Catastrophic events can be powerful catalysers of change. Almost overnight, the COVID-19 pandemic triggered increased dependence on digital technologies, which played a crucial role in helping firms weather the worst of the shock (Comin et al. 2021). Unsurprisingly the pace of digitalisation appears to have accelerated considerably during this time (Apedo-Amah et al. 2020, Bellmann et al. 2021). However, little is known about the relative pace of transition across different countries and markets and whether COVID is widening or narrowing digital divides across countries.
Using a new source of data on nearly the universe of websites and the technologies embedded within them, our study (Ragoussis and Timmis 2022) documents the rapid growth in the adoption of three types of website technologies in 2020: e-commerce, online payments, and digital advertising.
A new database sheds light on the dynamics of global transition online
Our analysis primarily relies on BuiltWith, a commercially available dataset containing information on 150 million active websites around the world. The data has been scraped in a consistent manner since mid-2018 using the source-code information embedded in company websites. The latter is continuously crawled such that virtually every website in the world is scraped every one to four weeks.
We allocate websites to 185 countries in several steps, using relevant fields of information provided in the data: company addresses and telephone numbers, the language of text within a website and their top-level domain.1 To focus on websites with commercial applications, we consider only secure websites, meaning the data to and from the site can be encrypted (a pre-requisite for any financial transaction).2
Most websites are basic, containing only readily and freely available functionality, whereas technologies that allow online transactions are rare. BuiltWith measures the presence of over 30,000 website technologies across different technology providers, which it aggregates into broad categories based on their function. In December 2019, whilst 69% of websites were mobile-friendly, only 14% of websites had e-commerce functionality, and 11% of them allowed for electronic payments.
To measure the impact of COVID-19 on technology adoption, we combine BuiltWith data with monthly measures of the timing and intensity of COVID-19 lockdowns across countries in a formal event study setting. We compare countries with different timings of COVID lockdowns as an event study, with the event defined as the first month that stay-at-home requirements are imposed.
The Great Upgrade
The timing of the first COVID lockdowns in early 2020 strongly predicts faster website technology upgrading, particularly e-commerce and online payments (see Figure 1). Firms responded rapidly to countries imposing stay-at-home orders or workplace closure restrictions – within two or three months of the first lockdowns – by upgrading their websites to sell online. Six months after COVID lockdowns are imposed, these events account for about a 1.5 percentage point increase in online payment use and a 0.7 percentage point growth in e-commerce. The magnitudes are substantial, with these increases accounting for about a third of the total increase in e-commerce or e-payments diffusion over 2020. The growth of website technology use represents not only a step-change in adoption levels but a shift upwards in adoption trend after COVID lockdowns (Figure 1). This is consistent with the role of network effects in online markets and/or expectations that COVID will persist into the medium term (Anayi et al. 2021).
Figure 1 Acceleration in the adoption of online payments and e-commerce after the first COVID lockdowns
Notes: The event study is estimated via a fully saturated model of all leads and lags of the COVID lockdown event (D_iγ) with country and time (month) fixed effects. Event time coefficients are reported from three periods before to six periods after lockdowns. The sample contains observations from September 2019 until December 2020. As is standard practice, coefficients are reported relative to t-1, the month before COVID lockdowns. 95% confidence intervals are plotted around the estimated coefficients.
Source: Ragoussis and Timmis (2022)
The effects are robust to alternative measures of COVID lockdown stringency, workplace closure requirements, stay-at-home requirements, or the number of COVID deaths, all of which predict growth in e-commerce. For example, moving from zero to an 80% stringency lockdown – the mean value of the lockdown stringency index in April 2020 – is associated with a 1.3 percentage point increase in e-commerce adoption and a 1 percentage point increase in online payments adoption by the end of 2020.
But is the general upgrading masking a widening or narrowing of digital divides across countries? Unpacking the trend across regions and countries brings further insights. While globally, the use of online payments, e-commerce and digital advertising technologies increased by 34%, 19%, and 17%, respectively, over 2020. Countries with the fastest growth in these technologies already had less intensive use in 2019 (Figure 2), echoing some evidence from other sources (Alfonso et al. 2021). The difference is non-trivial. Countries with 4 percentage point lower initial levels of e-commerce use in 2019 (equivalent to moving from the median country to the 25th percentile, or 75th percentile to the median) experience a 2.6% higher growth in e-commerce during 2020. Similarly, countries with 3.5 percentage point higher initial online payments usage (equivalent to moving from the median to 25th percentile) have a 2% higher growth in online payment diffusion the following year. Furthermore, we find that most of the growth in e-commerce by laggard countries is driven by the adoption of free-of-charge functionalities.
Figure 2 Laggard countries have faster growth in technology adoption during 2020
Notes: Country-level percentage growth in technology adoption between December 2020 and 2019 is shown against initial levels of technology use in December 2019. The sample is a balanced panel of incumbent websites that exist in both December 2019 and December 2020. A linear best fit line is included for readability.
Source: Ragoussis and Timmis (2022)
Importantly, however, because the starting point of e-commerce and e-payment use has been low in many countries, the acceleration of technology adoption has not been sufficient to narrow the divides across countries. Absolute differences in use kept growing during the pandemic – especially for e-commerce – despite faster percentage growth at the bottom of the distribution. Overall, absolute convergence in less-used technologies like e-commerce requires an even faster pace of digital adoption in laggard countries.
COVID-19 provided a shock that shifted the trend in digital development, taking advantage of increasingly accessible, low-cost technologies for online transactions. A number of developments have allowed this to materialise, such as the rise of e-commerce platforms or off-the-shelf e-commerce websites, mobile money, and relatively widespread mobile broadband. However, the gaps along the income distribution remain substantial and growing, which points to needs in areas such as the scarcity of digital skills to run an online business, the availability of banking and logistics services, or challenges around the access and use of data. Further research in this area is key to addressing specific constraints firms and consumers face in transitioning their activities online and sustaining the growth of digital markets.
Alfonso, V, C Boar, J Frost, L Gambacorta and J Liu (2021), “E-commerce in the pandemic and beyond”, Bank of International Settlements Bulletin 36.
Anayi, L, N Bloom, P Bunn, P Mizen, G Thwaites, C Young (2021), “Covid-19 and structural change”, VoxEU.org, 28 October.
Apedo-Amah, M C, B Avdiu, X Cirera, M Cruz, E Davies, A Grover, L Iacovone, U Kilinc, D Medvedev, F O Maduko, S Poupakis, J Torres, T T Tran (2020), “Unmasking the Impact of COVID-19 on Businesses: Firm Level Evidence from Across the World”, World Bank Policy Research Working Paper No 9434.
Bellmann, L, P Bourgeon, C Gathmann, P Gleiser, C Kagerl, E Kleifgen, C König, U Leber, D Marguerit, L Martin, L Pohlan, D Roth, M Schierholz, J Stegmaier, A Aminian (2021), “The pandemic has boosted firm investments in digital technologies”, VoxEU.org, 05 August.
Comin, D (2022), “Technology and Resilience. Which firms coped best with Covid-19”, VoxEU.org.
Ragoussis, A, and J Timmis (2022), “Global Transition Online”, World Bank Policy Research Working Paper No 9951.
1 We exclude all websites with generic top-level domains, “.org” “.edu” “.net” “.com” or “.int”, that do not have addresses, telephone numbers or a unique country language. Since “.com” is the primary domain ending for US companies, we completely exclude the US from our analysis.
2 Specifically, we focus on websites with a Secure Sockets Layer (SSL) certificate.