MEASURING THE VALUE OF DATA – THREE OBSERVATIONS Paul Schreyer OECD OECD workshop on the value of data to consumers 3 June 2021
Outline 1. Clarity of notion – the many meanings of ‘data’ 2. Productivity measurement and data assets 3. Free apps and GDP
1. Clarity of notion – the many meanings of ‘data’
3
Clarity of notion – the many meanings of ‘data’ • Term “Data” often used amorphously • From digital bits and bytes or individual case records to whole “business models” • Jim’s presentation; Carol/Cecilia/Johnathan/Massimiliano’s presentation already clear about this • OECD’s Going Digital • WP Measuring and Analysing the Digital Economy, WP on National Accounts, SNA update (eg Ker, Mitchell 2021)
2. Productivity Measurement and Data Assets
6
Productivity measurement and data assets • Productivity = Volume of output/volume of inputs • Service flow from data asset = additional input • What exactly is this service flow and what is its price? – Unlikely quantity of bits or bytes stored – Fall back = volume of inputs used in investing in data assets
Productivity measurement and data assets • Investment in new type of produced asset implies additional production process – own account • Increase in output and in inputs – Effects on measured MFP growth hard to predict – Measured LP levels will rise but unclear effects on LP growth
• Not possible to conclude that absence of data assets in productivity computations is reason for slowing productivity growth
3. Free apps and GDP
9
Consumer surplus • Discussion and interesting empirical work around free apps & services • Brynjolfsson, Collis, Diewert, Eggers and Fox (2018)
– Incentive compatible choice experiments – Willingness to forego digital services, in particular access to Facebook: – ≈ 500$/year: shadow price (marginal consumer surplus) not in GDP
Photo by Thought Catalog on Unsplash
10
25$ vs 500$ • Facebook sells advertising services • In 2017 approximative advertising revenues 25$/user/year < 500$/user/year (willingness to forego)
(i) Facebook does not act as a profit maximiser (unlikely), (ii) WTF estimates are vastly overstated (implausible), (iii)the 500$ value relates to a different act of production and consumption, not to the implicit barter transaction between consumers and Facebook 11
Own-account HH production • Production process by households who use: • time • capital services (hardware, software) including freely provided app
• to produce (typically, leisure) services associated with the use of social media • 500$ = value of own account production • No inconsistency with 25$ advertising revenues
Photo by David Boca on Unsplash
12
Conclusion
• No inclusion of willingness to pay/forego in GDP needed • But more systematic capture of HH production and innovation is important Photo by David Boca on Unsplash
13
Thank you!
Paul.Schreyer@oecd.org 14