A promising governmental response to COVID-19 was the deployment of contact tracing phone apps. Building on the widespread use of smartphones and tracking capabilities of Apple and Google, policymakers hoped to harness digital technology to fight the pandemic. Switzerland launched SwissCovid in June 2020, which was followed by similar apps in Europe.1 Research suggested that the app would be effective if used by 80% of the smartphone users, or about 60% of the population (Hinch et al. 2020). In practice, the number of active users of the SwissCovid app in Switzerland is about 1.8 million (21%) (Blasimme and Vayena 2020). Similar adoption rates are found in other advanced economies.2
A second wave of COVID-19 in Europe in September 2020 spiralled out of control, necessitating the imposition of lockdowns and other measures.3 Data demonstrate the current ineffectiveness of contact tracing. In Switzerland, the ratio of people quarantined following contact tracing relative to the number of cases was about 1 before the rise in cases, and then dropped to less than 0.5 (Figure 1a). This means that too few people were traced and quarantined following a positive test. In Italy and the Netherlands, the share of new cases discovered through contact tracing dropped to less than 20% at the end of October 2020 (Figure 1b). In theory, with perfect tracing, most cases can be traced (Blasimme and Vayena 2020).
Figure 1 The ineffectiveness of contact tracing to prevent new infections
Sources: Switzerland FOPH, Italy: Istituto Superiore di Sanità. Netherlands: National Institute for Public Health and the Environment
Why did contact tracing apps fail?
In a survey conducted by the University of Zurich,4 46.5% of respondents said they use the SwissCovid app, as opposed to actual use of 21%. Therefore, a large percentage of respondents thought they should report using the app even if they do not use it. Many respondents appear to understand the social importance of using the application, but are not using it.
Thirty percent of non-users thought the applications are not useful, while 23% claimed they did not have the right type of phone. Data on smartphone penetration for Switzerland show that between 6.5 and 7.5 million people have such phones.5 Assuming this is not an excuse, it seems that some have the right phone but need technical assistance to install the app.
Only 22% of non-users cited privacy concerns. The app usage rate would go up to 88% if this were the only reason for people not to use them. Public trust-building and data protection assurances can help reduce these numbers. However, we believe that economics can help us understand the more common reasons for not using the app.
Economic explanations for apps non-use
COVID-19 spreads through social contact. The virus exploits negative network externalities. The problem of externalities in economics is well-known – individuals calculate their cost and benefit from an action without internalising the cost their behaviour imposes on others. Well-known examples are pollution and smoking.
Using contact tracing apps exposes individuals to information that could cause them to have to quarantine. This involves an economic cost of forgone work and social isolation. If the contacted individuals develop symptoms, they would have to isolate anyway. However, since not everyone in contact with an infected person is infected, and not every infected person develops symptoms, costly quarantine seems unnecessary. From the individuals’ point of view, they would rather wait for the appearance of symptoms than quarantine themselves. However, this does not internalise the cost of potentially infecting others. Individuals also underestimate the future potential costs they might bear if infection rates increase to the point that they cause a lockdown that has macroeconomic costs that affect them.
Using a tracing app allows the negative network externalities of COVID-19 to be addressed by breaking the chain of infections. However, the social benefit of doing so does not directly benefit infected individuals. Common to network goods, the effectiveness of tracing apps increases with use.
Procrastination stems from shortsightedness – deferring action until tomorrow does not seem costly. However, over time, deferring action adds up to a considerable and sometimes irreversible cost (Akerlof 1991). The cost of procrastination is especially high in the case of negative network externalities. Multiple small decisions to delay app adoption add up to large social costs, leading to a pandemic with loss of lives and macroeconomic costs.
Incentives for using digital tracing apps
Economists propose taxes, financial incentives, and regulation to mitigate negative externalities. Taxes are uses to decrease the consumption and production of goods with negative externalities (e.g. carbon taxes, cigarettes). Though not unfeasible, it is less common to tax the non-use of goods. Therefore, we focus on carrots (i.e. financial incentives) and sticks (i.e. regulation).
Financial incentives are needed to overcome the disincentive to quarantine and to address the difference between the social and individual benefits of using the apps. The disincentive to quarantine can easily be met by guaranteeing individuals ‘paid leave’ during quarantine. Vouchers for delivery services (to cover delivery costs), and possibly for streaming applications such as Netflix, could also be provided.
Bridging the gap between the private and social benefits of using the app could be done by offering financial incentives to use the application. Fripmong and Helleringer (2020) show that offering financial incentives can increase the likelihood of using the app. Financial incentives could address respondents in the Swiss survey who thought the application is ineffective. These respondents have no ethical or privacy concerns. If they are indifferent to the app, a financial reward may make it beneficial for them to use it even if they think it has no social value. Finally, rewards also deal with procrastination, especially if they are limited-time offers.
As with pollution, monetary incentives are not always sufficient and are complemented with regulation. A mix of monetary incentives and regulation reduces the regulatory burden and enforcement costs by focusing on critical areas and, at the same time, does not rely entirely on rational market response.
We propose augmenting positive financial incentives with regulation requiring an active app to be presented before entering locations with high infection risk (e.g. restaurants, bars, sports and other entertainment venues, schools, and higher education institutions).
Using financial incentives to induce socially responsible behaviour has been challenged by behavioural economics. Backed by experimental evidence, behavioural economists argue that the monetisation of social behaviour can have the opposite effect. Turning social responsibility into a transaction may detract from its appeal to many (Gneezy and Rustichini 2000, Bowles 2008).
Unfortunately, COVID-19 provides a real-life experiment on the effectiveness of moral suasion and social responsibility in the world’s advanced societies. The failure of social distancing and the low uptake of COVID tracing apps cast a dark shadow on social responsibility. Evidence on app use shows that, at best, about 25% of the adult population is willing to participate actively in the effort to stop the spread of COVID-19.
A common explanation for the low use of apps is lack of trust in the government. Unfortunately, the evidence does not support this explanation. Cross-country data show that trust in government has almost no predictable power in explaining app penetration (Figure 2). While policies aimed at building trust between citizens and governments are a worthy, trust cannot be a significant reason for the failure of COVID-19 tracing apps use in many countries.
Figure 2 Trust in government does not account for COVID-19 app penetration rate
Sources: OECD Trust in government; App penetration: Australia, France, Germany, Ireland, Italy, Switzerland (5). Others from MIT Technology Review COVID tracing tracker database
Therefore, financial incentives should be an important part of the toolkit to increase the use of tracing apps. Behavioural economists advocate using nudge techniques (Thaler and Sunstein 2009) rather than more high-powered incentives. Nudging involves awareness campaigns and offering symbolic monetary incentives. Frimpong and Helleringer (2020) show that in the case of COVID apps, peanuts will not achieve the desired effect. Recent research on Ireland’s success in mitigating the use of plastic shopping bags shows that to be effective, the cost of using them cannot be trivial
Using financial incentives raises ethical and distributional issues. These could be addressed by offering paid leave compensation for quarantine progressively, paying out a lower percentage of high-income earners’ salaries. Incentives to use the app are progressive in nature. They benefit low-income wage earners relatively more than high-income earners. Moreover, ethical concerns could be addressed by valuable non-cash incentives, such as vouchers for a smartphone purchase or monthly mobile service subscriptions or vouchers that can be used in retail stores (Loi 2020). The vouchers provide stimulus to the retail sector that suffers disproportionally from social distancing.
Given what is at stake we should adopt a holistic policy that includes education, public campaigns, trust-building, accountability, and nudging. However, these measures take significant time to be effective, and COVID-19 is not waiting. Therefore, we recommend immediately using easy-to-implement financial incentives and regulation.
Akerlof, G A (1991), “Procrastination and obedience”, The American Economic Review 81(2): 1-19.
Blasimme, A, and E Vayena (2020), “What’s next for COVID-19 apps? Governance and oversight”, Science 370(6518): 760-762.
Frimpong, J A and S Helleringer (2020), “Financial Incentives for Downloading COVID–19 Digital Contact Tracing Apps“.
Bowles, S (2020), “Policies designed for self-interested citizens may undermine” the moral sentiments”: Evidence from economic experiments”, Science 320(5883): 1605-1609.
Gneezy, U, and A Rustichini (2000), “Pay enough or don’t pay at all.” The Quarterly Journal of Economics 115(3): 791-810.
Thaler, R H and C R Sunstein. (2009), Nudge: Improving decisions about health, wealth, and happiness, Penguin.
Loi, M (2020), “How to fairly incentivise digital contact tracing”, Journal of Medical Ethics.
2 See https://csm.netlify.app/
4 See https://csm.netlify.app/