Ce que la COVID nous a appris sur les plateformes numériques et la survie des petites entreprises

Authors: Xiaolan Zhou Yasuyuki Sawada Matthew Shum Elaine S. Tan Abstract The impact of digital platforms on the implementation of the United Nations’ (UN) Sustainable Development Goals (SDGs), especially the business sustainability of micro, small and medium-sized enterprises (MSMEs), under different containment policies during and after the COVID-19 pandemic has not yet been studied in detail. Using detailed administrative data from Alibaba Group’s online on-demand food delivery platform, we found that the digital platform contributed to the food security, wellbeing, employment and business sustainability of MSMEs both during and after the pandemic. We uncovered merchants’ heterogeneous responses to the COVID-19 pandemic during and after the implementation of different containment policies in China. On the extensive margin, the period of complete lockdown left long-term scarring on online merchants by decreasing the number of entrants, especially in the cooked food industry. In contrast, on the intensive margin, chain stores, especially those with large chain networks or multiapp stores, exhibited stronger resilience than their counterparts during and after lockdown. Thus, specialization (at the outlet level) and wider coverage (at the network level) emerged as key factors that enable business sustainability under challenging economic circumstances. Introduction In the last decade, the global economy has been characterized by unprecedented digitalization. Digital platforms are not only transforming the ways in which we interact, communicate and transact goods and services but also helping economies recover from unexpected negative shocks. The growing trend of e-commerce and the broader digitalization of economies has been accelerated by the global COVID-19 pandemic (ADB , ; Du et al. ). In late January 2020, the Chinese government imposed three stringent containment policies in various cities to prevent the rapid spread of the disease (Fang et al. ; Qiu et al. ), which considerably reduced the transmission rate and spread of COVID-19 (Alekseev et al. ; Bartik et al. ); prolonged the effects on the ecosystems of micro, small and medium-sized enterprises (MSMEs; Bartik et al. ; Bloom et al. ; Mathijs de et al. ; Cong et al. ; Dai et al. , ; Guo et al. ; Kong et al. ; Wang et al. ); and facilitated the digitization of MSMEs in China. However, the role played by digital platforms in supporting the United Nations (UN) Sustainable Development Goals (SDGs), the impact of different containment policies on digital merchants’ business sustainability and the effectiveness of different business strategies on sustainable MSME resilience remain underinvestigated. To at least partly bridge this gap in the literature, this study employs large-scale administrative data from Alibaba Group’s Ele.me in China, which is an online on-demand delivery service platform for food and nonfood groceries; most of the shops on Ele.me are MSMEs. This paper presents a quantitative evaluation of the impact of different containment policies on all selected shops on Ele.me, on shops in different product categories (e.g., cooked food, uncooked food, pharmacy), and on individual shops in sequence. We discuss the role of digital food delivery platforms in achieving the SDGs during and after the pandemic and then propose policy implications for policy-makers and merchants. This study shows that the digital food delivery platform has contributed not only to the food security, wellbeing, employment and business sustainability of MSMEs both during and after the pandemic but also to supporting the SDGs. 2021a b 2022 2020 2020 2022 2020 2020 2020 2021 2022 2021a b 2022 2021 2020 The contributions of this study are as follows. First, it examines sustainable development (SD) from the perspective of MSMEs; this focus is critical to the SD of China’s economy because the proportion of persons employed by MSMEs in all enterprises was 79.4% in 2018 (NBS ). Furthermore, the SD of MSMEs is relevant to SDG target 8.3. The administrative data used herein, which is rarely accessible, provides a unique opportunity to quantitatively and comprehensively examine the long-term impact of different COVID-19 containment policies on the business sustainability of MSMEs and, hence, on SD in general. In addition to the number of orders or sales, which are usually examined, we construct proxies for product variety and price discounts. We employ quadratic time trends to track the potential nonlinearity of policy effects, while many previous studies have only adopted binary variables of different containment policies to estimate the average policy “treatment” effects (Raj et al. ; Han et al. ; Guo et al. ). Second, our case study is related to the food industry and pharmacy industry, where digitalization is seen through merchants’ use of online delivery platforms; the findings can shed light on the role of digitalization in food security and health care issues during crises, which are related to SDG targets 2.c and 3.8, respectively. Our unique dataset allows us to derive one measure of the extent of digitalization driven by COVID-19, namely, the number of MSMEs in the food industry that entered one digital food delivery platform in China, which is linked to SDG target 8.3. Although this measure is limited to one platform, it can provide insights into the impact of containment policies on digitalization. Our finding is that these policies have heterogeneous outcomes on different product categories, which are different when a longer time period is studied; therefore, the pandemic period has not had a straightforward relationship with digitalization, as it has been mediated by different containment policies. Third, we examine the impact of the two business strategies, i.e., chain stores and multiapp stores, on the resilience of sustainable MSMEs. For MSMEs that entered a platform (i.e., digitalized), we find that the effects of containment policies vary according to the merchant type. Chain stores are found to perform better than nonchain stores, suggesting that even when both store types are digitalized, a larger network of supply chains brings greater gains to digitalization. Similarly, multiapp stores (which indicate a higher level of digitalization than stores without multiple apps) are found to perform better. 2019 2023 2022 2022 The rest of this paper proceeds as follows. First, we provide a literature review focused on tackling the UN SDGs through digitalization and the impact of COVID-19 on the sustainability and resilience of MSMEs to exogenous shocks. Second, we illustrate the utilized materials and methods, including the evolution of COVID-19 in China, data, nonmodel facts, and empirical models. Third, we present empirical results on the impacts of various containment policies and two business strategies. Fourth, we summarize our findings and discuss implications for policy-makers and merchants. Fifth and finally, we close with recommendations for achieving the UN SDGs through digital food delivery platforms and the limitations of this study and future research. Literature Tackling the UN SDGs through digitalization In 2015, the United Nations General Assembly adopted the 2030 Agenda for Sustainable Development, which established 17 SDGs and 169 targets (UN , ). The implementation of SDGs faces many challenges, such as localization and integration in governance. Digitalization is regarded as one of the most promising transformations for achieving the SDGs (Gouvea et al. ; Sachs et al. ; Seele and Lock ; Walker et al. ). Digital transformation and networking can help localize SDGs (ElMassah and Mohieldin ; Reuter ) and enhance governance toward the achievement of SD (Athey ; Lowman ). With the spread of big data and information and communication (ICT) technology in all areas of life, digitalization has dramatically increased in many scenarios, such as working from home, online education, digital health (Novillo-Ortiz et al. ; Asi and Williams ), smart cities (Jain et al. ), robotics, and emission reduction and other pro-environmental solutions (He et al. ), all of which have contributed to achieving the SDGs, especially during and after COVID-19 (Pan and Zhang ; Zhao et al. ). However, the relationship between digitalization and sustainability remains controversial (Castro et al. ; Vinuesa et al. ). Digitalization may undermine people’s legal rights to privacy, lead to a misuse of data and political surveillance (Seele and Lock ), degrade natural ecosystems (Bekaroo et al. ) and negatively impact the environment and health (LaPointe ). Digitalization can alleviate income inequality (Yin and Choi ); however, whether digitalization reduces inequality also depends on how it is shaped (Asongu and Odhiambo ). Most of the extant studies have adopted a qualitative approach to analyze the impact of digitalization on the achievement of SDGs, while few studies have adopted a quantitative methodology. Most quantitative studies on SDGs examine the SDG index (SDGI) at the country level (Zhao et al. ). To the best of our knowledge, no study has focused on the mechanisms by which digital food delivery platforms have helped achieve the SDGs during and after the pandemic. 2015a b 2018 2019 2017 2019 2020 2023 2017 2017 2018 2020 2023 2023 2020 2022 2021 2020 2017 2016 2018 2022 2019 2022 Impact of COVID-19 on the sustainability and resilience of MSMEs Resilience refers to the economic system’s ability to absorb shocks and reorganize in the face of socioeconomic changes. Sustainability is a broader concept than resilience and represents a long-term goal, which is the integration of social, environmental, and economic performance that benefits current and future generations (Geissdoerfer et al. ; Ranjbar et al. ; Elkington ). Exogenous shocks, such as natural disasters, usually propagate through supply chains or supplier-customer networks of firms (Brummitt et al. ; Inoue and Todo ; Carvalho et al. ). Studies focusing on COVID-19’s impact on sustainability can be divided into three dimensions: environmental performance (Adelodun et al. ; Amankwah-Amoah ; Freire-González and Vivanco ), social performance (Anholon et al. ; Christoffel et al. ; Sharma et al. ; Ryan et al. ; Iyengar ), and economic performance (Pan and Zhang ). A considerable number of studies have adopted qualitative rather than quantitative approaches. However, many studies on the economic impact of COVID-19 on MSMEs have been based on surveys or transaction data from financial accounts (Bartik et al. ; Bloom et al. ; Mathijs de et al. ; Cong et al. ; Dai et al. , ; Guo et al. ; Kong et al. ; Wang et al. ), and they have examined various important topics, such as the revenue and costs of small businesses, business expectations, layoffs, closures, entry, number of firms, digitalization, financial crunches, and firms’ factor productivity. However, detailed administrative data on MSMEs are either seldom available for research or have been reported at a low frequency (e.g., quarterly) in surveys (Guo et al. ; Kong et al. ). For the restaurant industry, some researchers have indirectly inferred restaurant visits or staffing from cell phone geolocation data, webpage views and reservations (Wang et al. ; Glaeser et al. ); only a small number of studies have employed the actual counts of orders obtained from an online platform (Raj et al. ). Additionally, the impact of COVID-19 on MSMEs varies considering the firm’s location and characteristics, such as the area’s income, whether a restaurant is a chain restaurant, whether the restaurant is located in the city center, or whether the state has different political preferences (Wang et al. ; Glaeser et al. ). To the best of our knowledge, no previous study has employed comprehensive high-frequency merchant-level administrative data to quantify the long-term impact of containment policies on merchants, users, delivery riders, product variety and price discounts on digital food delivery platforms. To bridge this important gap in the literature, we examine how shops in the food industry and pharmacy industry have responded to the different containment policies during and after lockdown. 2017 2021 1998 2017 2019 2021 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2021 2022 2021a b 2022 2021 2020 2022 2021 2022 2021 2023 2022 2021 Materials and methods The evolution of COVID-19 in China Table shows the chronology of the COVID-19 pandemic in China starting from December 2019, when several cases of unexplained pneumonia were identified in Wuhan. Subsequently, the disease spread to the entire city, to Hubei Province, and to other parts of the country within a short period. On 23 January 2020, China took a pivotal step to curb the spread of the virus by completely shutting down movement out of Wuhan, followed by seven cities in Hubei. Partial shutdowns, checkpoints and quarantine zones were implemented in other Chinese cities. By mid-March, the number of new cases per day was contained within the single digits, and nationwide epidemic prevention and control measures had achieved considerable success. The lockdown in Wuhan was lifted on 8 April 2020. On 31 December 2020, the National Health Commission of the People’s Republic of China announced that the COVID-19 vaccine would be free for all citizens and that 300 million doses would be administered nationwide by 1 May 2021. As of 31 August 2021, more than 60% of China’s population had full vaccine protection. 1 Data Administrative data from Alibaba Group’s Ele.me and Koubei Ele.me offers different types of data: number of shops, active users, full-time and part-time delivery riders at the district-week level, operational data at the merchant-week level, and shop characteristics. Consumers can review menus and order food or other products for delivery or takeout from participating restaurants or stores using the Ele.me application, Ele.me’s Alipay portal or Ele.me’s web browser. Ele.me collects a commission on orders placed on the platform from restaurants or stores and delivery charges received from customers. Another application related to Ele.me is Koubei, which provides users with information and reviews of local lifestyle services. The takeaway services of Koubei shops are directed toward Ele.me through links on Koubei. Ele.me was combined with Koubei as a local life service company in October 2018, and its main competitor in China is Meituan, which is an app that focuses on local lifestyle services. In 2020 Q2, the nationwide market shares of Meituan and Ele.me were 68.2% and 25.4% (Trustdata ), respectively. 2020 We employed 104-week-long detailed administrative data from Ele.me that ranged from the fifth week of 2019 to the fourth week of 2021, i.e., one year before and one year after the outbreak of the COVID-19 pandemic. Shops with at least one order during the entire sample period were randomly selected. An “effective” shop in a given week was defined as a shop with at least one order. Time-invariant shop characteristics included shop creation date and whether a shop is listed on Koubei, indicating the relative significance of the shop’s offline presence and location at the city-district level. The time-varying data included shop operation data such as the number of orders, gross merchandise value (GMV), subsidies provided by a shop or the platform, number of Stock Keeping Units (SKUs), shop’s rating score given by customers (0–5), categories of goods sold, whether a shop is a chain store, and a unique chain brand identification number. We used the ratio of subsidies to the GMV as a proxy for price discounts. Usually, the nominal or listed prices on Ele.me do not vary considerably over time, and the GMV is computed using nominal prices and delivery fees. Therefore, the ratio of subsidies to the GMV is a good proxy for price discounts. We further adopted the number of SKUs as a proxy for product variety. These shops are classified into 204 “narrow” categories by Ele.me. We grouped these “narrow” categories into four mutually exclusive and exhaustive “broad” categories, as shown in Table : cooked food, uncooked food, pharmacy and others. Cooked food was the largest category in terms of the net GMV (GMV minus subsidies). The order of shares of the main categories from largest to smallest was cooked food, uncooked food, nonfood grocery store items, pharmacy and others. 2 We further augmented the Ele.me data with public information on the number of weekly confirmed COVID-19 cases and the start and end dates of policies for each city at different stringency levels (Du et al. ). For the former, we obtained the daily number of new cases from the National Health Commission at the city-day level and aggregated the number of new cases at the city-week level. 2022 Lockdown The stringency levels were threefold: complete, partial and checkpoint lockdowns. Under complete lockdowns, residents were prohibited from leaving the city and could only leave their homes for essential activities during limited periods; both public and private transportation were forbidden. Partial lockdowns were less stringent, and residents could leave the city. Although public transportation was shut down, private modes of travel were permitted. At the least stringent level, movement restrictions, such as checkpoints, were limited to localized pockets of disease outbreaks. Even in cities without any containment policy, at the start of the pandemic outbreak or in the early period of Wuhan’s lockdown, people were strongly encouraged to obey the stay-at-home order. Usually, all delivered packages were left at the entrance of a community in cities with containment policies. We focused on all seven cities for which data are available: Wuhan and Yichang in Hubei Province, Hangzhou in Zhejiang Province, Harbin in Heilongjiang Province, Suzhou in Jiangsu Province, and Chongqing and Xi’an in Shaanxi Province. These seven cities were selected for the following reasons. First, containment policies varied over time, and stringency levels varied across the seven cities. Wuhan and Yichang were under complete lockdown policies from 24 January 2020 to 7 April 2020. Partial lockdowns were experienced in Hangzhou from 4 February 2020 to 27 February 2020; and in Harbin from 4 February 2020 to 16 March 2020. Suzhou had checkpoints from 31 January 2020 to 23 March 2020, and Chongqing had checkpoints from 6 February 2020 to 19 February 2020. Xi’an never experienced containment policies during this period. Second, all cities except Yichang are provincial capitals or municipalities; Yichang is a small city. The six large cities are economically comparable in terms of population and GDP; the gross regional products of Wuhan, Yichang, Hangzhou, Harbin, Suzhou, Xi’an and Chongqing in 2021 were $275, $77.8, $281, $83, $352, $166 and $432 billion (1.772, 0.502, 1.811, 0.535, 2.272, 1.069 and 2.789 trillion RMB) respectively. Third, these cities or municipalities do not share borders; therefore, the potential spillover effects are weak. We exploited the exogenous variations created by the pandemic to quantify its causal impact. The market’s geographical size is defined as a district in a city because the majority of orders are within a district. Consumers usually place orders from shops within three kilometers of their places. Very few shops provide citywide delivery services; moreover, these orders take much longer, charge much higher delivery fees and are often rejected by shops when they are busy. Nonmodel facts Entrants The lockdown period witnessed a higher share of entrants listed on Koubei (defined as multiapp stores), chain stores and uncooked food (Fig. ). On the supply side, shops listed on Koubei (most of which relied heavily on offline services before the pandemic and were better known to consumers due to their listing on Koubei) temporarily closed their brick-and-mortar stores because of stay-at-home orders and mandatory nonessential business closures and thus shifted toward multichannel retailing. Furthermore, shops listed on Koubei already had digitalization experience; therefore, it was easier for them to join another digital platform 1