By Diane Coyle and David Nguyen
How much do people value goods they do not have to pay for? Should this be taken into account in economic statistics?
Although there have always been public goods for which people do not pay directly, these questions have come to the fore now that ‘free’ digital goods are such an extensive part of daily life. Researchers have been exploring their value to consumers in the US, and often with a focus on Facebook and social media. In our new ESCoE Discussion Paper we considered for the UK a wide range of goods using an approach based on a survey to people covering their valuations for both digital and non-digital (like public parks) products and also some paid-for substitutes (such as newspapers and online news). What is more, the natural experiment of the first 2020 lockdown in between survey waves allowed us to look at how people’s valuations changed.
We used large YouGov surveys representative of the UK’s online population to ask people how much they would be willing to accept for the loss of each category for either one month or 12 months. We found that users place a high value on free digital goods, with mean valuations strongly correlated to usage (i.e. the proportion of respondents using them). The most valued are search and personal email, the least valued are more niche services such as online travel tools or learning resources. The mean annual value assigned to search and email was over £3,000 (the median valuation bands were £1001-2000 and £2001-5000 respectively). For others such as online ride hailing or online learning, the mean was around £200.
There were large changes in both usage and valuations between February and May 2020. However, online learning, gaming and online shopping for example saw significant increases, with smaller increases for other cases such as public parks. The mean annual valuation of online learning rose from £247 to £309 in just 10 weeks, and for parks from £1951 to £2063. The usage and valuations dropped for Google maps and also physical substitutes such as cinemas and newspapers. The changes are intuitive but confirm that the contingent valuation survey method can be used to measure such relative movements or at least form a good indicator of change.
Comparing online goods and offline substitutes, the online valuations are considerably higher, suggesting that there may be aspects of online use such as convenience, choice or time-saving that deliver consumer value. The ‘elasticity’ or responsiveness of willingness-to-accept (WTA) in response to usage varies widely between different goods – some, such as search and personal email, are highly valued at all usage levels. We did not test willingness-to-pay (WTP) for specific goods, but consistent with the contingent valuation literature the WTA values for free or public goods were much larger than actual revenues per user or comparable prices for marketed goods.
Our approach allowed us to look at distributional questions, and there were large differences in valuations in terms of gender, age, income and social group, and sometimes also region of the UK. For example, women required 40% more than men to be willing to give up Facebook for 12 months. Parks are most valued by people aged 20-49 and much more valued by Londoners than people elsewhere in the country. The distribution of values is important when thinking about constructing an aggregate welfare measure.
Our selection of goods allowed us to be able to compare stated values for online goods and physical substitutes. One of the pairings is online news and printed newspapers. In recent years there has been progressive substitution from print to online formats. In our February 2020 sample, 73% of respondents stated that they read news online and on average required £1,253 to give up access to online news for 12 months. Around 55% of respondents said they read printed newspapers and on average placed a WTA valuation on this of £566. Readership of online news is the lowest (65%) among those aged 65+, as are annual valuations (£832). At the same time this age group has the highest share of reading printed newspapers (66%) with the highest average valuation (£895). Valuations on online news are the highest for respondents aged 18-24 (£1,685) and twice as high as for the over-65s. Over the 10 weeks between end of February and mid-May 2020, stated usage of printed newspapers declined from 55% to 47%, while use of online news increased from 73% to 75%. At the same time average valuations of printed news decreased by 24% from £566 to £430, while the valuation of online news decreased minimally from £1,253 to £1,243. It is not clear whether this would be reversed when things move back to some ‘new’ normal, thus further surveys would be useful identifying such changes.
In the absence of other methods for estimating the consumer surplus value of free digital goods, our research indicates that the stated value approach – widely used in environmental and cultural economics – gives valuable insights. However, there are a number of issues to tackle before the estimates could be used for aggregate measurement of economic welfare, as has been suggested. The series of pilots and surveys we carried out suggest stated values are large but not tightly anchored and have large standard errors. There would be further benefit from having a sub-sample in the form of a panel to allow for better period-on-period and like-for-like respondent comparisons.
More significantly, it is not clear how to define and partition the universe of goods to survey. For example, stated values for social media do not add to the sum of stated values for each social media platform named separately. Should survey questions ask about categories like ‘social media’ or specific services like Facebook, Twitter and SnapChat? What about new goods – how many people now working from home had heard of Zoom before March 2020? If we were to include valuations of free digital goods to measure aggregate welfare, why not include other free but non-digital goods also? The large distributional differences in usage and values would also need to be taken into account for an economic welfare aggregate: how should different groups be weighted? Finally, it is not obvious how to impose an adding-up constraint in terms of the time spent using free digital goods and other goods, whereas with paid-for goods this constraint is provided by actual monetary expenditures.
You can download and read a copy of the Discussion Paper in full here.
Diane Coyle is Bennett Professor of Public Policy at the University of Cambridge and a Research Associate at ESCoE.
David Nguyen is a Senior Economist at National Institute of Economic and Social Research (NIESR) and a Research Associate at ESCoE.
ESCoE blogs are published to further debate. Any views expressed are solely those of the author(s) and so cannot be taken to represent those of the ESCoE, its partner institutions or the Office for National Statistics.