Technological change, the productivity of formal and informal businesses, and the impact on labor market

Sodokin, Koffi; Djafon, Joseph Kokouvi; Couchoro, Mawuli; Kounetsron, Yao Mensah; Agbodji, Akoété Ega Article Technological change, the productivity of formal and informal businesses, and the impact on labor market Cogent Economics & Finance Provided in Cooperation with: Taylor & Francis Group Suggested Citation: Sodokin, Koffi; Djafon, Joseph Kokouvi; Couchoro, Mawuli; Kounetsron, Yao Mensah; Agbodji, Akoété Ega (2023) : Technological change, the productivity of formal and informal businesses, and the impact on labor market, Cogent Economics & Finance, ISSN 2332-2039, Taylor & Francis, Abingdon, Vol. 11, Iss. 2, pp. 1-46, https://doi.org/10.1080/23322039.2023.2268790 This Version is available at: https://hdl.handle.net/10419/304241 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Cogent Economics & Finance ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/oaef20 Technological change, the productivity of formal and informal businesses, and the impact on labor market Koffi Sodokin, Joseph Kokouvi Djafon, Mawuli Kodjovi Couchoro, Yao Mensah Kounetsron & Akoété Ega Agbodji To cite this article: Koffi Sodokin, Joseph Kokouvi Djafon, Mawuli Kodjovi Couchoro, Yao Mensah Kounetsron & Akoété Ega Agbodji (2023) Technological change, the productivity of formal and informal businesses, and the impact on labor market, Cogent Economics & Finance, 11:2, 2268790, DOI: 10.1080/23322039.2023.2268790 To link to this article: https://doi.org/10.1080/23322039.2023.2268790 © 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. Published online: 20 Oct 2023. Submit your article to this journal Article views: 1503 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=oaef20 DEVELOPMENT ECONOMICS | RESEARCH ARTICLE Technological change, the productivity of formal and informal businesses, and the impact on labor market Koffi Sodokin *, Joseph Kokouvi Djafon , Mawuli Kodjovi Couchoro , Yao Mensah Kounetsron and Akoété Ega Agbodji ABOUT THE AUTHORS Koffi Sodokin is an Associate Professor at the Faculty of Economics of the University of Lome and member of the Research Center for Applied Economics and Management of Organisations (CREAMO). He has over ten years of experience teaching and supervising MSc students. His research focuses on money, finance, macroeconomics, and microeconomics of development. Joseph Kokouvi Djafon is a research assistant in the Research Center for Applied Economics and Management of Organisations (CREAMO) at the University of Lome-Togo. He has over 1 years of experience in field research focused on African countries. His research area includes Micro economics, Development Economics, Banks and Finance. Mawuli K. Couchoro is a Full Professor and the Dean of the Faculty of Economics at the University of Lome and member of the Research Center for Applied Economics and Management of Organisations (CREAMO). He has over 15 years of teaching and research experience and supervision of Ph.D. and MSc students. His research interests include money, finance, macroeconomics, and microeconomics of development. Yao Kounetsron is a Full Professor at the Institute of Enterprise Administration (IAE) of the University of Lome and member of the Research Center for Applied Economics and Management of Organisations (CREAMO). He has over 20 years of teaching and research experience and supervision of Ph.D. and MSc students. His research interests include accounting, finance and management of organisations. Akoété Ega Agbodji is a Full Professor at the Faculty of Economics of the University of Lome and member of the Research Center for Applied Economics and Management of Organisations (CREAMO). He has over 20 years of teaching and research experience and supervision of Ph.D. and MSc students. His research interests include microeconomics of development. PUBLIC INTEREST STATEMENT Technological innovations, such as e-commerce, are transforming economies worldwide, but their impacts can be complex and uneven across sectors. This study provides valuable insights into how online business adoption affects companies and employment in Togo, focusing on the differences between the formal and informal sectors. This research reveals that engaging in online commerce substantially increases productivity, especially in informal enterprises. This highlights the potential of digital tools to enhance competitiveness. However, the study also shows that online businesses can reduce labor demand, thereby threatening job losses. These findings have important policy implications. They indicate that online platforms and mobile technology can empower informal businesses and promote inclusive growth as Internet access expands in developing countries. However, the research also underscores the need for strategies to smooth economic transition through training programs and social protection. As commerce moves online globally, understanding its multifaceted impact is crucial. This study advances the knowledge on how e-commerce adoption influences productivity, employment, and inequality across the formal and informal sectors. Evidence can guide policies to maximize the benefits of digital transformation and ensure that its gains are broadly shared. Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 1 of 46 Received: 24 July 2023 Accepted: 05 October 2023 *Corresponding author: Koffi Sodokin, Research Center for Applied Economics and Management of Organisations (CREAMO), University of Lome, Lome 1515, Togo E-mail: [email protected] Reviewing editor: Goodness Aye, Agricultural Economics, University of Agriculture, Makurdi Benue State, Nigeria Additional information is available at the end of the article © 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. Abstract: Digital transformation, both omnipresent and influential, has deeply impacted various sectors with a particular focus on the economy. The introduction and adoption of tools that facilitate technological changes and online business practices have emerged as game changers. They have endowed corporations with enhanced internal agility and improved employee communications. Online business, with its vast potential, is becoming increasingly crucial in developing countries, where Internet accessibility is steadily growing. This study explores the impact of e-commerce on company productivity by considering both formal and informal sectors. It leverages data from the 2018 General Business Census of Togo. By applying endogenous switching regression and smoothed instrumental variable quantile regression tools, this study demonstrates that online businesses can significantly increase productivity, particularly in firms within the informal sector. However, this finding highlights the potential risk of job loss. This study concludes that support strategies are essential for promoting the integration of online companies, increasing productivity, and protecting jobs. Subjects: Development Studies; Economics and Development; Economics Keywords: online business; productivity; digital transformation, employment JEL Classfiication: L25; O33; M15; O55 1. Introduction In recent decades, Information and Communication Technology (ICT) has significantly affected various sectors including the economy. However, quantifying its effect on competitiveness and performance remains challenging (Keček et al., 2019; Neirotti & Pesce, 2019; OECD, 1998). The impact of ICT on various sectors, including the economy, has been widely recognized . However, quantifying its effect on competitiveness and performance remains a challenge (Barsoum & Elfeky, 2017; Frank et al., 2018; OECD, 1998; Swamy, 2020). With the rapid increase in internet usage and online business adoption, digital transformation has become crucial to Africa’s economy, especially in the informal sector. Automation and digitalization have the potential to affect employment and exacerbate inequalities, particularly among low-skilled workers (Frank et al., 2018). Although the influence of digital transformation is acknowledged in global economies, its implications for the African informal sector remain largely unexplored. Frank et al. (2018) examine the impact of automation on employment in urban areas. They found that small cities may face greater adjustments, such as worker displacement and job content substitutions, whereas large cities exhibit increased occupational and skill specialization, reducing the potential impact of automation (Frank et al., 2018). Frank et al. (2018) also demonstrated the connection between urban agglomeration and automation's influence on employment, providing empirical evidence for this relationship. Another study by Autor and Salomons (2018) investigated the labor-displacing effects of automation and productivity growth. They find that the labor share-displacing effects of productivity growth have become more pronounced over time, suggesting that automation has become less labor-augmenting and more labor-displacing (Autor & Salomons, 2018). However, comprehensive evidence of the labor-displacing channel of automation remains limited (Autor & Salomons, 2018). In terms of methodology, Frank et al. (2018) use task groups to assess the resilience of different tasks to job displacement from automation in cities (Frank et al., 2018). They also utilized alternative estimates for the probability of job automation provided by the Organization for Economic Cooperation and Development (OECD) (Frank et al., 2018). Autor and Salomons (2018) employed data on industries and countries to estimate the employment and labor share impacts of productivity growth and automation (Autor & Salomons, 2018). Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 2 of 46 With the rapid increase in internet usage (World-Bank, 2016, 2020) and online business adoption (Swamy, 2020), digital transformation has become crucial in Africa’s formal and informal sector (Hootsuite, 2019). Arntz et al. (2016) argue that automation and digitalization may not destroy many jobs in the Organization for Economic Co-operation and Development (OECD) countries but could exacerbate inequalities and disproportionately affect low-skilled workers. While the influence of digital transformation is clearly recognized in global economies, its specific implications for the African informal sector remain largely unexplored. Rapid technological changes in Internet use and the adoption of online businesses have raised critical questions regarding the potential impact of digital transformation on business performance and employment. This study investigates the impact of ICT on business productivity and employment in Togo by asking how online activities affect companies and labor in the formal and informal sectors. The study presents four key findings. First, online operations increase productivity in informal sector firms, with permanent employees having a negligible impact. Second, firms that do not engage in online marketing experience negative turnover variations, thus emphasizing the need for greater online adoption. Third, the impact of online business varies across sectors and productivity variations, with informal sector firms benefiting from higher productivity. Internet access plays a critical role in a firm’s productivity, particularly in the informal sector. Finally, additional investigations show that companies operating online require fewer jobs than those that do not. The adoption of online business had a minimal impact on labor supply across all firms and sectors. This was also the case for informal sector enterprises, where formal sector enterprises recorded a substantial increase. This study offers three main contributions. First, it establishes a connection between the adoption of online business and the stages of innovation diffusion (Rogers, 1962), outlining potential impacts on productivity and employment. This perspective provides a theoretical framework for understanding the dynamics of the link between new technology adoption and the economy using survey data from Togo. Second, it underscores the role of e-commerce in promoting inclusive growth and narrowing the digital divide, highlighting the significance of mobile and digital technologies for business productivity and the labor market. Third, by examining both the formal and informal sectors, this study deepens our understanding of online business adoption and its impact on firm productivity, especially in the informal sector and employment, offering valuable insights for policymakers and business leaders. 2. A brief review of the literature Previous studies have analyzed the impact of online business adoption on company productivity from different perspectives. Some researchers have focused on economic functions, with Dewan and Min (1997) and Konana et al. (1999) exploring how online businesses can enhance efficiency and lower costs. Others, such as Raymond and Bergeron (1996) and Kekwaletswe (2015), have examined the effects of information and communication technology (ICT) investments on company productivity and organizational change. Two major research streams have emerged: one centered on organizational change and productivity improvements enabled by ICTs (Jorgenson & Stiroh, 1999), and another investigating how ICT investments drive organizational change within companies (Leavitt & Whisler, 1958). 2.1. Online business and its impact on firm productivity Numerous studies have investigated the relationship between online business adoption and productivity gains. Research shows that enhancing website performance (Bilgic & Duan, 2019) and utilizing automation (Păvăloaia & Necula, 2023) can reduce costs and improve the user experience. Advanced technologies such as the Internet of Things (IoT), big data, and artificial intelligence (AI) can optimize supply chains (Misra et al., 2020) and streamline business processes (Daskalakis & Golowich, 2022). Further productivity improvements stem from innovations in urban logistics (Cano et al., 2022; Rai & Dablanc, 2022), blockchains (Taherdoost & Madanchian, 2023), website accessibility (Najadat et al., 2021), and rural e-commerce opportunities (Ballerini et al., Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 3 of 46 2023). User interfaces and usability are critical components (Lesage, 2015). The rapid growth of e-commerce (Eurostat, 2019), highlighted during the COVID-19 pandemic (WTO, 2020), provides opportunities for traditional retailers (Xia & Monroe, 2010) and enables enhanced business performance (Cosgun & Dogerlioglu, 2012) through improved information management (Damanpour & Damanpour, 2001). Consumers benefit from greater product variety, time savings, and lower prices (Khan, 2016), whereas society benefits from reduced traffic, lower air pollution, and increased rural access (Shahriari et al., 2015). Digital transformation influences several aspects of business and company management (Kraus et al., 2022). Digital innovation and the adoption of management software are significant drivers of this evolution (Endres et al., 2022). The influence of digital transformation extends to enhancing an organization’s ability to respond to market turbulence through the establishment of a digital technology infrastructure (Li et al., 2021). The success of digital transformation is also closely tied to business and management commitments, with IT departments playing a supporting role (Ko et al., 2021). Investments in digital technologies, employee digital skills, and digital transformation strategies have been identified as vital for improving performance and sustainability, especially in small and medium-sized enterprises (SMEs) (Teng et al., 2022). Furthermore, digital transformation has been shown to promote green innovation in enterprises, reflecting its broader societal and environmental impacts of digital transformation (Feng et al., 2022). The journey of digital transformation is not without challenges, and organizations face various obstacles in becoming digitally transformed. These challenges can be addressed through a clear understanding of the meaning of digital transformation and implementation of potential solutions (Shahi & Sinha, 2020). Hypothesis 1: The adoption of online business activities has a positive effect on company productivity (Damanpour & Damanpour, 2001; Xia & Monroe, 2010; Kraus et al., 2022). 2.2. Potential limitations of online business influence on business productivity The literature presents mixed findings regarding the effects of online business adoption on firm productivity. Some studies suggest potential obstacles, with Leavitt and Whisler (1958) noting that the Internet and ICTs could harm small businesses. Other research points to challenges such as management issues (Farooq et al., 2019; Thaichon et al., 2018), the need for digital transformation (Hategan et al., 2021), sustainability concerns (Oláh et al., 2018), privacy risks (Maseeh et al., 2021; Rita & Ramos, 2022), data management (Akter & Wamba, 2016), and various consumer attitudes (Alrousan & Jones, 2016; Rosário & Raimundo, 2021). Additional barriers include adoption challenges for SMEs (Abed et al., 2015; Sila, 2015; Sin et al., 2016; Viu-Roig & Alvarez-Palau, 2020), differing levels of customer loyalty (Mangiaracina et al., 2015; Wang et al., 2020) and emotional intelligence (Huang et al., 2021). However, business performance can be significantly improved through commercial websites and online marketplaces (Davies et al., 2019). Retail innovation continues (Pantano & Priporas, 2016), although research on social media marketing effectiveness remains mixed (Kapoor et al., 2018; Kartika, 2021). The transition to online commerce can disrupt consumer habits and destabilize companies during the adaptation period. This disruption is multifaceted and can be understood through various lenses. First, the perceived risk associated with online shopping, especially during unprecedented events such as the COVID-19 pandemic, can create a barrier to consumer acceptance of online commerce, leading to disruptions in traditional shopping habits (Habib & Hamadneh, 2021). Second, the virtual store environment itself can alter consumer behavior, as real consumers may react differently in online settings than in physical stores, leading to unexpected changes in purchasing patterns (Dahlen & Lange, 2002). Third, factors such as product type and individual consumer characteristics can significantly influence the intention to shop online, adding complexity to the transition from offline commerce to online commerce (Chiang & Dholakia, 2003). Fourth, the equilibrium between offline and online selling channels can be affected, as altering consumer Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 4 of 46 preferences for online purchasing shows a shift in traditional supply chain dynamics, posing challenges to companies in maintaining balance (Yu et al., 2015). Fifth, the strategic behavior of e-commerce businesses, particularly in industries such as electronics, can influence factors such as price competition and service quality, further contributing to the destabilization of traditional business models (Svobodová & Rajchlová, 2020). Together, these factors paint a complex picture of the transition to online commerce, in which both consumer habits and business stability can be significantly affected during the adaptation period (Wang et al., 2020). Hypothesis 2: The adoption of online commerce may have a negative or limited effect on the productivity of certain companies (Abed et al., 2015; Endres et al., 2022) and a disruptive effect of social networks (Qu et al., 2013; Song et al., 2022). 2.3. The effect of internet and online business on labor market The impact of online businesses on the labor market has been a major research focus. Studies have revealed structural changes in retail, job losses, and the need to adapt to e-commerce (Terzi, 2011). Research has found mixed effects, with some studies emphasizing job creation (Américo & Veronico, 2018; Paul et al., 2022) and others highlighting the displacement and importance of retraining. Employment effects vary across developing countries (Gherghina et al., 2021), and are heavily dependent on the local context and policies in place (Li et al., 2023). Training workers is essential for adapting to the evolving labor market driven by e-commerce expansion. Some scholars caution that digital transformation through online business and Internet adoption could negatively impact informal sectors in developing economies (Leavitt & Whisler, 1958). Despite these potential benefits, adopting digital technologies may render some traditional roles obsolete or require significant change (Autor et al., 2003). Trends in automation, artificial intelligence, and machine learning can displace human workers, particularly in roles susceptible to automation (Frey & Osborne, 2017). This may result in job loss and higher unemployment, particularly among workers with limited skills and education (Acemoglu & Restrepo, 2018). Comparative analyses of OECD countries further highlight the need to understand comprehensively how digital transformation can impact employment (Arntz et al., 2016). The rise of interconnected entrepreneurial ecosystems offers new opportunities in the context of digital transformation (Barykin et al., 2020; Candelo et al., 2021). According to Bouncken and Kraus (2022), this evolution is characterized by an interconnection between various stakeholders in an ecosystem, such as governments, the private sector, society, universities, and entrepreneurs. These actors work together to create a social and economic environment conducive to innovation and entrepreneurship (Hernández-Chea et al., 2021; Komninos et al., 2021). This means that companies are no longer solely focused on distinguishing themselves individually from their competitors but also rely on shared resources, network externalities, knowledge transfers, and government support. Digital transformation plays a key role in the interconnection of entrepreneurial ecosystems (Stroumpoulis & Kopanaki, 2022). Technological advancements, such as increased internet connectivity and high-speed broadband connections, facilitate the design and testing of new technologies (Feng et al., 2022; Xue et al., 2022). Moreover, information and communication technologies enable stronger links between resources and actors at the local, regional, and international levels (Shahi & Sinha, 2020; Soto-Acosta, 2020). The digitization of business processes offers new opportunities and imposes new challenges for companies, such as the need to develop digital knowledge and rethink business models (Furr et al., 2022; Ritala et al., 2021). Digitization promotes collaboration and complementarity between companies within ecosystems while allowing for rapid feedback and autonomous digital processes (Secundo et al., 2020). It also enables customers to play a more active role in defining the demand. However, some traditional sectors may be destabilized, requiring adaptation (Song et al., 2022). Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 5 of 46 Hypothesis 3a: Online commerce leads to job losses in certain sectors (Frey & Osborne, 2017; Bouncken & Kraus, 2022). Hypothesis 3b: Online commerce creates new jobs in other sectors (Américo & Veronico, 2018; Kusi-Appiah & Essandoh, 2023). This brief literature review clearly shows that online businesses can enhance their business productivity. However, this study has several limitations. The impact on employment can also be mixed and context dependent. Therefore, the present study was conducted. Additional investigations using microdata from Togo better understand the ins and outs of these modern facts in the African context, where economies are characterized by the coexistence of formal sectors and many economic microstructural activities (the so-called informal sector) (Sodokin, 2007; Sodokin et al., 2023). 3. Methodological approach 3.1. Theoretical model and assumptions Drawing from Rogers (1962) innovation diffusion model, companies adopting online business as innovation are risk-takers, aiming to enhance competitiveness. Online business speed depends on factors, such as the complexity of the online sales process, setup costs, and market competition. Early adopters may gain a competitive advantage, whereas latecomers may face market-share losses and higher entry costs. To model the impact of online business on productivity and the labor market, we include in the Rogers(1962) innovation diffusion model a binary variable OP (online business), which takes the value of one if the company has adopted online business and zero otherwise. Based on Rogers (1962) innovation diffusion model, we formulate a theoretical model of the link between online commerce and productivity, as follows: where C0t represents the productivity of the company after adopting online business. To model a company’s decision to adopt a digital tool for online business, we use the following logistic function: where P represents the probability that a company adopts an online business, X represents the factors that influence the adoption decision, such as the complexity of online sales, the cost of setting up an online business website, and competition in the market, and β is a regression coefficient that measures the impact of these factors on the adoption probability. Finally, we combined these two formulations to obtain an expression for the variation in productivity depending on the online business decision. To combine the two formulations, we express productivity change as a function of the binary variable OP, the probability of adoption (P), and factors that influence the adoption decision (X). We introduce parameter α, which represents the impact of online business on productivity changes. The combined expression can be written as Substituting the logistic function for P, we get: Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 6 of 46 To simplify, the expression of ΔC by grouping the terms Ct1, equation (4) becomes: This expression shows the changes in productivity ΔCð Þ as a function of the binary variable OP, the factors that influence adoption decision Xð Þ, the impact of these factors on adoption probability βð Þ, and the effect of adoption on changes in productivity αð Þ. By analyzing the values of α and β and the specific factors in X, a company can better understand how its productivity may change based on its decision to adopt an online business. In summary, the model incorporates a binary variable, OP to model the impact of e-commerce adoption on revenue variation . The combined expression represents productivity variation as a function of OP, the probability of adopting P, and the factors influencing the decision to adopt an online business. To analyze the changes in labor productivity ΔCð Þ as a function of the binary variable, we first take the partial derivatives of equation (5) with respect to OP. Taking the partial derivative with respect to OP, we treat the other variables as constants. Equation (6) calculates the partial derivative with respect to OP. This allowed for the analysis of the marginal effect of the variation in OP on ΔC. The derivative of αOP with respect to OP is α: Equation (7) is the derivative of a linear function, which is constant and equal to coefficient α:. This simplifies the first term in Equation (6). For the second term, we use the chain rule for derivatives:. Equation (8) is the composite function derivation rule for the second term in Equation (6), where F is the following function: Equation (9) defines the composite function F for applying the derivation rule of Equation (8). Because F does not depend on OP, its derivative with respect to OP is zero: In equation (10), the derivative of the constant is zero, and Because F does not depend on OP, its derivative with respect to OP is zero: Thus, the derivative of the second term is simply: Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 7 of 46 Table 1. Descriptive statistics Variables Description Mean Stand. Dev Minimum Maximum Changes in labor productivity Changes in turnover between 2017 and 2016 per worker −.1011 2.5576 −29.3199 24.8385 *Turnover in 2017 The logarithm of the turnover in 2017 12.2888 3.4768 0 26.1230 *Turnover in 2016 The logarithm of the turnover in 2016 12.8632 3.0965 0 25.4499 Formal Formal (1 if the firm operates in informal sector and 0 if it operates in the informal sector .1491 .3561 0 1 Online business operation Online business operation (1 if the company performs business operations on the internet and 0 otherwise) .0562 .2303 0 1 Internet Detention of an internet connection (one if the company owns one and 0 otherwise) .0690 .2535 0 1 Age of the firm Age of the firm 5.3391 7.1584 0 206 Age of the manager Age of the manager 39.0494 15.1499 12 99 Instruction of the manager Education of the manager (one for primary, 2 for secondary, 3 for superior, and 0 otherwise) 1.5728 .8525 0 3 Region An area where the firm is located (1 for Lomé and 0 otherwise) .6289 .4831 0 1 Bookkeeping Bookkeeping (one if the firm has bookkeeping and 0 otherwise) .2365 .4249 0 1 Permanent staff The permanent staff (logarithm of permanent staff) .9043 .4717 0 7.2896 Weekly work hours The logarithm of the weekly work hours 3.9076 0.6818 0 5.1240 The size of firm in 2016 The size of the firm in 2016 (2 for large firm, 1 for medium firm, and 0 for small firm) .0369 .2110 0 2 Access to technology issue Access to the technology issue if the firm has issued to access to technology and 0 otherwise) .1778 .3823 0 1 Concurrence issue The concurrence issue (1 if firm ha issues with competition and 0 otherwise) .6201 .4854 0 1 Transport infrastructure issues The transport infrastructure issue (one if the firm has issues with transport infrastructure and 0 otherwise) .2591 .4382 0 1 (Continued) Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 14 of 46 less likely to steer their firms toward online businesses, as are largeand medium-sized companies. Thus, conducting business transactions online is predominant in the domain of small companies and this phenomenon is most pronounced among firms operating in the informal sector. 4.2. The impact of online business adoption on changes in firm productivity Table 4 summarizes the firms’ marginal gains when conducting online businesses. We examine these gains at the firm level and separately for firms in formal and informal sectors. Column (1) of Table 4 presents the treatment effects for firms conducting online business. Column (2) displays the treatment effects for firms that do not engage in online business. Finally, Column (3) estimates the treatment effects for all firms regardless of their online business activities. Considering the Average Treatment Effects on the Treated (ATTs), informal sector firms benefit the most from online business, as their changes in labor productivity have more than tripled. Formal-sector firms nearly tripled their labor productivity. This result demonstrates that online business operations have varied impacts depending on whether a firm operates in the formal or informal sector, which is corroborated by other treatment effects. There are several possible explanations for this observation. One possibility is that online businesses allow firms to reach a wider audience and sell their products and services to a larger number of customers. This can lead to increased sales and profits (Benner & Waldfogel, 2020). Another possibility is that online business allows firms to operate more efficiently. For example, firms can use the internet to automate tasks, communicate with suppliers and customers, and manage their inventories. This could lead to lower costs and higher profits. The findings of this study suggest that the Internet is a valuable tool for businesses (Munirathinam, 2019; Wang et al., 2020). Firms that use the internet to conduct business may be able to increase their productivity and profits. Furthermore, the study finds that the impact of online business transactions on productivity is greater for informal enterprises than formal enterprises. This is likely to be because informal firms are small and have limited resources. Variables Description Mean Stand. Dev Minimum Maximum Issue to access credit Issue to access credit (one if the firm has issue of access to credit and 0 otherwise) .3652 .4815 0 1 Age of the firm * Instruction of the manager Crossed variable between age of the firm and instruction of the manager 8.5809 15.2813 0 525 Instruction * Technology Crossed variable between Instruction of the manager and access to technology issue .3150 .7639 0 3 Region * Infrastructure Crossed variable between region and transport infrastructure issue .1574 .3642 0 1 Bookkeeping * Technology Crossed variable between bookkeeping and access to technology issue .0438 .2046 0 1 Staff * Concurrence Crossed variable between permanent staff and concurrence issues .5499 .5586 0 7.2027 Note. * We took the log (1+X) to deal with zero values. Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 15 of 46 Table 2. Difference in means results All firms Formal firms Informal firms Variables Online business operation (1) No Online business operation (2) Diff mean (2)-(1) Online business operation (1) No Online business operation (2) Diff mean (2)-(1) Online business operation (1) No Online business operation (2) Diff mean (2)-(1) Mean Mean Mean Changes in labor productivity .134 −.123 −0.257*** .111 −.002 −0.113 .157 −.141 −0.297** Turnover in 2017 13.420 11.823 −1.597*** 15.624 13.826 −1.798*** 12.331 11.685 −0.645*** Turnover in 2016 14.117 12.538 −1.579*** 15.778 14.173 −1.605*** 13.029 12.387 −0.642*** Formal .341 .067 −0.275*** - - - - - - Internet .565 .039 −0.525*** .746 .156 −0.589*** .471 .031 −0.440*** Age of the firm 6.975 5.219 −1.756*** 1.207 7.974 −2.233*** 5.305 5.023 −0.282** Age of the manager 4.990 38.926 −2.064*** 48.192 46.129 −2.063*** 37.275 38.413 1.138*** Instruction of the manager 2.307 1.530 −0.777*** 2.706 2.132 −0.574*** 2.101 1.487 −0.614*** Region .781 .585 −0.197*** .836 .781 −0.055*** .753 .571 −0.182*** Bookkeeping .544 .150 −0.394*** - - - .283 .087 −0.196*** Permanent staff 1.322 .871 −0.451*** 1.870 1.324 −0.547*** 1.039 .838 −0.200*** Weekly work hours 3.876 3.911 0.035*** 3.840 3.895 0.055*** 3.890 3.912 0.022** The size of firm in 2016 .199 .015 −0.184*** .439 .124 −0.316*** .016 .004 −0.012*** Access to technology issue .320 .168 −0.152*** .318 .215 −0.103*** .321 .164 −0.156*** (Continued) Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 16 of 46 Table 2. (Continued) All firms Formal firms Informal firms Variables Online business operation (1) No Online business operation (2) Diff mean (2)-(1) Online business operation (1) No Online business operation (2) Diff mean (2)-(1) Online business operation (1) No Online business operation (2) Diff mean (2)-(1) Mean Mean Mean Concurrence issue .658 .617 −0.041*** .740 .714 −0.026* .618 .610 −0.008 Transport infrastructure issues .349 .253 −0.096*** .337 .288 −0.049*** .354 .250 −.104*** Issue to access credit .459 .359 −0.101*** .475 .405 −0.070*** .451 .355 −0.096*** Age of the firm * Instruction of the manager 17.362 8.063 −9.300*** 28.181 17.585 −10.596*** 11.799 7.386 −4.412*** Instruction of the manager * Technology .762 .283 −0.478*** .879 .505 −0.374*** .705 .267 −0.438*** Region * Infrastructure .275 .149 −0.126*** .299 .217 −0.083*** .264 .144 −0.120*** Bookkeeping * Technology .181 .034 −0.147*** .318 .215 −0.103*** .113 .021 −0.093*** Staff * Concurrence .858 .528 −0.331*** 1.411 .950 −0.461*** .588 .496 −0.092*** Notes. *** p< 0.01, ** p< 0.05, * p< 0.1 Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 17 of 46 Table 3. Estimates of the endogenous switching regression model. Determinants of online business operation and firms change in labor productivity All firms Formal firms Informal firms Variables Selection No Online business operation Online business operation Selection No Online business operation Online business operation Selection No Online business operation Online business operation Turnover in 2016 0.2402*** 0.1933*** (0.0069) (0.0089) Permanent staff 0.1984*** 0.1468*** 0.4855*** (0.0339) (0.0371) (0.0745) Age of the manager −0.0052*** −0.0038** −0.0076*** (0.0012) (0.0015) (0.0019) The size of firm in 2016 −0.4721*** −0.3849*** 0.1227 −1.1084** −.0603 (0.0599) (0.0596) (0.3676) (.5404) (2.6653) Age of the firm .0405 .1045 .0319 .0766 .1379** .0211 (.0269) (.0712) (.0423) (.0740) (.0653) (.2739) Instruction of the manager .0994** .4470** .1542 .1663 .1858** .5144 (.0484) (.1974) (.1570) (.2540) (.0752) (.4859) Age of the firm * Instruction of the manager −.0078 −.0283 −.0066 −.0201 −.0632** −.0322 (.0107) (.0246) (.0168) (.0256) (.0300) (.1881) Region .0952 −.3142 −.7800 1.0247 .0796 −.5834 (.0698) (.3596) (.5025) (.8443) (.0696) (.4642) Access to technology issue −.4575*** .7989 −1.0963 .5113 −.4186** 1.7384* (.1705) (.6762) (.7076) (1.1537) (.1781) (.9781) Internet .1518 .3911** .4992** .0528 −.1887 .6473** (.1240) (.1939) (.2445) (.2506) (.1592) (.3032) Concurrence issue −.2342** .0776 −.1872 .2224 .0870 −1.3682** (Continued) Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 18 of 46 Table 3. (Continued) All firms Formal firms Informal firms Variables Selection No Online business operation Online business operation Selection No Online business operation Online business operation Selection No Online business operation Online business operation (.0972) (.2609) (.2904) (.3427) (.1336) (.6036) Bookkeeping .3010*** −.0196 .0338 −.7159 (.0815) (.2679) (.1148) (.4549) Transport infrastructure issues −.0392 −.1767 −.2631 1.0827 −.0644 −.2196 (.1096) (.5292) (.7312) (1.2473) (.1084) (.6512) Access to credit −.0515 −.0396 .0041 .1214 −.0592 −.1948 (.0582) (.1857) (.1767) (.2235) (.0616) (.3012) Instruction of the manager * Technology .1843** −.6877** .4456 −.2097 .1648* −1.1096** (.0936) (.2996) (.2855) (.4121) (.0999) (.4607) Region * Infrastructure −.1233 .2919 .1575 −1.4049 −.1473 .6966 (.1330) (.5684) (.7579) (1.2710) (.1369) (.7398) Staff * Concurrence .3577*** .2664** .3773** .1218 −.0426 2.0901*** (.0875) (.1277) (.1516) (.1235) (.1489) (.6003) Bookkeeping * Technology .0998 .8824** .2821 1.5016** (.1770) (.4419) (.2553) (.6996) Constant −4.2753*** −.0878 −3.5286*** −3.3635*** 1.4594** −3.1940*** −1.7607*** −.4102** −9.4308*** (0.0890) (.1080) (.5941) (0.1468) (.6278) (1.0972) (0.0878) (.1627) (1.2069) Rho 0 .833 .919 .181 Rho 1 .527 .518 .861 Sigma 0 2.857 3.992 2.524 (Continued) Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 19 of 46 Table 3. (Continued) All firms Formal firms Informal firms Variables Selection No Online business operation Online business operation Selection No Online business operation Online business operation Selection No Online business operation Online business operation Sigma 1 2.943 2.453 5.032 LR independence test of the equation χ2 = 698.44 Pvalue=.000 χ2 = 287.23 Pvalue = .000 χ2 = 30.31 Pvalue = .000 The overall significance test (Wald chi2) χ2 = 146.45 Pvalue = .000 χ2 = 40.55 Pvalue = .000 χ2 = 33.84 Pvalue = 0.004 Observations 9,835 9,835 9,835 1,719 1,719 1,719 8,116 8,116 8,116 Notes. The standard deviations are shown in parentheses. *** p< 0.01, ** p< 0.05, * p< 0.1. Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 20 of 46 When evaluating the Average Treatment Effects on the Untreated (ATUs), there is a negative average variation in labor productivity for firms not conducting business online. This effect is strongest for informal sector firms, as they have lost more than twice the change in productivity compared to previous years. Finally, when estimating the impact of online business operations on all firms, a negative average treatment effect (ATE) is observed. Thus, although many firms conduct business online, their gains do not offset the losses incurred by firms that do not engage in online businesses. This highlights the need to encourage more formal and informal sector firms to conduct business online as the potential gains indicated by ATTs are substantial. These results highlight the overall positive influence of online business operations on cahnges in firms’ productivity in Togo. However, the impact varies across formal and informal businesses and their initial productivity levels. To further analyze this, it is appropriate to examine the distribution of productivity gains across different quantiles in more detail using instrumental variable quantile regressions. 4.3. Smoothed instrumental variable quantile regression of the impact of online business operations on firm’s productivity 4.3.1. Lorenz curve of the distribution of changes in productivity Figure 1(a.a) represents the dynamics of productivity dispersion across all companies; the blue points seem to fluctuate around zero. This suggests that the individual productivity of all companies varies considerably, and is negative for some. This could be due to factors such as variations in company performance, economic shocks, management issues, or seasonal fluctuations. The red curve, representing cumulative productivity, shows a general upward trend for all the companies in the sample, although it seems to have struggled to exceed zero. This suggests that the total productivity of businesses increases over time; however, this increase is slowed by the presence of companies with negative productivity. This could reflect a compensation effect in which productivity gains from some companies are offset by productivity losses from others. In Figure 1(a.b), individual productivity appears to have a greater dispersion in formal companies than in all companies, indicating greater variability in individual productivity. As for cumulative productivity, the curve seems to be flatter and closer to zero for formal businesses than for all businesses, indicating slower growth in cumulative productivity compared with the entire group. In Figure 1.a.c, which represents the dynamics of productivity in informal businesses, the curve of individual productivity fluctuates around zero. However, dispersion appears to be slightly lower than in the case of formal businesses, indicating slightly lower variability in individual productivity. The Table 4. The impact of online business on changes in firm productivity Impacts Online business operation No Online business operation (1) (2) (3) Categories ATT ATU ATE All firms 889 8,946 3.068*** −1.856*** −1.524*** (0.058) (0.004) (0.008) Formal firms 466 1253 3.194*** −1.422*** −1.110*** (0.060) (0.006) (0.009) Informal firms 7693 423 1.175*** −8.590*** −7.923*** (0.026) (0.006) (0.013) Notes. The standard deviations are shown in parentheses. ***p < 0.01, **p < 0.05, * p < 0.1. Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 21 of 46 cumulative productivity curve also shows an upward trend, but the curve seems to be slightly farther from zero compared to that in the case of formal businesses, indicating slightly faster growth in cumulative productivity. Figure 1 (b) presents the Lorenz curve for companies, categorizing them based on their engagement in online business and operations in the informal sector. First, when examining all firms, a distinct difference in productivity can be observed between those involved in online commerce and those who are not (Figure 1(b-b)). From the 30th quantile onward, the dotted yellow curve surpasses the solid blue curve, signifying a lower productivity gap among companies conducting online businesses. However, if we distinguish between firms, this pattern is the opposite for firms in formal sectors (Figure 1(b-d)), whereas it is the same for firms in informal sectors (Figure 1(b-f)). Consequently, the Lorenz curve diagrams highlight the uneven impact of online business operations across the examined productivity quantiles. 4.3.2. Instruments validity test In Table 5, we perform econometric tests on the instruments used to account for endogeneity. We employed permanent workforce, firm size, access to credit, firm bookkeeping, and crossed variables for bookkeeping and technology access issues as the instruments. Firms are more inclined to operate online to compensate for a lack of staff. However, a company’s permanent workforce rarely changes from one year to the next, and thus has little effect on changes in productivity. Similarly, access to credit can have long-term effects on firm performance. However, in the short run, the impact on changes in firm productivity was limited (see Tables 6, 7). Nonetheless, this may be synonymous with an environment that is conducive to infrastructure development, particularly a.a (All Firms) )smriFlamrofnI(c.a)smriFlamroF(b.a Figure 1a. Non parametric curve of firms’ productivity and its cumulative. Source. Authors’ computation based on the 2018 General Business Census data of Togo (INSEED, 2019). Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 22 of 46 in telecommunications. Good bookkeeping may be a sign of efficient business management; however, it is only relevant if the gains from good management are invested well. This reinvestment involves increasing the number of online businesses by controlling for other variables that can affect firm productivity. The econometric tests carried out in Table 5 are in line with our a priori assumptions about the instruments and endogenous nature of the online business variable. The tests confirm the endogeneity of the online business variable, because we reject the null hypothesis that the standard and instrumented regressions are not significantly different. Fischer’s statistics show that these instruments correlate with our endogenous variables. Nevertheless, as Nelson and Startz (1990) show through a simple Monte Carlo experiment, the properties of our estimator can be problematic in a finite sample (Nelson & Startz, 1990). Particularly at certain quantiles, when we compare the standard errors of the standard regression with those of the regression with instruments, we notice that the latter is much higher. This may be a sign of the weakness of the instruments and the presence of bias in instrumented regressions (Cameron & Trivedi, 2005). However, the simulations conducted by Staiger and Stock (1997) in their study show that, if the Fischer statistic b-a: Lorenz curve for all firms b-b: Lorenz curve for all firms by online business b-c: Lorenz curve for formal firms b-d: Lorenz curve for formal firms by online business b-e: Lorenz curve for informal firms b-f: Lorenz curve for informal firms by online business Figure 1b. Lorenz curves for all samples and by participation variable. Source. Authors’ computation based on the 2018 General Business Census data of Togo (INSEED, 2019). Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 23 of 46 Table 8. Online business and internet access impact on changes in productivity using smoothed instrumental variable quantile regression Informal firms (Online business) Informal firms (Robustness check: internet access) Standard quantile regression Smoothed instrumental variable quantile regression Standard quantile regression Smoothed instrumental variable quantile regression Variables 0.30 0.60 0.90 0.30 0.60 0.90 0.30 0.60 0.90 0.30 0.60 0.90 Online business operation 0.007 0.037 0.172 1.330* 2.191 4.305*** (0.037) (0.027) (0.187) (0.745) (1.814) (1.025) Internet access −0.028 −0.002 0.096 2.131 14.868*** 6.009*** (0.035) (0.026) (0.184) (1.671) (4.850) (1.534) Age of the firm −0.025 −0.001 0.058 −0.024 0.009 0.099 −0.025 0.000 0.058 −0.022 −0.010 0.034 (0.018) (0.013) (0.090) (0.036) (0.041) (0.088) (0.018) (0.013) (0.092) (0.036) (0.045) (0.085) Age of the manager 0.000 −0.000 −0.000 0.001*** −0.000 0.000 0.000 −0.000 −0.000 0.001 −0.001 −0.001 (0.001) (0.000) (0.003) (0.001) (0.001) (0.002) (0.001) (0.000) (0.003) (0.001) (0.001) (0.002) Instruction of the manager 0.000 −0.005 0.169 −0.018 0.027 0.071 0.002 −0.002 0.172 −0.028 0.002 −0.044 (0.020) (0.015) (0.104) (0.053) (0.044) (0.092) (0.020) (0.015) (0.106) (0.053) (0.070) (0.106) Age of the firm * Instruction of the manager 0.009 0.005 −0.025 0.008 −0.002 −0.041 0.009 0.004 −0.026 0.004 −0.001 −0.019 (0.008) (0.006) (0.042) (0.027) (0.020) (0.038) (0.008) (0.006) (0.043) (0.022) (0.022) (0.038) Region −0.023 −0.043*** −0.023 −0.057*** −0.046 −0.163* −0.025 −0.042*** −0.013 −0.024 −0.010 0.011 (0.020) (0.014) (0.100) (0.020) (0.042) (0.086) (0.020) (0.014) (0.102) (0.023) (0.039) (0.079) Access to technology issue −0.025 0.002 0.067 −0.084*** −0.034 −0.441*** −0.023 0.003 0.096 −0.076** −0.084 −0.413*** (0.024) (0.018) (0.122) (0.029) (0.083) (0.109) (0.024) (0.018) (0.124) (0.033) (0.079) (0.092) Concurrence issue −0.017 −0.002 −0.003 −0.013 −0.018 0.149** −0.021 −0.001 −0.014 −0.005 −0.008 0.114* (0.017) (0.012) (0.085) (0.016) (0.034) (0.075) (0.017) (0.012) (0.087) (0.022) (0.036) (0.069) Transport infrastructure issues 0.069** −0.033 0.044 0.034 0.004 −0.074 0.068** −0.031 0.032 −0.116*** −0.141 −0.385** (0.030) (0.022) (0.153) (0.029) (0.032) (0.084) (0.030) (0.022) (0.157) (0.043) (0.108) (0.151) Region * Infrastructure −0.113*** 0.027 −0.039 −0.128*** −0.061 −0.103 −0.110*** 0.027 −0.028 0.010 0.035 0.130 (Continued) Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 30 of 46 Table 8. (Continued) Informal firms (Online business) Informal firms (Robustness check: internet access) Standard quantile regression Smoothed instrumental variable quantile regression Standard quantile regression Smoothed instrumental variable quantile regression Variables 0.30 0.60 0.90 0.30 0.60 0.90 0.30 0.60 0.90 0.30 0.60 0.90 (0.038) (0.028) (0.192) (0.043) (0.043) (0.172) (0.038) (0.028) (0.197) (0.045) (0.095) (0.175) Turnover in 2016 −0.019*** −0.006** −0.337*** −0.024*** −0.049 −0.316*** −0.019*** −0.005** −0.329*** −0.051*** −0.105** −0.313*** (0.003) (0.002) (0.016) (0.007) (0.064) (0.028) (0.003) (0.002) (0.016) (0.009) (0.050) (0.030) Permanent staff 0.161*** 0.005 0.215 0.168*** 0.005 0.197 (0.032) (0.024) (0.165) (0.032) (0.024) (0.169) The size of the firm in 2016 −1.676*** −0.127 1.489** −1.679*** −0.124 1.446* (0.147) (0.109) (0.750) (0.148) (0.108) (0.767) Issue to access credit 0.023 0.004 −0.055 0.022 0.004 −0.045 (0.017) (0.013) (0.087) (0.017) (0.013) (0.089) Bookkeeping 0.052* 0.044* 0.270* 0.050 0.044* 0.275* (0.031) (0.023) (0.159) (0.031) (0.023) (0.163) Bookkeeping * Technology −0.027 0.034 −0.100 −0.020 0.044 −0.084 (0.063) (0.047) (0.324) (0.064) (0.047) (0.331) Constant −0.072 0.109** 4.609*** 0.112 0.705 75.320*** −0.073 0.105** 4.517*** 0.141 1.715** 111.992*** (0.064) (0.047) (0.325) (0.097) (1.108) (0.416) (0.064) (0.047) (0.332) (0.101) (0.858) (0.398) Observations 8,116 8,116 8,116 8,116 8,116 8,116 8,116 8,116 8,116 8,116 8,116 8,116 Notes. The standard deviations are shown in parentheses. ***p < 0.01, **p < 0.05, * p < 0.1. Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 31 of 46 Table 9. Estimates of the endogenous switching regression model. Determinants of online business and labor demand All firms Formal firms Informal firms Variables Selection No Online business operation Online business operation Selection No Online business operation Online business operation Selection No Online business operation Online business operation Turnover in 2016 0.0341*** 0.0282*** 0.0035** (0.0026) (0.0044) (0.0014) Age of the manager 0.0001 0.0025*** −0.0027*** (0.0005) (0.0008) (0.0004) The size of firm in 2016 1.2878*** 0.7977*** 0.4221*** (0.0322) (0.0367) (0.0715) Age of the firm −.0016*** .0006 −.0012 .0341** .0006 −.0018 (.0005) (.0052) (.0037) (.0172) (.0005) (.0023) Instruction of the manager .0088*** .1250*** .1036*** .2171*** .0176*** .0186* (.0032) (.0258) (.0190) (.0735) (.0030) (.0105) Age of the firm * Instruction of the manager .0043*** .0087*** .0071*** −.0031 .0019*** .0010 (.0003) (.0020) (.0015) (.0060) (.0003) (.0011) Region .0070* .0820*** −.0780** −.1519 −.0002 .0363*** (.0039) (.0312) (.0335) (.1071) (.0036) (.0136) Access to technology −.0296*** .0158 .0034 −.0151 .0394 (.0103) (.0734) (.0890) (.0094) (.0278) (Continued) Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 32 of 46 Table 9. (Continued) All firms Formal firms Informal firms Variables Selection No Online business operation Online business operation Selection No Online business operation Online business operation Selection No Online business operation Online business operation Concurrence issue −.0002 −.0191 .0277 −.0638 −.0008 −.0035 (.0035) (.0244) (.0227) (.0609) (.0032) (.0087) Bookkeeping .1765*** .4323*** .1301*** −.0019 (.0049) (.0322) (.0056) (.0139) Transport infrastructure issues −.0128** .0380 .0384 .0696 −.0185*** .0406** (.0060) (.0367) (.0479) (.1690) (.0054) (.0173) Access to credit −.0016 −.0242 −.0141 −.0362 −.0041 .0037 (.0035) (.0229) (.0209) (.0546) (.0033) (.0084) Instruction of the manager * Technology .0360*** −.0025 .0640* .1517 .0251*** −.0153 (.0058) (.0366) (.0365) (.1048) (.0054) (.0136) Region * Infrastructure .0018 −.0139 −.0996* −.0069 .0184** −.0571*** (.0078) (.0469) (.0535) (.1784) (.0072) (.0198) Bookkeeping * Technology .0540*** .0099 −.4392 .0267** .0358* (.0105) (.0576) (.2909) (.0121) (.0209) Constant −1.6853*** .7309*** 1.9079*** −1.3185*** .7426*** 2.3925*** −1.3183*** .7386*** −.6477*** (0.0406) (.0063) (.0718) (0.0756) (.0560) (.2400) (0.0270) (.0058) (.0395) Rho 0 −.8800 −.9325 −.8837 (Continued) Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 33 of 46 Table 9. (Continued) All firms Formal firms Informal firms Variables Selection No Online business operation Online business operation Selection No Online business operation Online business operation Selection No Online business operation Online business operation Rho 1 −.8379 −.8318 .9969 Sigma 0 .3879 .8048 .3183 Sigma 1 1.0078 1.2839 .8611 LR independence test of the equation χ2 = 3826.36 Pvalue=.000 χ2 = 992.59 Pvalue=.000 χ2 = 3677.26 Pvalue = .000 The overall significance test (Wald chi2) χ2 = 5047.20 Pvalue =.000 χ2 = 498.56 Pvalue=.000 χ2 = 1589.87 Pvalue = .000 Observations 41732 41732 41732 4646 4646 4646 37086 37086 37086 Notes. The standard deviations are shown in parentheses. ***p < 0.01, **p < 0.05, * p < 0.1. Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 34 of 46 Table 10. Estimates of the endogenous switching regression model. Determinants of online business and labor supply All firms Formal firms Informal firms Variables Selection No Online business operation Online business operation Selection No Online business operation Online business operation Selection No Online business operation Ebusiness operation Turnover in 2016 0.0146*** 0.0011 (0.0024) (0.0038) Permanent staff 0.4833*** 0.1989*** 0.3601*** (0.0186) (0.0255) (0.0278) Age of the manager −0.0025*** 0.0002 −0.0052*** (0.0006) (0.0010) (0.0007) The size of firm in 2016 0.2471*** 0.1579*** 0.4301*** −.0861 −.1637 (0.0409) (0.0410) (0.1268) (.0566) (.1572) Age of the firm .0005 .0044 −.0001 .0024 −.00002 .0025 (.0010) (.0049) (.0040) (.0122) (.0010) (.0056) Instruction of the manager −.0146** −.0234 .0036 .0208 −.0187*** −.0234 (.0060) (.0221) (.0202) (.0529) (.0064) (.0252) Age of the firm * Instruction of the manager −.0014** −.0039* −.0014 −.0027 −.0007 −.0002 (.0006) (.0020) (.0016) (.0044) (.0006) (.0027) Region −.0704*** −.0587** −.1382*** −.1091 −.0671*** −.0253 (.0075) (.0297) (.0368) (.0800) (.0077) (.0309) Access to technology issue −.0322* −.0839 −.0048 −.0381* −.0740 (.0192) (.0655) (.0882) (.0198) (.0678) (Continued) Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 35 of 46 Table 10. (Continued) All firms Formal firms Informal firms Variables Selection No Online business operation Online business operation Selection No Online business operation Online business operation Selection No Online business operation Ebusiness operation Internet .0987*** .0135 .0846*** .0006 .1209*** .0287 (.0163) (.0199) (.0325) (.0438) (.0185) (.0219) Concurrence issue .1051*** .4295*** −.0674* .2687*** .0956*** .2336*** (.0126) (.0327) (.0364) (.0687) (.0135) (.0430) Bookkeeping .0473*** −.1132*** .0741*** −.0142 (.0102) (.0263) (.0130) (.0324) Transport infrastructure issues −.0472*** −.0271 −.0467 −.1832 −.0476*** −.0030 (.0114) (.0423) (.0557) (.1364) (.0115) (.0418) Access to credit −.0024 −.0422** −.0177 −.0585 .0008 −.0287 (.0068) (.0195) (.0221) (.0405) (.0070) (.0213) Instruction of the manager * Technology .0004 .0151 −.0294 .0168 .0040 .0092 (.0111) (.0315) (.0367) (.0768) (.0116) (.0334) Region * Infrastructure .0613*** .0566 .1224** .1852 .0563*** .0392 (.0149) (.0476) (.0611) (.1427) (.0155) (.0488) Staff * Concurrence −.0472*** −.4159*** .0703*** −.1831*** −.0334** −.2457*** (.0126) (.0254) (.0202) (.0337) (.0141) (.0417) (Continued) Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 36 of 46 Table 10. (Continued) All firms Formal firms Informal firms Variables Selection No Online business operation Online business operation Selection No Online business operation Online business operation Selection No Online business operation Ebusiness operation Bookkeeping * Technology −.0708*** .0948** −.0800*** .0371 (.0227) (.0456) (.0290) (.0529) Constant −2.0968*** 3.9954*** 6.1458*** −1.0257*** 4.2805*** 5.2526*** −1.8110*** 3.9945*** 6.2883*** (0.0392) (.0129) (.0762) (0.0735) (.0597) (.1731) (0.0356) (.0129) (.0885) Rho 0 −.006 .908 −.012 Rho 1 −.955 −.942 −.969 Sigma 0 .616 .760 .614 Sigma 1 1.085 1.087 1.113 LR independence test of the equation χ2¼338:10 Pvalue = .000 χ2 = 427.52 Pvalue = .000 χ2 = 397.41 Pvalue = .000 The overall significance test (Wald chi2) χ2 = 309.79 Pvalue = .000 χ2= 50.96 Pvalue = .000 χ2= 293.06 Pvalue = 0.004 Observations 40,179 40,179 40,179 3,866 3,866 3,866 36,374 36,374 36,374 Notes. The standard deviations are shown in parentheses. ***p< 0.01, **p< 0.05, *p< 0.1. Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 37 of 46 Table 11. Impact of online business on job market Categories Impacts Labor demand Labor supply Labor demand Labor supply Online business operation No Ebusiness operation Ebusiness operation No Ebusiness operation ATT ATU ATE ATT ATU ATE All firms 2,819 38,913 2,491 37,688 −0.191*** 1.322*** 1.219*** 0.010** 1.973*** 1.841*** (0.012) (0.001) (0.002) (0.004) (0.001) (0.003) Formal firms 1,640 35,446 1,634 34,679 −0.453*** 1.254*** 1.138*** 0.523*** 0.941*** 0.913*** (0.011) (0.001) (0.002) (0.003) (0.001) (0.001) Informal firms 3,467 1,179 857 3,009 −0.453*** −1.312*** −1.254*** 0.053*** 2.170*** 2.025*** (0.003) (0.0003) (0.001) (0.005) (0.001) (0.003) Notes. The standard deviations are shown in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1. Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 38 of 46 Regarding labor supply, for all firms across sectors, the adoption of online business had a minimal effect, with ATT indicating a 1% increase. This trend was also confirmed for firms in the informal sector, whereas firms in the formal sector showed a more pronounced increase. This might appear counterintuitive as the adoption of online business can be seen as a technological evolution that reduces the number of employees required (Frey & Osborne, 2017). However, adopting online business, much like innovation, entails the emergence of new tasks within firms, the development of accompanying skills, or even prompting employees to work overtime (Arntz et al., 2016). Our results reveal that ATUs and ATEs are significantly higher than ATTs, suggesting that non-online businesses have longer working hours than ATTs do. Combined with previous results (see Tables 4 and 6) showing productivity gains from online business adoption, we can infer that online business allows companies, particularly those in the informal sector, to utilize their labor supply more efficiently. 4.5. Discussion of the results Our empirical estimations strongly support the first hypothesis of our theoretical model (Assumption 1), suggesting that firms that engage in online business activities experience significant productivity changes. By employing two distinct econometric methodologies, the endogenous switching regression method and the smoothed instrumental variable quantile regression method, our findings are robust (Caliendo & Kopeinig, 2008). The use of proxy variables such as Internet access further strengthens the robustness of our results. Online business represents substantial innovation that enables firms to become more efficient. These stages correspond to the phases of innovators and early adopters in Rogers (1962) diffusion model. However, innovation diffusion across the Togolese economy is non-uniform. Firms in the informal sector primarily benefit from this innovation as they achieve significant performance gains through online business operations (Nagayets, 2005). By facilitating the diversification of products and services, Internet business operations enable companies to reach many potential customers, particularly those in remote areas (UNCTAD, 2021). This allows for improvements in productivity (Khan, 2016). E-commerce is emerging as a vital innovation that propels Togolese businesses, especially those in the informal sector, to enhance their economic performance. This development aligns with the broader trend in developing countries, where e-commerce has become a driver of economic growth and poverty reduction (UNCTAD, 2021; WorldBank, 2016). The potential of e-commerce to bridge the digital divide and promote inclusive growth in Togo is further supported by studies such as Aker and Mbiti (2010), who emphasize the importance of mobile and digital technologies in enhancing the productivity of African firms. The adoption of online business operations in Togo led to significant changes in firm productivity, particularly in the informal sector. This innovation can boost economic performance and contribute to inclusive growth, especially as the digital divide narrows and more firms adopt e-commerce strategies. The second panel shows that online business operations and adoption negatively affect employment. Indeed, these findings do not seem to align with the results on the impact on business productivity. The apparent contradiction between these two phenomena can be attributed to the fact that the adoption of information and communication technologies (ICTs), such as Internet operations, can have opposing effects on business productivity and employment. Internet use can enhance business productivity and efficiency, leading to increased productivity. Companies can expand their reach, lower transaction and communication costs, access new markets, and streamline their operational processes through internet use. This viewpoint is supported by research by Jorgenson and Stiroh (1999) and Brynjolfsson and Hitt (2000), who investigate the impact of ICTs on business performance and productivity. However, the adoption of technologies, such as the Internet, can lead to a decline in permanent jobs. This can be attributed to several factors: (i) The adoption of the Internet and digital technologies can result in the increasing automation of work processes, meaning that certain tasks performed by permanent employees can now be executed Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 39 of 46 change, completeness of financing microstructures, and impact on well-being and income inequality. Telecommunications Policy, 47(6), 102571. https://doi. org/10.1016/j.telpol.2023.102571 Song, Y., Escobar, O., Arzubiaga, U., & De Massis, A. (2022). The digital transformation of a traditional market entrepreneurial ecosystem. Review of Managrialt Science, 16(1), 65–88. https://doi.org/10.1007/ s11846-020-00438-5 Soto-Acosta, P. (2020). COVID-19 pandemic: Shifting digital transformation to a high-speed gear. Information Systems Management, 37(4), 260–266. https://doi.org/10.1080/10580530.2020.1814461 Staiger, D., & Stock, J. (1997). Instrumental variables regression with weakinstruments. Econometrica, 65 (3), 557–586. https://doi.org/10.2307/2171753 Stephen, A. T., & Toubia, O. (2010). Deriving value from social commerce networks. Journal of Marketing Research, 47 (2), 215–228. https://doi.org/10.1509/jmkr.47.2.215 Stock, J., & Watson, M. (2022). Introduction to econometrics (4th ed.). Pearson. https://www.pearson.com/ en-us/subject-catalog/p/introduction-toeconometrics/P200000006421/9780136879787 Stroumpoulis, A., & Kopanaki, E. (2022). Theoretical Perspectives on sustainable supply chain data Management and digital transformation: A literature review and a conceptual framework. Sustainability, 14(8), 4862. https://doi.org/10.3390/su14084862 Svobodová, Z., & Rajchlová, J. (2020). Strategic behavior of E-Commerce businesses in online industry of Electronics from a customer perspective. Administrative Sciences, 10(4), 78. https://doi.org/10. 3390/admsci10040078 Swamy, L. (2020). The digital economy: New business models and key features. International Journal of Research in Engineering, 3(7), 118–122. https://jour nal.ijresm.com/index.php/ijresm/article/view/33 Taherdoost, H., & Madanchian, M. (2023). Blockchainbased e-commerce: A review on applications and challenges. Electronics, 12(8), 1889. https://doi.org/ 10.3390/electronics12081889 Tan, C. (2016). Lifelong learning through the SkillsFuture movement in Singapore: Challenges and prospects. International Journal of Lifelong Education, 36(3), 278–291. https://doi.org/10.1080/02601370.2016. 1241833 Teng, X., Wu, Z., & Yang, F. (2022). Research on the relationship between digital transformation and performance of SMEs. Sustainability, 14(10), 6012. https:// doi.org/10.3390/su14106012 Terzi, N. (2011). The impact of e-commerce on international trade and employment. Procedia - Social & Behavioral Sciences, 24, 745–753. https://doi.org/10. 1016/j.sbspro.2011.09.010 Thaichon, P., Surachartkumtonkun, J., Quach, S., Weaven, S., & Palmatier, R. (2018). Hybrid sales structures in the age of e-commerce. Journal of Personal Selling and SAles Management, 38(3), 277–302. https://doi.org/10. 1080/08853134.2018.1441718 UNCTAD. (2021). Digital economy report 2021: Crossborder data flows and development-for whom the data flow. https://unctad.org/publication/digitaleconomy-report-2021 Viu-Roig, M., & Alvarez-Palau. (2020). The impact of e-commerce-related last-mile logistics on cities: A systematic literature review. Sustainability, 12(16), 6492. https://doi.org/10.3390/su12166492 Wang, Y., Hong, A., Li, X., & Gao, J. (2020). Marketing innovations during a global crisis: A study of China firms’ response to COVID-19. Journal of Business Research, 116, 214–220. https://doi.org/10.1016/j. jbusres.2020.05.029 Wooldridge, J. (2015). Introductory econometrics: A modern approach. South-western cengage learning. https://www.cengage.ca/c/introductoryeconometrics-a-modern-approach-7e-wooldridge /9781337558860/ World-Bank. (2016). World development report 2016: Digital dividends. World Bank Publications. https:// www.worldbank.org/en/publication/wdr2016 World-Bank. (2020). World development report 2020: Trading for development in the age of global value chains. World Bank. https://doi.org/10.1596/9781-4648-1457-0 WTO. (2020). E-commerce, trade and the COVID-19 pandemic. WTO Working Papers No. 2020/05. Xia, L., & Monroe, K. (2010). Is a good deal always fair? Examining the concepts of transaction value and price fairness. Journal of Economic Psychology, 31 (6), 884–894. https://doi.org/10.1016/j.joep.2010. 07.001 Xue, L., Zhang, Q., Zhang, X., & Li, C. (2022). Can digital transformation promote green technology innovation? Sustainability, 14(12), 7497. https://doi. org/10.3390/su14127497 Yin, Z., & Choi, C. (2022). Does e-commerce narrow the urban-rural income gap? Evidence from Chinese provinces. Internet Research, 32(4), 1427–1452. https://doi.org/10.1108/INTR-04-2021-0227 Yin, Z., & Choi, C. (2023). The effect of trade on the gender gap in labour markets: The moderating role of information and communication technologies. Economic Research, 36(1), 2443–2462. https://doi.org/10.1080/ 1331677X.2022.2100434 Yu, Y., Han, X., Liu, J., & Cheng, Q. (2015). Supply chain equilibrium among companies with offline online selling channels. International Journal of Production Research, 53(22), 6672–6688. https://doi.org/10. 1080/00207543.2015.1055350 Zandi, M. (2009). The economic impact of the American recovery and reinvestment act. Sodokin et al., Cogent Economics & Finance (2023), 11: 2268790 https://doi.org/10.1080/23322039.2023.2268790 Page 46 of 46