The question of how to value digital services for which users pay no upfront cost has become increasingly important in two areas of policy interest. One is the accurate measurement of productivity, with a number of researchers suggesting that GDP, and hence productivity figures, are omitting a potentially large source of economic value (e.g. Brynjolfsson et al. 2019a, b). The other is competition policy, as authorities focus on market power in digital services (e.g. CMA 2020).
One way to explore consumer valuations of ‘free’ digital goods is to use contingent valuation methods, which are widely used in environmental economics (e.g. Carson et al. 2001). We had results from a YouGov survey representative of the UK online population and carried out in late February eliciting willingness to accept (WTA) the loss of 30 goods and services – a mixture of zero price online goods, market substitutes for some of these, and zero price offline goods such as access to a public park. Repeating the same survey in mid-May allows us to treat the lockdown in the UK as a natural experiment to test changes in these consumer valuations (Coyle and Nguyen 2020).
Consumer spending patterns clearly changed substantially during lockdowns (Chronopoulos et al. 2020). In the UK, as in other countries, online consumption in general increased (eg Relihan et al 2020). ONS figures show that internet shopping as a proportion of retail spending increased to a record 32.8% share in May 2020, and Ofcom reported that Brits were spending a quarter of their waking hours online.
Comparing our surveys between the end of February, before the lockdown, and mid-May (a ten-week period), there were some significant changes for some goods in the proportion of respondents saying they used or did not use them. The proportion reporting that they shop online for groceries increased from 50% to 60% in the second survey wave. The share of people not using Skype, Facebook Messenger, Netflix and WhatsApp decreased by around 5 percentage points. Other goods that saw a decline in non-usage rates were Facebook, online learning, mobile games, Amazon and Twitter. On the other hand, the usage of various other goods has declined. For example, while in February around 41% reported they do not read (physical) newspapers, this had increased to 48.5% by mid-May.
Figure 1 Proportion of respondents who do not use…
The reported valuations differ widely between the goods, from the low hundreds of pounds (12-month willingness to accept loss for online ride hailing or online learning) to over £3,000 (personal email, online search, TV set). We also saw some significant changes comparing pre- and post-lockdown. There were statistically significant increases in average valuations in the case of online groceries, online learning, WhatsApp, Netflix, Facebook, public parks, and TV sets, and significant decreases in a number of categories such as cinema visits (which were of course unavailable), Google maps, Twitter, and print newspapers. The changes in valuations were positively correlated with changes in usage.
In order to compare the relative valuations of different age groups across goods, we compared the ratio of valuations reported by those aged under 50 to those aged 50 or above (Figure 2). On average, younger people tend to place a higher valuation on most of the goods included.
Figure 2 Annual valuations: Young compared to old
Changes in valuations differed across age groups. For example, while valuations of Facebook decreased by a small amount for those aged 25-65, they increased by 26% for those aged 18-24 and by 38% for those aged above 65. In the case of online grocery shopping, the valuations increased for all age groups apart from those aged 18-24, rising by 127% for people aged 65 and above, while for people between 25-64 they increased by 37%.
We also found significant differences between men and women and between social groups, both in the level of willingness to accept and in the changes in valuation between February and May. For example, in February around 51% of men and 49% of women did not use online grocery shopping. Over the lockdown period, these proportions decreased to 44% for men and 38% for women. Valuation for online grocery shopping increased relatively more for women (+51%) than for men (+41.5%). The changes in valuations across social groups differed substantially, to the extent that for most categories they were opposite in sign. For instance, the valuations for LinkedIn and online learning decreased for grades A to C2, but increased considerably for grades D and E. In the case of online learning the WTA loss of access increased by more than 400% for semi-skilled, unskilled and manual workers (grade D).
Consistent with other studies, in our surveys consumers on average assign high willingness-to-accept valuations to many of these goods, particularly when benchmarked against revenue figures for the services (e.g. Sunstein 2019). The methodology can nevertheless give consistent rankings among goods.
It is also a useful way to assess changes in valuations. During the lockdown, there were significant and rapid changes in the contributions different goods and services make to consumer valuations, with differences by age group, social class, and gender. The lockdown has been a natural experiment suggesting the extent to which digital goods and physical goods are substitutes. As many of the goods we considered are free to use, these valuation changes give useful insights into economic welfare and activity that are not captured by changes in prices. They offer a forward-look at which services may be most valuable in a post-pandemic world if more activity continues to take place online. They also provide important, policy-relevant insights into distributional questions.
References
Brynjolfsson, E, A Collis and F Eggers (2019a), “Using massive online choice experiments to measure changes in well-being”, Proceedings of the National Academy of Sciences 116(15): 7250-7255.
Brynjolfsson, E, A Collis, W E Diewert, F Eggers and K J Fox (2019b), “GDP-B: Accounting for the Value of New and Free Goods in the Digital Economy”, NBER Working Paper No. 25695.
Carson, R T, N E Flores and N F Meade (2001), “Contingent Valuation: Controversies and Evidence”, Environmental and Resource Economics 19: 173–210.
Chronopoulos, D K, L M Wilson and O S John (2020), “Consumer Spending Responses to the Covid-19 Pandemic: An Assessment of Great Britain”, Covid Economics 34: 145-186.
CMA (2020) Online Platforms and Digital Advertising.
Coyle, D and D Nguyen (2020), “Valuing goods online and offline: the impact of Covid-19”, Covid Economics 33: 110-124.
Relihan, L E, M Ward, C W Wheat and D Farrell (2020), “The Early Impact of COVID-19 on Local Commerce: Changes in Spend Across Neighborhoods and Online”, Covid Economics 28: 1-28.
Sunstein, C (2019), “Valuing Facebook”, Behavioural Public Policy, 1-12.
