Board, Washington, D.C.

Byrne, D.M. and C.A. Corrado (2016), ‘ICT prices and ICT services: What do they tell us about productivity and technology?’, forthcoming working paper.

Byrne, D.M., S.D. Oliner and D.E. Sichel (2015), ‘How fast are semiconductor prices falling?’, NBER working paper No. 21074 (April), Cambridge, MA.

Cardona, M., T. Kretschmer and T. Strobel (2013), ‘ICT and productivity: Conclusions from the empirical literature’, Information Economics and Policy, 25 (3), 109-25.

Caroli, E. and J. van Reenen (2001), ‘Skill-biased organizational change? Evidence for a panel of British and French establishments’, The Quarterly Journal of Economics, 116 (4), 1449-92.

Chen, W., T. Niebel and M. Saam (2014), ‘Are intangibles more productive in ICT-intensive industries? Evidence from EU countries’, ZEW Discussion Papers No. 14-070 (September), Centre for European Economic Research (ZEW), Germany.

Corrado, C. (2011), ‘Communication capital, Metcalfe’s law, and US productivity growth’, Economics Program Working Paper No. 11- 01, New York: The Conference Board, Inc.

Corrado, C., J. Haskel and C. Jona-Lasinio (2014), ‘Knowledge spillovers, ICT, and productivity growth’, The Conference Board, CEPR, and IZA Working Paper (June).

Corrado, C., J. Haskel and C. Jona-Lasinio (2015), ‘Intangibles, ICT and industry productivity growth: Evidence from the EU’, forthcoming in D. Jorgenson, K. Fukao and M. Timmer (eds), The World Economy: Growth or Stagnation? Cambridge, UK: Cambridge University Press.

Corrado, C., C. Hulten and D. Sichel (2005), ‘Measuring capital and technology: An expanded framework’, in C. Corrado, J. Haltiwanger and D. Sichel (eds), Measuring Capital in the New Economy, Studies in Income and Wealth, Volume 65, Chicago, IL: University of Chicago Press, pp. 11-46.

Corrado, C., C. Hulten and D. Sichel (2009), ‘Intangible capital and US economic growth’, Review of Income and Wealth, 55 (3), 661-85.

Corrado, C. and K. Jager (2014), ‘Communication networks, ICT, and productivity growth in Europe’, Economics Program Working Paper No. 14-04, New York; The Conference Board, Inc.

Doms, M. (2005), ‘Communication equipment: What has happened to prices?’, in C. Corrado, J. Haltiwanger and D. Sichel (eds), Measuring Capital in the New Economy, Studies in Income and Wealth, Volume 65, Chicago, IL: University of Chicago Press, pp. 323-62.

Faggio, G., K.G. Salvanes and J. van Reenen (2010), ‘The evolution of inequality in productivity and wages: Panel data evidence’, Industrial and Corporate Change, 19 (6), 1919-51.

Fixler, D., J.S. Greenlees and WJ. Lane (2001), ‘Telecommunications indexes in the US Consumer Price Index’, paper presented at the Sixth Meeting of the International Working Group on Price Indices (April), Canberra, Australia.

Flamm, K. (2007), ‘The microeconomics of microprocessor innovation’, paper presented at 2007 NBER Summer Institute Productivity Workshop.

Forman, C., A. Goldfarb and S. Greenstein (2003a), ‘The geographic dispersion of commercial Internet use’, in S. Wildman and L. Cranor (eds), Rethinking Rights and Regulations: Institutional Responses to New Communication Technologies, Cambridge, MA: MIT Press, pp. 113-45.

Forman, C., A. Goldfarb and S. Greenstein (2003b), ‘Which industries use the Internet?’, in M.R. Baye (ed.), Organizing the New Industrial Economy, Advances in Applied Microeconomics, Volume 12, Amsterdam: Elsevier, pp. 47-72.

Gordon, R.J. (2014a), ‘The demise of US economic growth: Restatement, rebuttal, and reflections’, NBER Working Paper No. 19895, Cambridge, MA: NBER.

Gordon, R.J. (2014b), ‘A new method of estimating potential real GDP growth: Implications for the labor market and the debt/GDP ratio’, NBER Working Paper No. 20423, Cambridge, MA: NBER.

Greenstein, S. (2000), ‘Building and delivering the virtual world: Commercializing services for Internet access’, Journal of Industrial Economics, 48 (4), 391-411.

Greenstein, S. and F. Nagle (2014), ‘Digital dark matter and the economic contribution of Apache’, Research Policy, 43 (4), 623-31.

Griliches, Z. (1994), ‘Productivity, R&D and the data constraint’, American Economic Review, 81 (1), 1-23.

Hausman, J. (1999), ‘Cellular telephone, new products, and the CPI’, Journal of Business & Economic Statistics, 17 (2), 188-94.

Hilbert, M. and P. Lopez (2011), ‘The world’s technological capacity to store, communicate, and compute infor­mation’, Science, 332 (6025), 60-65.

Holt, L. and M. Jamison (2009), ‘Broadband and contribution to economic growth: Lessons from the US experience’, Telecommunications Policy, 33 (10-11), 575-81.

Hulten, C. (2009), ‘Growth accounting’, NBER Working Paper No. 15341, Cambridge, MA: NBER.

Ichniowski, C. and K. Shaw (2003), ‘Beyond incentive pay: Insiders’ estimates of the value of complementary human resource management practices’, Journal of Economic Perspectives, 17 (1), 155-60.

Inklaar, R., M.P. Timmer and B. van Ark (2008), ‘Market services productivity across Europe and the US’, Economic Policy, 23 (53), 139-94.

Isaacson, Walter (2014), The Innovators, New York: Simon & Schuster.

Jackson, D.J. (2011), ‘What is an innovation ecosystem?’, White Paper, Arlington, VA: NSF Engineering Research Center.

Karamti, C. (2007), ‘Hedonic price indexes for mobile telephony services in France’, Working Paper No. ESS- 07-07, Paris: Telecom Paris.

Levin, J.D. (2011), ‘The economics of Internet markets’, NBER Working Paper No. 16852, Cambridge, MA: NBER.

Mun, S.B. and M.I. Nadiri (2002), ‘Information technology externalities: Empirical evidence from 42 US indus­tries’, NBER Working Paper No. 9272, Cambridge, MA: NBER.

Roller, L.H. and L. Waverman (2001), ‘Telecommunications infrastructure and economic development: A simultaneous approach’, American Economic Review, 91 (4), 909-23.

Rosenthal, S., M. Russell, J. Samuels, E.H. Strassner and L. Usher (2014), ‘Integrated industry-level production account for the United States: Intellectual property product and the 2007 NAICS’, paper presented at the Third WORLD KLEMS conference, Tokyo, Japan, 19-20 May.

Solow, Robert M., (1987), ‘We’d better watch out’, review in the New York Times Book Review, 12 July 12, p. 36. Stiroh, K.J. (2002), ‘Are ICT spillovers driving the new economy?’, Review of Income and Wealth, 48 (1), 33-57. Stranger, G. and S. Greenstein (2007), ‘Pricing on the on-ramp to the Internet: Price indexes for ISPs during the 1990s’, in E.R. Berndt and C.R. Hulten (eds), Hard-to-Measure Goods and Services: Essays in Honor of Zvi Griliches, Studies in Income and Wealth, Volume 67, Chicago, IL: University of Chicago Press, pp. 197-233.

Syverson, C. (2011), ‘What determines productivity?’, Journal of Economic Literature, 49 (2), 326-65.

The Conference Board (2015), The Conference Board Total Economy Database™ May 2015. The Conference Board, accessed 1 March 2016 at http://www.conference-board.org/data/economydatabase/.

Triplett, J.E. (2004), ‘Handbook on hedonic indexes and quality adjustments in price indexes: Special applica­tion to information technology products’, STI Working Paper No. 2004/9, Paris: Organisation for Economic Co-operation and Development.

Van Ark, B. (2014a), ‘Productivity and digitalization in Europe: Paving the road to faster growth’, Policy Brief, Brussels and New York: The Lisbon Council and The Conference Board.

Van Ark, B. (2014b), ‘Total factor productivity: Lessons from the past and directions for the future’, Working Paper Series No. 271, Brussels: National Bank of Belgium.

Williams, B. (2008), ‘A hedonic model for Internet access service in the Consumer Price Index’, Monthly Labor Review, 131 (7), 33-48.

7.