Policy Research Working Paper 10716 Cross-Country Empirical Analysis of GovTech Platforms on Citizen Engagement Maimouna Diakite Abdoul-Akim Wandaogo Governance Global Practice March 2024 Policy Research Working Paper 10716 Abstract Countries worldwide are implementing GovTech reforms to regressions on the matching sample as alternatives. Addi- modernize the public sector and achieve better performance tional robustness checks were performed using alternative while responding to citizens’ needs. At its core, GovTech GovTech Maturity Index 2022 data and by considering the represents a whole-of-government approach to public sector possibility of a slower diffusion of the technology. A sensi- modernization, which emphasizes three critical aspects: (i) tivity analysis, considering the role of governance, political citizen-centric public services that are universally accessible; and institutional factors, as well as the level of development, (ii) a whole-of-government approach to digital government is likewise performed. The results show a significant and transformation; and (iii) simple, efficient, and transparent positive impact of GovTech platforms on citizen engage- government systems. Within this context, strengthening cit- ment. Similarly, democracy and the equal distribution of izen engagement is crucial to ensure accountability, improve political power have strong and positive effects on citizen public policy quality, and enhance service delivery. Accord- engagement. By contrast, public sector corruption nega- ingly, this study aims to be the first cross-country empirical tively and significantly impacts citizen engagement. The assessment of the impact of GovTech platforms, which can findings also provide evidence that GovTech platforms are allow citizens to: (i) participate in policy decision-making more effective in fostering citizen engagement in high-in- and (ii) provide feedback on public service delivery. Using a come economies and in countries where the government is large sample of 176 countries, the study assesses the impact efficient, institutional and social fragility is low, and there is of the implementation of national platforms that allow cit- no conflict or only low-intensity conflict. The results of an izens to participate more effectively. This research employs Africa-focused analysis indicate that African countries that entropy balancing as the main identification strategy, as have adopted such digital platforms likewise experience an well as propensity score matching and ordinary least squares increase in citizen engagement. This paper is a product of the Governance Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at mdiakite2@worldbank.org and awandaogo@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Cross-Country Empirical Analysis of GovTech Platforms on Citizen Engagement Maimouna Diakite and Abdoul-Akim Wandaogo1 JEL Classification: C1, H11, 033, O55 Keywords: Citizen engagement, GovTech, e-government, digital platforms, Impact analysis, entropy balancing, weighted least squares, propensity score matching, ordinary least squares, control function, Africa. 1 Maimouna Diakite is an Economist Consultant at the World Bank’s Governance Global Practice and Presidential CEoG (Chief Economists of Government) Fellow at ACET/World Bank. Abdoul Akim Wandaogo is an Economist Consultant at the World Bank’s Governance Global Practice and Research Associate at the "Centre d'Etudes et de Recherches sur le Développement International (CERDI)". The authors thank Tracey Marie Layne, Gael Raballand, Michael Jelenic, Cem Dener, Saki Kumagai, João Ricardo Vasconcelos, Uwe Serdült, Tiago Peixoto, the participants of the World Bank Africa fellow research seminar, the anonymous reviewers and participants of ICEGOV 2023, and participants of the AFW GovTech community of practice for valuable comments/contributions and support. The findings, interpretations, and conclusions expressed in this paper are those of the authors alone, and do not represent or reflect the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. I. Introduction Countries worldwide are implementing GovTech reforms to modernize the public sector and achieve better performance while responding to citizens’ needs. At its core, GovTech represents a whole-of- government approach to public sector modernization, which emphasizes three critical aspects: (i) citizen- centric public services that are universally accessible; (ii) a whole-of-government approach to digital government transformation; and (iii) simple, efficient, and transparent government systems. Building upon the foundations of e-Government 2 and Digital Government, 3 GovTech encompasses the effective use of a variety of platforms and technologies, which include, among others: (i) artificial intelligence and machine learning, cloud computing, and the internet of things; (ii) public data platforms promoting the use of open public data by individuals and firms; (iii) local GovTech ecosystems supporting local entrepreneurs and start-ups to develop new products and services for government; and (iv) greater use of public-private partnerships to draw upon private sector skills, innovations, and investments to address public sector challenges. 4 Digital platforms are typically based on the internet and can take many forms, such as websites, mobile applications, social media, cloud computing services, online marketplaces, content management systems, and more (Senshaw & Twinomurinzi, 2022; Bonina et al. 2021; World Bank, 2020; OECD, 2019 and Asadullah et al. 2018). As documented by the existing literature, the adoption of GovTech platforms and other digital government tools has the potential to improve government effectiveness. Such tools have been shown to: (i) facilitate internal and external collaboration between different government entities (Islam, Trautmann, and Buxmann 2016); (ii) reduce costs of storage, duplication, transmission, and errors (Fichman, Dos Santos, and Zheng 2014; Negroponte, 1995); and (iii) enable engagement with private actors and thereby promote transparency, democracy, and freedom of action (Falk, Römmele, and Silverman 2017). As the literature demonstrates, however, the specific form of the platform depends on the type of task its participants are trying to accomplish (Bonina, Carla et al., 2021; Cusumano et al., 2019; Jacobides et al., 2018; Constantinides et al., 2018). This study is particularly interested in digital GovTech platforms dedicated to: (i) consulting citizens in making policy decisions; and (ii) soliciting and collecting feedback on public service delivery. With respect to the former, this study is concerned with digital platforms such as e-petition platforms, participatory budget platforms, and public consultation platforms, among others. With respect to the latter, this study is concerned with digital platforms that allow citizens to provide feedback on service delivery, including the availability of a platform that allows citizens to provide complements, make suggestions, make information requests, or complain about the delivery of public services. Taken together, such public consultation, participatory, and user feedback platforms can positively affect levels of citizen engagement 5 within countries, as they contribute to closing the communication loop between 2 E-Government refers to the principles of: (i) a user-centered approach, but which is supply driven; (ii) one-way communications and service delivery; (iii) ICT-enabled procedures, but which often are analogue in design; (iv) sliced ICT development and acquisition; (v) greater transparency; and (vi) government as the primary provider. 3 Digital Government refers to the principles of: (i) procedures that are digital by design; (ii) user-driven public services; (iii) Government as a Platform (GaaP); (iv) open by default; (v) data-driven public sector: and (vi) proactive administration. 4 GovTech: The New Frontier in Digital Government Transformation. Equitable Growth, Finance & Institutions Notes. World Bank (2020). 5 Citizen Engagement (CE) is defined at its most basic as “two-way interactions” between citizens and governments and covers a wide range of activities that empower citizens to make better decisions, give inputs into projects affecting their communities, and help them be heard in public matters. Citizen Engagement takes many forms, called mechanisms, and serves many functions as evidenced by their names: Grievance Redress Mechanism, Community Score Cards, Participatory Budgeting, and Social Audit, 2 government and citizens. Such “CivicTech” 6 tools are based on the use of digital technology to improve governance and the interactions between government and citizens (World Bank, 2022). Importantly, evidence shows that such technologies can promote citizens’ trust and help governments achieve improved development results (World Bank, 2022; Waddington et al., 2019; Bae et al., 2019; Diergarten and Krieger, 2015; Lindstedt and Naurin, 2010; Aidt, 2009; Shim and Eom, 2008). However, while the existing studies have demonstrated linkages between digital tools and citizen engagement, there has been no empirical assessment of the link between these variables using econometric modeling at a cross-country level. In this lacuna, the present research paper seeks to estimate the extent to which and under which conditions GovTech platforms support improvements to citizen engagement. In doing so, this study attempts to fill a gap in the literature by using original data for a large sample of 176 countries (for which data are available) in order to analyze the importance of GovTech platforms in improving citizen engagement. As such, this is a cross-sectional study using primarily: (i) the World Bank’s 2020 GovTech Maturity Index data (extracted from the 2022 GTMI version) and (ii) the 2021 Varieties of Democracy (V-dem) citizen engagement data to assess the impact of the implementation of national platforms that allow citizens to participate in policy decision-making as well as to provide feedback on service delivery. A time lag has been introduced to allow for consideration of the uptake and diffusion, as the effect of introducing digital platforms to support citizen engagement may take sufficient time to appear. Importantly, this research considers the effects of contextual and demographic factors on this relationship, including, inter alia: political regime type; level of development; perceptions of public sector corruption; coverage of broadband subscriptions; distribution of political power among social groups; population distribution between 15 and 44; human capital development; measures of government effectiveness; and indices of conflict and fragility. This analysis utilizes entropy balancing as the primary estimation strategy. In recent years, this method has been used for impact analyses on cross-sectional or panel data in various fields (Kinda and Thiombiano, 2024; Balima and Sy, 2021; Sawadogo, 2020; Gottschalk et al., 2020; Baborska, 2020; Pek et al., 2019). Importantly, this method has several advantages compared to the other matching methods and avoids multicollinearity issues. Propensity Score Matching (PSM) and Ordinary Least Squares (OLS) regressions on the matching sample are used as alternative strategies. Additional robustness checks were performed using alternative GTMI data (GTMI 2022) and by considering the possibility of a slower diffusion of the technology. In addition, a sensitivity analysis considering the role of governance as well as political and institutional factors in the success of GovTech initiatives in boosting citizen participation is performed. Also, an Africa-focused analysis was conducted to test whether the adoption of digital platforms in an African context has an impact on citizen engagement. This study finds a significant and positive impact of the implementation of GovTech platforms allowing citizens to participate in policy decision-making and provide feedback on public service delivery. These results are robust with respect to the use of alternative data and identification strategies and the rate of diffusion of technology. In addition, the findings show that democracy and the equal distribution of political power have strong and positive effects on citizen engagement. Likewise, public sector corruption has a negative and significant impact on citizen engagement. The analysis also provides evidence that for example. (World Bank, 2023. Amplifying People’s Voices - Opportunities for Mainstreaming Citizen Engagement Through Digital Technologies (English). Washington, D.C.) 6 CivicTech refers to any innovation in the use of digital technology to improve governance and the interactions between government and citizens. Depending on the nature and engagement between government and citizens, CivicTech activities can be categorized into three levels: (1) openness and transparency; (2) participation and engagement; and (3) collaboration, co- design, co-creation, and co-production (World Bank, 2022). This study concerns the second and third levels of CivicTech activities. 3 GovTech platforms are more effective in fostering citizen engagement in high-income economies and in countries where the government is efficient, institutional and social fragility is low, and there is no conflict or only low-intensity conflict. The results of the Africa focused analysis indicate that African countries that have adopted digital platforms likewise experience an increase in citizen engagement. The remainder of this research paper is organized as follows: Section II provides a brief review of the existing literature; Section III presents the data used and an overview of the adoption of GovTech platforms and the level of citizen engagement worldwide; Section IV describes the main methodological framework and displays the main findings; section V describes the robustness checks and displays the findings; Section VI presents the sensitivity analysis; Section VII presents the additional analysis focused on Africa; and Section VIII concludes with recommendations. II. Literature Review Literature shows, in general, a positive effect of the use of digital technologies on government performance (World Bank, 2022; Waddington et al., 2019; Bae et al., 2019; Diergarten and Krieger, 2015; Lindstedt and Naurin, 2010; Aidt, 2009; Shim and Eom, 2008). For instance, implementation of the Proactive Listening Initiative (an online/mobile phone user feedback platform) in the Dominican Republic has supported growing resolution rates of reported issues and reduction of corruption reports by 70%, with close to 100% of the feedback provided indicating good or excellent levels of satisfaction, according to Peixoto and Fox (2016). The analysis of Haikin et al. (2015) assessing the impact of the online voting platforms in Brazil highlighted that at minimum, a 12.2% increase in citizen participation is directly attributable to online voting. Also, by assessing the impact of digital tools on the inclusivity of citizen engagement during the COVID-19 pandemic, Hofstra et al. (2023) found that digital tools, including interactive platforms, facilitate citizen participation. However, the consensus on the impact of digital platforms on citizen engagement is also debated and notable caveats exist. David (2017) explored both the theoretical and empirical literature on civic engagement in the context of e-government and found that citizen engagement through e-government has the same shortcomings and caveats as citizen participation, and the use of ICT itself brings with it additional challenges. According to this study, quality control measures must be followed to use ICT for civic engagement to produce the expected benefits. Similarly, Lee-Geiller and Lee (2019) conducted a qualitative meta-analysis of four strands of literature and developed the Democratic E-governance Website Evaluation Model. From the literature review they conducted, they found that governments have used their websites mainly for nascent patterns of communication, particularly information sharing and communication between the citizens and their governments. Spada et al. (2015) studied the effects of technology on citizen engagement through online voting in Brazil, finding that technology appears only more likely to engage people who are younger, male, of higher income and educational attainment, and more frequent social media users. Other literature has studied the contextual factors needed to ensure that digital technologies and platforms translate into improved citizen engagement. World Bank (2022) highlights the potential of e- government to make the interactions between government and citizens easier by allowing governments to reach more people and provide a wide range of information. However, the study identifies key foundations needed to be in place for such initiatives to succeed, such as digital infrastructure availability, digital skills of public servants, as well as the legal and regulatory framework. Gulati et al. (2014) assessed the relationship of political factors and policy initiatives to improved e-government services and e- participation capabilities using the online service index and the e-participation index of the United 4 Nations. Their findings show that there is greater e-government capability in countries that have more effective public sector governance and administration as well as policies that advance the development and diffusion of information and communication technologies. The comparative study of Serdült et al. (2023) extended the existing index of digital political participation in the Swiss cantons to the national and international levels and found that existing measures of digital political participation overestimate opportunities for digital political participation in countries with rather low levels of substantive digital political participation. According to Peixoto and Steinberg (2019), technologies can substantially impact how citizens engage and how power is sought; however, only when institutions change their rules to match the capabilities of modern tools and the expectations of the public. Previous research also highlights the importance of critical analogue enabling factors, including the necessary Access to Information and media enabling environment that foster citizen engagement as well as the effect that ICT enabled platforms can have in this context. In this regard, literature on the efficacy of Freedom of Information (FoI) and Access to Information (ATI) legislation supports the importance of government transparency in facilitating citizen engagement. FoI and ATI laws have been found to be associated with both increased perceptions and detection rates of corruption (Costa, 2012) as well as improved government decision-making and public awareness (Hazell, Worthy, and Glover, 2010). While such legislation may provide a useful prerequisite for citizens to hold government to account, their effectiveness has been found to increase in contexts of higher media freedom, political pluralism, and CSO activism (Vadlamannati and Cooray, 2016). Linking this to the use of ICT-enabled platforms, Peixoto (2013) highlights the importance of technological capacity to process politically important data as well as the existence of a free press and internet in doing so. Another strand of research suggests that more democratic institutions and processes may impact the relationship between digital technologies and citizen participation. Kneuer and Harnisch (2016) analyzed “when” and “how” e-government and e-participation diffuse among regime subtypes. Their results show that e-government and e-participation adoption differs substantially between and among regime types over time and in kind. Democracies, on average, adopt more e-government and e-participation tools than autocracies, and they do so earlier, whereas deficient democracies adopt much fewer e-techniques than established democracies, and they do so later. In this regard, the literature also argues that elite capture, lack of civic capacities, or other local factors could undermine the potential gains from citizen participation and citizen associations could even be un-representative or manipulated (Gaventa and Barrett 2010; Bonfiglioli 2003, Golooba-Mutebi 2004, Crook and Sturla Sverrisson 2001). Additional research reveals that greater use of e-government websites may even have negative effects on citizens' satisfaction and perceptions of public sector trustworthiness (Porumbescu 2016). Welch et al. (2005) used US survey data to assess the link between e-government and citizen satisfaction. Their findings show that, in general, the use of online services increases citizens’ satisfaction, which in turn increases their trust in the government; however, individuals with more concern about the government's responsiveness tend to be less satisfied. Moreover, greater use of information can also enable bad governments to better realize bad policy objectives (Jacobs, 2017). As such, Ragnedda (2018) and Roos, Ank and Albert (2022) found that digital citizen engagement could lead to social exclusion due to unequal digital literacy and access to the internet. Thus, factors such as political regime, civil liberty and justice, quality of the institution, and internet usage could affect citizen engagement and participation. III. Data and Statistical Analysis 1. Data and Sources 5 The sample includes 176 7 countries for which data was available. Dependent Variables The dependent variables (which are used one by one) are the V-dem 8 variables for: (i) Civil Society Participation Index (CSPI); and (ii) engaged society (ES). The CSPI is a V-Dem mid-level index developed using a range of V-Dem indicators, which measures, the extent to which civil society organizations (CSOs) are involved in government decision-making, 9 and the values range from 0 (low) to 1 (high). The V-Dem variable ES assesses how wide and how independent are public deliberations when important policy changes are being considered, and the values range from 0 (no public deliberation) to 5 (large numbers of public deliberation). This study uses the ordinal scale version of these two variables. 10 Variables of Interest The study uses data from the GovTech Maturity Index (GTMI) dataset, for the digital platforms’ implementation variable. Developed by the World Bank Group in 2020 and 2022, GTMI covers 198 11 economies and includes 48 key indicators. For this study, GTMI indicators I-30 and I-31 are used. I-30 12 considers whether there are national web-based platforms that allow citizens to participate in policy decision-making, whereas I-31 13 assesses the existence of government platforms that enable citizens to provide feedback on service delivery. I-30 and I-31 are dummy variables that take the value 1 if the platforms exist and 0 otherwise. The present study constructs a GovTech variable taking 1 when the countries have implemented at least one of I-30 or I-31 and 0 otherwise. The main analysis focuses on 2020 GTMI data extracted from the 2022 version of the dataset, which includes both 2020 and 2022 data. The 2020 GTMI data were collected remotely by the GTMI team during the COVID-19 period and were displayed in the 2022 dataset after converting to some of the relevant 2022 sub-indicators and including new sub-indicators. By contrast, the 2022 GTMI data were collected mainly based on an online survey completed by government officials. There are more positive responses in the 2022 GTMI data regarding the implementation of GovTech platforms (15 for I-30, 34 for I-31, and 39 for the GovTech variable). However, some improvements in the 2022 data compared with that of 2020 may be related to undetected relevant websites that the GTMI team were unable to spot remotely. To consider these aspects, this paper uses both GTMI 2020 data (in the main analysis) and GTMI 2022 (in a robustness check) extracted from the 2022 version of the dataset. 7 This is the number of countries for which data are available for the dependent variable and the variables of interest. This number can be reduced in the different estimations due to data unavailability for the matching and/or control variables. 8 Varieties of Democracy is a database of multidimensional and disaggregated data related to measuring democracy and other factors. It contains three general types of data: (i) variables coded by country experts; (ii) factual data based on existing sources, pre-calculated indices based on these two data sources; and (iii) variables from external sources. 9 The questions that constitute this indicator are whether policy makers regularly consult key CSOs; how much participation is by individuals in CSOs; whether women are prevented from participating; and whether the nomination of legislative candidates within the party organization is highly decentralized or is done through party primaries. 10 Annex A presents the complete list the variables used, their definitions, and the data sources. 11 The number of countries covered by this study is limited to 176 due to a lack of information on the other variables (dependent and control variables). 12 This GTMI indicator measures whether there are national platforms that allow citizens to participate in policy decision- making. It takes the value 1 if such a platform exists and 0 otherwise. 13 This GTMI indicator measures whether there are government platforms that allow citizens to provide feedback (e.g., compliments, complaints, suggestions, info requests) on service delivery. This captures the availability of a platform that allows citizens to provide feedback, make suggestions, collect information, or complain about the delivery of public services. It also takes 1 if such a platform exists and 0 otherwise. 6 Control Variables The study has extracted the control variables from various sources, including the World Development Indicators (WDI) and other relevant sources. The set of control variables includes the political regime or democracy measure, including Polity2 of the Polity V dataset 14 as political regime characteristics may lead to differentiated effects across countries. Following the literature (Hilger et al., 2020; Vicari et al., 2015), the study also uses a set of control variables from WDI, including socio-economic and demographic characteristics (e.g., the share of the population aged between 15 and 44, the human capital index, and GDP per capita 15). The study also uses the power distributed by social group variable from V-Dem. In addition, the study uses the fixed internet broadband subscriptions per 100 inhabitants variable as a proxy for internet penetration (Boulianne, 2009; Howard, 2006; Putman, 2000) from the International Telecommunication Union (ITU) dataset. 2. Overview of the Adoption of GovTech Platforms and the Level of Citizen Engagement Worldwide In the context of this study, a statistical analysis (including stylized facts) was conducted related to the adoption of GovTech platforms and the level of citizen engagement worldwide, focusing on the 198 countries covered by the GTMI. Table 1 summarizes information on the implementation of digital platforms. In the sample of 198 countries, 82 (41.4%) have Implemented a digital platform allowing citizens to participate in policy decision-making, and 75 countries (37.8%) have implemented a platform enabling citizens to provide feedback. However, 112 countries (more than 56%) still need a digital platform for citizens to participate in decision-making or give feedback. Figure 1 highlights that countries that have implemented GovTech, have a relatively higher average level of citizen engagement than countries that have not implemented it. The average differences between adopters and non-adopters are respectively 0.14 points and 0.15 points in 2021 and 2022 for the Civil Society Participation Index and 0.68 points and 0.87 points for the Engaged Society variable. A geographical representation of the countries shows that most countries that have yet to adopt this platform are in Africa and Eastern Europe. In comparison, most North and South American countries as well as Western European countries have platforms in place allowing citizens to participate in decision-making and/or give their opinions. Table 1: Cross-tabulation of I30 and I31 adoption Figure 1: Average Citizen Participation by (2020) 16 whether GovTech is implemented or not (2020 GMTI Data and 2021 V-dem Data) I-31 4 3 0 1 Total 2 0 112 4 116 1 I-30 0 1 11 71 82 CSP_2021 CSP_2022 ES_2021 ES_2022 Total 123 75 198 Non_GovTech GovTech 14 The Polity V project codes the authority characteristics of states in the world system for purposes of comparative, quantitative analysis. 15 This allows for the removal of the implicit income effects and reduces the gap between the values of the different variables. 16 Information concerning the GTMI 2020 data (extracted from the 2022 version of the dataset) is provided above in the subsection Data and Sources. 7 IV. Estimation Strategy Based on the literature on impact analysis and the structure of the dependent variable, this study employs the entropy-balancing method (Hainmueller, J. 2012 and Hainmueller, J., and Xu, Y. 2013) as the main empirical strategy, which combines matching and regression analysis. It allows for selection matches of countries having implemented GovTech platforms allowing citizens to participate in policy decision- making and provide feedback on public service delivery and subsequent estimation of the impact of the implementation of these platforms on citizen engagement. In recent years, this method has been used for impact analyses on cross-sectional and panel data in various fields (Kinda and Thiombiano, 2024, 2023; Balima and Sy, 2021; Sawadogo, 2020; Gottschalk et al., 2020; Baborska, 2020; Pek et al., 2019; Neuenkirch and Neumeier, 2016). Importantly, this method has several advantages compared to the other matching methods. 17 This study is based on the idea that the adoption of digital platforms represents a treatment. It is a cross-sectional study with country-level observations. Observations with GovTech platforms constitute the treatment group, and those without digital platforms represent the control group. The measure of interest is the well-known average treatment effect on the treated (ATT) defined as follows: = � () � = ] − � () � = ] (1) Where is the outcome variable, that is, the citizen engagement/participation variable, indicates whether a unit (country) is exposed to treatment (has adopted a digital platform) = or not = . Accordingly, � () � = ] is the expected level of citizen engagement after the adoption of a digital platform and � () � = ] is the counterfactual outcome, namely the level of citizen engagement that a country having adopted a digital platform would have reach if it had not adopted a digital platform. Given that the counterfactual outcome is not observable, an appropriate proxy is needed to identify the average treatment effect on the treated. If the adoption of a digital platform was a random event, � () � = ] would be a suitable proxy. However, the decision to adopt a digital platform may be influenced by some observable factors, including contextual, socio-economic, and demographic factors. Thus, this study compares countries having adopted digital platforms with countries without digital platforms that are similar as possible with regard to observable characteristics that: (i) are correlated with a country decision to adopt a digital platform; and (ii) influence the level of citizen engagement. Under the condition that countries without digital platforms (the control group) are similar as possible to the treated countries, the difference in the level of citizen engagement () is caused by the adoption of digital platforms. Accordingly, Equation (1) can be rewritten as follows: () = � () � = , = ] − � () � = , = ] (2) is a vector of relevant covariates, which are described in the data section; � () � = , = ] is the expected level of citizen engagement for countries having adopted a digital platform; and � () � = , = ] is the expected level of citizen engagement for the treated countries’ best matches (synthetic control units). As stated above, this study uses entropy balancing to select the best matches for the treated countries. 17 As no concern for functional form misspecification exists, it outperforms conventional matching and regression methods, as it avoids multicollinearity issues because of the orthogonality of the treatment variable to the covariates due to the reweighting mechanism. It then balances the covariates between the treatment group and the control group. 8 Formally, to estimate the average treatment effect on the treated with the entropy balancing reweighting scheme, this study follows two consecutive steps. In the first step, the control group that best matches the treatment group is created by reweighting the covariates. This first step consists of measuring the significance of the mean difference between the covariates of the control group and the treated group and then reweights the covariates (using the weighting variable) to ensure that the mean difference is not significant between both groups. In the second step, the effect of the treatment variable (GovTech platforms) on the dependent variable (citizen engagement) is estimated through a regression analysis in the reweighted data by using the weighted least squares estimator. In the second step, the study also controls for the entropy balancing covariates to increase the efficiency of the estimates. An alternative method employed to estimate the effect of the adoption of digital platforms on citizen engagement is the propensity score matching method (PSM) (Rosenbaum and Rubin, 1983), which is a popular technique commonly used in impact evaluations. PSM is adapted to non-randomized studies and permits correct sample selection biases related to the differences between the control and treatment groups. Thus, it enables to reduce the imbalance in the covariates’ distribution. As stated earlier, the estimation strategy introduces a one-year lag for the variable of interest to assess the effect of the adoption of digital GovTech platforms on citizen engagement given that this effect is expected to appear later. Another advantage of shifting the interest variable by lag is that it contributes to dealing with reverse causality, which is the source of endogeneity. In this case, according to Data and Agrawal (2004) and Wandaogo (2022), if the effect of the lag variable on the dependent variable is significant, the result is not due to the two-way effect. Main Findings Before estimating the impact of GovTech platforms on citizen engagement, this study first focuses on the performance of entropy balancing in building the synthetic control group. Table 2 presents the sample means of all matching covariates and the results of the synthetic group construction. Columns 1 and 2 show the mean covariates for countries with and without GovTech platforms, respectively. All the differences in means between these two groups are significant (at the 10 percent threshold), meaning that the contextual, socio-economic, and demographic factors are different between countries having adopted digital platforms and those without such platforms. GDP per capita, human capital index, democracy, equality (i.e. equal distribution of political power), and internet broadband subscription, are higher in countries that have implemented digital GovTech platforms that allow citizens to participate in policy decision-making and provide feedback on service delivery. By contrast, corruption is higher in countries without GovTech platforms. These significant differences in means, show the importance of selecting a proper control group before estimating the treatment effect; otherwise, the predicted treatment impact might be biased. After creating the synthetic control group by using entropy balancing, the differences in means are virtually nonexistent (column 6) and not statistically significant (column 7) for all the matching variables. This demonstrates the efficacy of entropy balancing in creating a perfect control group for the treatment group. 9 Table 2: Descriptive Statistics Before and After Weighting and Covariate Balancing (1) (2) (3) (4) (5) (6) (7) A B C=A-B D E=Weight*B F=A-E G Without Webal* GovTec - Difference p_value Without- Difference p_value h GovTec GovTech h Democracy (Polity 2) 5.45 3.677 1.773 0.073 5.45 0 1.000 GDP per capita (ln) 9.841 8.815 1.026 0.000 9.841 0 1.000 Fixed broadband 12.56 0 1.000 21.07 8.51 0.000 21.07 subscriptions Public sector -0.2697 0 1.000 0.4688 0.7385 0.000 0.4688 corruption Population 15 - 44 20.94 22.08 -1.14 0.003 20.94 0 1.000 Equal distribution of 0.392 0.001 1.000 2.438 2.046 0.013 2.437 political power Human capital index 0.6249 0.4924 0.1325 0.000 0.6249 0 1.000 Note: This table presents the sample means of the covariates before weighting for the treated units in column (1) and the control group in column (2). Column (5) shows the synthetic control group obtained by balancing the entropy. Columns (3) - (4) and (6) - (7), respectively, show the mean difference and the p_value of the mean difference t-test before and after weighting. As noted, this study estimates the causal effect of GovTech platforms on two forms of citizen engagement variables, namely, the V-Dem Civil Society Participation Index (CPSI) and the Engaged Society (ES) indicator using the weighted least squares estimator as a second step of the entropy balancing. The results of the estimations are presented in Table 3. Most of the estimations revealed a significant and positive impact of GovTech platforms on citizen engagement. The coefficient of the GovTech variable is not significant in only one baseline regression excluding the matching covariates. An omitted variable bias is to be expected in the baseline estimations. As such, the introduction of control variables could enhance the impact of GovTech platforms on civil engagement. The results obtained through the multiple regressions that control for the covariates are positive at the 5 percent threshold regardless the form of civil engagement. These results are similar to those of Tolbert and Mossberger (2006) and Hofstra et al. (2023), who found that e-government can improve governments’ interactions with citizens and perceptions of responsiveness. As such, the adoption of digital platforms can: (i) help governments to reach a wider audience and foster meaningful participation of citizens/CSOs in the policy-making process by providing them with an accessible channel to engage with policy makers and participate in policy discussion and consultation while eliminating the geographical and environmental barriers; (ii) allow CSOs to provide real-time feedback to decision-makers and help to establish independent feedback mechanisms, ensuring that public deliberations are not influenced or biased by vested interests; and (iii) enable more people from diverse backgrounds to share their opinions, perspectives, concerns, and suggestions thus creating more inclusive and responsive governance systems at a reduced cost. The magnitude of the impact varies according to the form of citizen engagement. The estimated coefficient of the GovTech variable is 0.09 for the two estimations for which the study employed the civil society participation index as the dependent variable. CSPI captures the extent to which CSOs are consulted by policy makers and people (especially women) are involved in CSOs, and the process of nomination of the legislative candidates. The significantly positive coefficient of the GovTech variable is higher in the regression where the study used the engaged society indicator as the dependent variable 10 and reaches 0.31. This suggests that countries with digital GovTech platforms experience a 0.31 points increase in public deliberations when important policy changes are being considered compared with their peers that have not adopted a GovTech platform. As regards the covariates, the results show a positive and significant effect of democracy on citizen engagement at the 1 percent threshold in both estimations. More democracy means more independent and wide public deliberations, freedom of expression and association, suggesting more opportunities to participate in decision-making. Citizen participation is generally agreed to be an essential ingredient of a healthy democracy (Bullock, 2014) thus, the level of citizen engagement might be closely related to the level of democracy in a country. The results also indicate a positive and significant impact of the equal distribution of political power variable on citizen engagement. Inequality leads to the exclusion of certain vulnerable people from meaningful participation in the decision-making process. So, when a region or a social group is systematically excluded based on socioeconomic status, gender, ethnicity, caste, race, religion, and other factors, this means less participation of these people in political activities and public consultations. Like the findings of Lee, Chang, and Berry (2011), the results show that public sector corruption has a negative impact on citizen engagement. However, the study highlights significant impacts of corruption only on the civil society participation variable. There are no significant impacts of the remaining control variables. Table 3: Weighted Least Squares (WLS) Estimations’ Results Civil society participation Engaged society (2021) (2021) (1) (2) (3) (4) GovTech 0.0908* 0.0908*** 0.306 0.306** (0.0485) (0.0318) (0.206) (0.143) Democracy (Polity 2) 0.0254*** 0.109*** (0.00338) (0.0152) GDP per capita (ln) 0.0102 0.0929 (0.0363) (0.163) Fixed broadband 0.00206 0.00624 subscriptions (0.00287) (0.0129) Public sector -0.137** -0.253 corruption (0.0575) (0.259) Population 15 - 44 0.00803 0.00579 (0.0103) (0.0463) Equal distribution of 0.0752*** 0.260** political power (0.0247) (0.111) Human capital index -0.509 -0.760 (0.319) (1.436) Constant 0.684*** 0.432 2.819*** 1.017 (0.0343) (0.406) (0.146) (1.829) Observations 145 145 145 145 R-squared 0.024 0.601 0.015 0.549 Covariates No Yes No Yes Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 11 V. Robustness Checks The main findings of this study show that the adoption of GovTech platforms that allow citizens to participate in policy decision-making and provide feedback on service delivery positively impacts citizen engagement, as determined through the entropy balancing method. In this section, additional analyses are performed to check the robustness of the main results by considering the rate of diffusion of technology, alternative data for the variable of interest, and alternative estimation strategies. 1. Considering the Rate of Diffusion of Technology In the main analysis, the study introduced a one-year lag for the variable of interest to consider the effect of the uptake and diffusion of GovTech platforms, as the effect of introducing a digital platform on citizen engagement can be expected to appear later. In the first robustness check, it was assumed that the uptake and diffusion could take time to manifest; thus, the model introduced instead a two-year lag for the citizen engagement variable to estimate the effect of the adoption of digital platforms. The results of this additional analysis are consistent with the previous findings and even more justify the adoption of digital GovTech platforms to foster citizen engagement. In all the estimations conducted, a significant and positive impact of GovTech platforms was found on citizen engagement at the 5 percent threshold at least (Table 4). The coefficients of the GovTech variable are higher compared to those obtained in the main analysis for both the baseline and multiple regressions. Accordingly, as for other forms of GovTech reforms, there is a need for a diffusion time before observing the expected positive impact of the implementation of digital platforms on citizen engagement. Table 4: Results of the estimations considering the rate of diffusion of technology Civil society participation Engaged society (2022) (2022) (1) (2) (3) (4) GovTech 0.130** 0.130*** 0.539*** 0.539*** (0.0523) (0.0343) (0.202) (0.141) Observations 145 145 145 145 R-squared 0.042 0.608 0.047 0.558 Covariates No Yes No Yes Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 2. Using Alternative GTMI Data As specified in section III, the GovTech variable employed in the study as the variable of interest was constructed using 2020 GTMI data extracted from the 2022 version of the GTMI dataset. However, the 2020 data were collected remotely by the GTMI team during the COVID-19 pandemic. By contrast, the 2022 GTMI data were collected mainly based on an online survey completed by government officials. To control for the potential biases related to these aspects, this study conducts a second robustness check consisting of using the 2022 GTMI data as an alternative to the 2020 data. The study also considers the rate of diffusion of technology in this analysis. The findings show a positive and significant impact of the adoption of GovTech platforms on citizen engagement (Table 5). Here again, the predicted coefficients for the GovTech variable are higher compared to those obtained in the main analysis. These results provide evidence that GovTech platforms 12 positively affects citizen engagement regardless of the version of GTMI data employed, the form of civil engagement, the type of regression (baseline or multiple), and the technology diffusion rate. Table 5: Results of the estimations conducted through the 2022 GTMI data Civil society participation (CSPI) Engaged society (ES) CSPI (2021) CSPI (2022) ES (2021) ES (2022) (1) (2) (3) (4) (5) (6) (7) (8) GovTech (2022 data) 0.155*** 0.155*** 0.163*** 0.163*** 0.283 0.283** 0.545*** 0.545*** (0.0503) (0.0342) (0.0558) (0.0370) (0.211) (0.142) (0.202) (0.148) Observations 145 145 145 145 145 145 145 145 R-squared 0.062 0.589 0.056 0.605 0.012 0.576 0.049 0.515 Covariates No Yes No Yes No Yes No Yes Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 3. Using Alternative Estimation Strategies: Propensity Score Matching (PSM) This third robustness check employs propensity score matching (PSM) using various matching techniques as an alternative estimation method. The study also takes into account the rate of diffusion of technology in estimating the causal effect of GovTech platforms on citizen engagement as in the two first robustness checks. The Propensity Scores (PS) are estimated using the Logit model. Figure 2 presents the distribution of the PS across countries in the sample before and after matching. Figure 2: Density plot for PS before and after matching for GovTech Table 6 showcases the evaluation results using various matching techniques alongside statistics for standard diagnostic tests. To begin, the pseudo-R2 gauges the degree to which the control variables account for the likelihood of GovTech platforms adoption, consequently producing balanced scores (Sianesi, 2004). Caliendo and Kopeinig (2008) argue that satisfactory model performance is linked to “relatively low” values of the pseudo-R2. Given that none of the pseudo-R2 values surpass 0.05, meaning that the matches generate balanced scores, and the outcomes align with the common support hypothesis. Next, the standardized bias test assesses the conditional independence assumption for observables. The P-value should be higher than all the conventional thresholds. Producing p-values ranging from 0.182 to 0.956 indicates that there is no statistical distinction between the characteristics of GovTech adopters and non-adopters after matching. The Rosenbaum upper bound sensitivity test examines the presence of unobservable factors that might impact the estimated citizen engagement resulting from GovTech 13 platform adoption. Critical values vary from 1.17 to 1.85, suggesting that the results remain robust with respect to the conditional independence hypothesis. 18 The estimated effect of adopting GovTech platforms on citizen engagement ranges from 0.12 to 0.18 for the CSP and from 0.45 to 0.85 for engaged society, contingent upon the matching method and the year (2021 or 2022). This underscores that GovTech platform adoption contributes to an increase in citizen engagement. The Average Treatment Effects (ATE) are between 0.12 and 0.19 for CSP and between 0.54 and 0.76 for engaged society, signifying that GovTech platforms enhance citizen engagement for both adopters and non-adopters. These results are similar to the main results, confirming that the result is not sensitive to alternative matching methods. Table 6: The effects of GovTech platforms on civil engagement using PSM (1) (2) (3) (4) (5) (6) Radius 1-nearest 2-nearest 3-nearest Local caliper Kernel Treatment variable: GovTech neighbor neighbor neighbor Linear matching matching matching matching matching matching (r=0.05) Average Treatment Effect of the Treated (ATT) Civil Society Participation 2021 0.1781* 0.1672** 0.1500* 0.1161* 0.1347 0.1205* (0.0959) (0.0800) (0.0766) (0.0673) (0.0988) (0.0722) Civil Society Participation 2022 0.2469** 0.2203** 0.1979** 0.1634** 0.1656 0.1683** (0.0997) (0.0877) (0.0819) (0.0820) (0.1076) (0.0826) Quality of the marching Pseudo-R2 0.041 0.019 0.016 0.009 0.041 0.009 Rosenbaum upper bound 1.53 1.58 1.63 1.58 1.66 1.71 sensitivity test Standardized bias (p-value) 0.242 0.760 0.823 0.956 0.242 0.955 Engaged society 2021 0.7250** 0.6500** 0.5292** 0.4502* 0.6304* 0.4706* (0.3386) (0.3055) (0.2639) (0.2689) (0.3608) (0.2724) Engaged society 2022 0.8500** 0.7937** 0.7083*** 0.6435** 0.7654** 0.6602** (0.3352) (0.3123) (0.2711) (0.2582) (0.3587) (0.2736) Quality of the marching Pseudo-R2 0.046 0.018 0.011 0.009 0.05 0.009 Rosenbaum upper bound 1.85 1.23 1.29 1.58 1.27 1.29 sensitivity test Standardized bias (p-value) 0.182 0.793 0.930 0.956 0.134 0.963 Standard errors in parentheses. ***significance level at 1 percent; **significance level at 5 percent; *significance level at 10 percent. Bootstrap replication=500. Observations/Treated=145/80 4. Ordinary Least Squares (OLS) Regressions on the Matching Sample 18 The test is performed at a level of 5%. It should be higher than 1, at least. Using the simulation-based sensitivity analysis presented by Ichino et al (2008), the robustness of the estimates is also tested in the event of a failure of the conditional independence hypothesis. The test suggests that any unobserved factor correlated with each of the covariates used in this study would not be sufficient to reduce the estimated mean treatment effect to zero. 14 This study also conducts subsequent OLS regressions on the matching sample as another step of PSM (as the use of WLS for entropy balancing). This method, which is less common in the literature (DuGoff et al. 2014; Ho et al. 2007), allows conducting both baseline and multiple regressions on the matching sample as an alternative to entropy balancing. The PSM sample was generated using the nearest neighbor matching technique and a Logit model. The regression coefficients presented for the OLS models estimated using the PSM sample indicate a significant and positive impact of GovTech platforms on citizen engagement, irrespective of the form of the regression (Table 7). As for the main analysis through the entropy balancing method (with the weighted least squares estimator), the results reveal a higher impact of GovTech platforms on citizen engagement measured through the engaged society variable than that measured through the civil society participation index. The significance of the coefficients of the GovTech variable, irrespective of the method employed, shows that the findings are not sensitive to alternative identification strategies. Table 7: Results of the estimations conducted through OLS regressions on the matching sample Civil society participation (CSPI) Engaged society (ES) CSPI (2021) CSPI (2022) ES (2021) ES (2022) (1) (2) (3) (4) (5) (6) (7) (8) GovTech 0.178*** 0.0870** 0.247*** 0.130*** 0.725*** 0.321** 0.850*** 0.460*** (0.0475) * (0.0552) (0.0326) (0.204) (0.137) (0.196) (0.132) (0.0292) Observations 160 160 160 160 160 160 160 160 R-squared 0.082 0.689 0.113 0.722 0.074 0.625 0.106 0.635 Covariates No Yes No Yes No Yes No Yes Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Overall, the results of the robustness checks indicate that the adoption of GovTech platforms positively affects citizen engagement, regardless of the rate of diffusion of technology, the version of GTMI data used, and the estimation strategy employed. VI. Sensitivity/Heterogeneity Analysis The present study also includes a sensitivity analysis focused on the civil society participation variable to consider the role of factors needed to ensure that the implementation of digital platforms translates into improved citizen engagement, as highlighted by World Bank (2022), Peixoto and Steinberg (2019), and Gulati et al. (2014). The idea is that factors such as government effectiveness, conflict and fragility, and the level of development can magnify or alleviate the effect of GovTech platforms on citizen engagement. The study also considers the level of development as a potential source of heterogeneity to test if it plays a role in the success of GovTech platforms in fostering citizen participation. Following Lin and Ye (2009) and Sawadogo (2020), the study uses a control function regression methodology for the heterogeneity analysis developed by Wooldridge (2002). The method involves including the propensity scores firstly estimated as a control function in an OLS regression. The statistical significance of the propensity scores would reflect the presence of self-selection in the models. The results show that GovTech platforms are more effective in fostering citizen engagement in developed (high-income) economies, and in countries where the government is efficient; institutional and social fragility is low; and there is no conflict or only low-intensity conflict (Table 8). More specifically, the coefficient of the interaction term between the government effectiveness variable and GovTech is positive and statistically significant at the 1 percent threshold, and the coefficient of GovTech becomes smaller 15 when the study includes this interaction term. This suggests that the effect of GovTech platforms on citizen engagement is sensitive to the quality of public services, civil service, as well as policy formulation and implementation, which strongly impact the success of GovTech initiatives in boosting citizen participation. These findings are similar to those of Gulati et al. (2014). The results also highlight the negative impact of the presence and intensity of conflict and fragility on the effectiveness of GovTech platforms to foster citizen engagement. Furthermore, the negative and statistically significant interaction term between GovTech platforms and the developing countries dummy indicates that GovTech reforms may face challenges in effectively promoting citizen engagement in developing countries. However, this finding highlights the need for tailored strategies and targeted interventions to address the unique contexts and challenges faced by developing countries. It underscores the importance of refining GovTech implementation to align with the specific needs and conditions of these countries, presenting an opportunity for more context-sensitive and impactful digital governance solutions in the future. Concerning the efficacy of the empirical strategy, none of the predicted coefficients for the propensity score are statistically significant, suggesting the absence of self-selection biases. Table 8: Results of the Sensitivity/Heterogeneity Analysis (1) (2) (3) (4) (5) (6) (7) (8) (9) GovTech 0.178*** 0.178*** 0.209*** 0.208*** 0.184*** 0.184*** 0.111** 0.105** 0.272*** (0.0475) (0.0476) (0.0482) (0.0482) (0.0476) (0.0476) (0.0492) (0.0525) (0.0705) Pscore 0.107 0.0241 0.0371 0.0720 0.0720 -0.131 -0.0291 -0.0291 (0.123) (0.124) (0.122) (0.122) (0.122) (0.145) (0.152) (0.152) GovTech*Conflict -0.204** (intensity level) (0.0807) GovTech*Conflict -0.126** (cumulative intensity) (0.0510) GovTech* Conflict-affected - situations 0.249*** (0.0342) GovTech* Fragile and - Conflict-affected situations 0.166*** (0.0228) GovTech* Government 0.149*** effectiveness (0.0424) GovTech*High income 0.167** countries (0.0784) GovTech*Developing -0.167** countries (0.0784) Constant 0.597*** 0.526*** 0.581*** 0.572*** 0.549*** 0.549*** 0.684*** 0.616*** 0.616*** (0.0371) (0.0893) (0.0885) (0.0878) (0.0883) (0.0883) (0.0998) (0.106) (0.106) Observations 160 160 160 160 160 160 160 160 160 R-squared 0.082 0.087 0.116 0.116 0.102 0.102 0.150 0.114 0.114 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 VII. Analysis Focused on Africa 16 In this section, an additional Africa-focused analysis was conducted to test whether the adoption of digital GovTech platforms in an African context has an impact on citizen engagement. The sample includes 8 African countries having implemented digital GovTech platforms and 112 control countries without GovTech platforms. This choice was made to retain non-African countries without GovTech platforms in the control group to increase the number of observations. The entropy balancing was once again applied as identification strategy. According to the findings, the adoption of a GovTech platform translates into improved citizen engagement in Africa. For example, the positive and significant coefficient of the GovTech variable in the first column of Table 9 means that African countries having adopted digital platforms allowing citizen to participate in policy decision-making and provide feedback on public service delivery experience 0.13 points increase in citizen engagement. However, the results were not significant when using the 2021 citizen engagement data compared to those obtained when using the 2022 data. This suggests that there is a need for a diffusion time before observing the positive and significant impact of GovTech on citizen engagement through public deliberations. Also, the magnitude of the impact of GovTech is higher for the engaged society variable. Table 9: Results of the Analysis focused on Africa Civil society participation (CSPI) Engaged society (ES) CSPI (2021) CSPI (2022) ES (2021) ES (2022) (1) (2) (3) (4) (5) (6) (7) (8) GovTech 0.139** 0.139*** 0.137** 0.137*** 0.337 0.337 0.852*** 0.852*** Africa (0.0584) (0.0465) (0.0536) (0.0442) (0.253) (0.224) (0.225) (0.185) Observations 84 84 84 84 84 84 84 84 R-squared 0.064 0.436 0.074 0.401 0.021 0.267 0.149 0.452 Covariates No Yes No Yes No Yes No Yes Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 VIII. Conclusion and Recommendations Citizen engagement and participation play a pivotal role in ensuring accountability, improving public policy quality, and enhancing service delivery. The study’s results highlight a significant and positive impact of the implementation of GovTech platforms on citizen engagement. This robust finding holds across alternative data, identification strategies, and different rates of technology diffusion. Moreover, the study identifies democracy and the equal distribution of political power as strong, positive factors affecting citizen engagement, while public sector corruption negatively affects citizen participation. In terms of global application, GovTech platforms prove more effective in fostering citizen engagement in developed economies with efficient governments; low institutional and social fragility; and little to no conflict. An additional Africa-focused analysis reinforces the positive impact of digital platforms on citizen engagement in African countries. These results suggest that GovTech platforms hold the potential to enhance citizen engagement/participation, particularly in democratic, egalitarian societies, and low- corruption contexts. However, the success of the GovTech reforms hinges on government effectiveness, institutional quality, political stability, and level of development. Therefore, critical policy recommendations emerge from these findings, and include the following: 17 • Improving government effectiveness through enhanced public policy formulation and implementation, along with capacity building for civil servants, stands as a foundational recommendation. This involves investing in targeted capacity building and training programs, updating essential infrastructure, and implementing streamlined processes, such as those focusing on the efficiency of public service delivery. By bolstering the capabilities of government agencies in these specific areas, countries can ensure more efficient and responsive governance. This, in turn, directly impacts citizens' engagement by facilitating smoother interactions and delivering improved services. • Improving institutional quality through transparency and anti-corruption efforts is a crucial step in fostering an environment conducive to citizen participation. This recommendation entails implementing and strengthening anti-corruption institutions and measures, ensuring transparent decision-making processes, and holding public officials accountable, namely through reporting or public disclosure requirements. These efforts contribute to building trust in institutions, encouraging citizens to actively engage in civic processes, and voice their concerns. Ultimately, these measures enhance the overall effectiveness of GovTech reforms. • Fostering political stability and an improved social contract are identified as a key factor in encouraging citizen participation. This recommendation involves implementing policies that promote political stability, conflict resolution, inclusive governance, and state legitimacy. Concrete examples could include mechanisms for resolving conflicts peacefully and ensuring representation from diverse groups. Stable political and democratic environments provide citizens with the confidence to engage actively in civic activities, contributing to a more robust and participatory democracy. • Ensuring inclusive, reliable, and affordable access to electricity and the internet is pivotal, especially in developing nations and African countries. This recommendation emphasizes infrastructure development to provide widespread access to essential utilities. Improved connectivity not only facilitates the use of GovTech platforms, but also contributes to overall economic development and societal progress. This comprehensive approach ensures inclusiveness as citizens, regardless of location or economic status, could actively participate in decision-making processes. • Promoting digital literacy through education and training emerges as a critical step in preparing citizens for effective engagement with GovTech platforms. This recommendation involves integrating digital literacy programs into educational curricula, offering training sessions for the general population, and fostering partnerships with NGOs and private entities. Digital literacy empowers citizens to navigate and utilize GovTech tools, making their engagement more meaningful and impactful. References Aidt, T. S. (2009). Corrup�on, ins�tu�ons, and economic development. Oxford review of economic policy, 25(2), 271-291. Asadullah, A., Faik, I., & Kankanhalli, A. (2018). Digital Pla�orms: A Review and Future Direc�ons. PACIS, 248. 18 Baborska, R., Hernandez, E., Magrini, E., & Morales-Opazo, C. (2020). The impact of financial inclusion on rural food security experience: a perspec�ve from low-and middle-income countries. Review of Development Finance, 10(2), 1-18. Bae Y-J., Choi S. W, Kim M. J., & Kim S. (2019). Leveraging ICT Pla�orms to Foster Ci�zen Engagement For Enhanced Public Accountability: The Korean Experience. World Bank, Washington, DC. © World Bank. htps://openknowledge.worldbank.org/handle/10986/31837 License: CC BY 3.0 IGO. Balima, H., & Sy, A. (2021). IMF-supported programs and sovereign debt crises. IMF Economic Review, 69, 427-465. Bonfiglioli, A. (2003). Empowering the Poor: Local Governance for Poverty Reduc�on. United Na�ons Capital Development Fund, New York. Bonina, C., Koskinen, K., Eaton, B., & Gawer, A. (2021). Digital pla�orms for development: Founda�ons and research agenda. Informa�on Systems Journal, 31(6), 869-902. Brambilla, I., & Tortarolo, D. (2018). Investment in ICT, produc�vity, and labor demand: the case of Argen�na. World Bank Policy Research Working Paper, (8325). Bullock, K. (2014). Ci�zen Par�cipa�on and Democracy. In: Ci�zens, Community and Crime Control. Crime Preven�on and Security Management. Palgrave Macmillan, London. htps://doi.org/10.1057/9781137269331_2. Caliendo, M., and S. Kopeinig (2008). “Some Prac�cal Guidance for the Implementa�on of Propensity Score Matching”. Journal of Economic Surveys, 22(1): 31–72. Chadwick, A., & Howard, P. N. (2009). The Routledge Handbook of Internet Poli�cs. Routledge. conceptual model for evalua�on. Government Informa�on Quarterly 36 (2019) 208–225. Constan�nides, P., Henfridsson, O., & Parker, G. G. (2018). Introduc�on – Pla�orms and infrastructures in the digital age. Informa�on Systems Research, 29(2), 381–400. htps://doi.org/10.1287/isre.2018.0794 Cordella, A., & Bonina, C. M. (2012). A public value perspec�ve for ICT enabled public sector reforms: A theore�cal reflec�on. Government information quarterly, 29(4), 512-520. Costa, S. (2013). Do freedom of informa�on laws decrease corrup�on?. The Journal of Law, Economics, & Organiza�on, 29(6), 1317-1343. Crook, R. C., & Sverrisson, A. (2001). Decentralisa�on and Poverty Allevia�on in Developing Countries: A Compara�ve Analysis, or, is West Bengal Unique?. IDS Working Paper 130, Brighton: IDS. Cusumano, M. A., Gawer, A., & Yoffie, D. B. (2019). The business of pla�orms: Strategy in the age of digital compe��on, innova- �on, and power, New York: HarperBusiness. Da Cruz N. F., Tavares A. F., Marques R. C., Jorge S., & de Sousa L. (2016). Measuring local government transparency. Public Management Review 18(6): 866-893. htps://doi.org/10.1080/14719037.2015.1051572. Dahlgren, P. (2009). Media and Poli�cal Engagement: Ci�zens, Communica�on, and Democracy. Cambridge University Press. Data*, A., & Agarwal, S. (2004). Telecommunica�ons and economic growth: a panel data approach. Applied Economics, 36(15), 1649-1654. David, N. (2018). Democra�zing Government: What We Know About E-Government and Civic Engagement. In: Alcaide Muñoz, L., Rodríguez Bolívar, M. (eds) Interna�onal E-Government Development. Palgrave Macmillan, Cham. htps://doi.org/10.1007/978-3-319-63284-1_4. Davison R., Vogel D., Harris R. W., & Jones N. (2000). Technology Leapfrogging in Developing Countries: An Inevitable Luxury? Electronic Journal of Informa�on Systems in Developing Countries, 1, 1-10. htps://doi.org/10.1002/j.1681-4835.2000.tb00005.x. Diergarten, Y., & Krieger, T. (2015). Large-scale land acquisi�ons, commitment problems and interna�onal law. Law and Development Review, 8(1), 217-233. 19 Dobrolyubova, E., Klochkova, E., & Alexandrov, O. (2019). Digitaliza�on and effec�ve government: what is the cause and what is the effect?. In International Conference on Digital Transformation and Global Society (pp. 55-67). Springer, Cham. DuGoff, E.H., Schuler, M., and Stuart, E.A. (2014). Generalizing Observa�onal Study Results: Applying Propensity Score Methods to Complex Surveys. Health Services Research, 49(1): 284–303. Economics, 121, 110-119. Falk, S., Römmele, A., & Silverman, M. (2017). The promise of digital government. In Digital Government (pp. 3-23). Springer, Cham. Fichman, R. G., Dos Santos, B. L., & Zheng, Z. (2014). Digital innova�on as a fundamental and powerful concept in the informa�on systems curriculum. MIS quarterly, 38(2), 329-A15. Foth, M., Forlano, L., Satchell, C., & Gibbs, M. (Eds.). (2011). From social butterfly to engaged citizen: Urban informatics, social media, ubiquitous computing, and mobile technology to support citizen engagement. MIT Press. Gandía, J. L., Marrahí, L., & Huguet, D. (2016). “Digital transparency and Web 2.0 in Spanish city councils.” Government Informa�on Quarterly 33(1): 28–39. htps://doi.org/10.1016/j.giq.2015.12.004. Gaventa, J., & Barret, G. (2010). So, What Difference Does It Make? Mapping the Outcomes of Ci�zen Engagement. IDS Working Papers, Volume 2010, Issue 347, Pages: i-ii, 01-72, October 2010. Gil de Zúñiga, H., Jung, N., & Valenzuela, S. (2012). Social media use for news and individuals' social capital, civic engagement and poli�cal par�cipa�on. Journal of Computer-Mediated Communica�on, 17(3), 319-336. doi:10.1111/j.1083-6101.2012.01574.x Golooba-Mutebi, F. (2004). Reassessing Popular Par�cipa�on in Uganda. Public Administration and Development, 24.4: 289–304. Gotschalk, S., König, H. H., & Bretschneider, C. (2020). The associa�on between informal caregiving and behavioral risk factors: a cross-sec�onal study. International Journal of Public Health, 65, 911-921. Gula�, G.J.J., Williams, C.B., & Yates, D.J. (2014). Predictors of online services and e-par�cipa�on: A cross- na�onal comparison. Government Informa�on Quarterly 31 (2014) 526–533. Haikin M., Sjoberg F. M., & Mellon J. (2015). Impact of Online Vo�ng on Par�cipatory Budge�ng in Brazil. In Peixoto, Tiago; Sifry, Micah L.. 2017. Civic Tech in the Global South: Assessing Technology for the Public Good. Washington, DC: World Bank and Personal Democracy Press. htp://hdl.handle.net/10986/27947. Hazell, R., Worthy, B., & Glover, M. (2010). The impact of the freedom of informa�on act on central government in the UK. Does FOI Work. Hilger A., El Ghandour I., Vircoulon T. and, Diakite M. 2020. Opportuni�es and Constraints to Open Government Partnership (OGP) Membership in Francophone and Lusophone African Countries. World Bank Group. htps://documents.worldbank.org/en/publica�on/documents- reports/documentdetail/684091604328807532/overview-of-the-research-ac�vi�es Ho, D. E., Imai, K., King, G., & Stuart, E. A. (2007). Matching as Nonparametric Preprocessing for Reducing Model Dependence in Parametric Causal Inference. Political Analysis, 15(3), 199–236. htps://doi.org/10.1093/pan/mpl013. Ho, D.E., Imai, K., King, G., and Stuart, E.A. (2007). Matching as Nonparametric Preprocessing for Reducing Model Dependence in Parametric Causal Inference. Political Analysis, 15(3): 199–236. Hö�en, S., & Streich, B. (2013). Mobile par�cipa�on: Ci�zen engagement in urban planning via smartphones. In Citizen E-participation in urban governance: Crowdsourcing and collaborative creativity (pp. 199- 225). IGI Global. Hofstra R., Michels A., & Meijer A. (2023). Online democra�c par�cipa�on during COVID-19. Informa�on Polity, Volume 28, Issue 3, 2023. Pp 395–410. htps://doi.org/10.3233/IP-211540 Holum, M. (2022). Ci�zen par�cipa�on: Linking government efforts, actual par�cipa�on, and trust in local poli�cians. International Journal of Public Administration, 1-11. 20 Howard, P. N. (2006). New media campaigns and the managed citizen. Cambridge University Press. Ichino, A., F. Mealli, and T. Nannicini. (2008). “From Temporary Help Jobs to Permanent Employment: What Can We Learn from Matching Es�mators and Their Sensi�vity?” Journal of Applied Econometrics, 23(3), 305–27. Islam, N., Trautmann, K., & Buxmann, P. (2016). Tradi�on Meets Modernity–Learning from Start-ups as a Chance to Create Digital Innova�on in Corpora�ons. Jacobides, M., Cennamo, C., & Gawer, A. (2018). Towards a theory of ecosystems. Strategic Management Journal, 39(8), 2255–2276. Jacobs, B. (2017). Digitaliza�on and Taxa�on. In IMF 2017 Digital Revolutions in Public Finance (2017). htps://doi.org/10.5089/9781484315224.071. Kinda, H., & Thiombiano, N. (2024). Does transparency mater? Evalua�ng the Impacts of the Extrac�ve Industries Transparency Ini�a�ve (EITI) on Deforesta�on in Resource-rich Developing Countries. World Development, 173, 106431. Kleinhans, R., Van Ham, M., & Evans-Cowley, J. (2015). Using social media and mobile technologies to foster engagement and self-organiza�on in par�cipatory urban planning and neighbourhood governance. Planning practice & research, 30(3), 237-247. Kneuer, M., & Harnisch, S. (2016). Diffusion of e-government and e-par�cipa�on in Democracies and Autocracies. Global Policy Volume 7. Issue 4. November 2016. Lee-Geiller, S., & Lee, T.D. (2017). Using government websites to enhance democra�c E-governance: A conceptual model for evalua�on. Government Informa�on Quarterly 36 (2019) 208–225. Lee, C., Chang, K., & Berry, F. S. (2011). Tes�ng the development and diffusion of e-government and e- democracy: A global perspec�ve. Public Administra�on Review, 71(3), 444–454. Lin, S., & H. Ye (2009). “Does Infla�on Targe�ng Make a Difference in Developing Countries?” Journal of Development Economics, 89(1): 118–23. Lindstedt, C., & Naurin, D. (2010). Transparency is not enough: Making transparency effec�ve in reducing corrup�on. International political science review, 31(3), 301-322. Ndou, V. (2004). E-Government for Developing Countries: Opportuni�es and Challenges. Electronic Journal on Information Systems in Developing Countries, 18, 1-24. Negroponte, N. (1995). Being Digital, New York: Alfred A. Knopf. Neuenkirch, M., & Neumeier, F. (2016). The impact of US sanc�ons on poverty. Journal of Development OECD (2019), Measuring pla�orm mediated workers, OECD Publishing, Paris. htps://www.oecd- ilibrary.org/docserver/170a14d9- en.pdf?expires=1701711551&id=id&accname=guest&checksum=027601172974F2EC6A6F1BC4155 B998C Ohmer, M., & Beck, E. (2006). Ci�zen par�cipa�on in neighborhood organiza�ons in poor communi�es and its rela�onship to neighborhood and organiza�onal collec�ve efficacy. J. Soc. & Soc. Welfare, 33, 179. Peixoto, T. (2012). The Uncertain Rela�onship between Open Data and Accountability: A Response to Yu and Robinson's the New Ambiguity of Open Government. UCLA l. Rev. Discourse, 60, 200. Peixoto, T., & Fox J. (2016). When Does ICT-Enabled Ci�zen Voice Lead to Government Responsiveness?. World Development Report. Digital Dividends Background Paper. Peixoto, T., & Steinberg, T. (2019). Ci�zen Engagement: Emerging Digital Technologies Create New Risks and Value. World Bank, Washington, DC. © World Bank. htps://openknowledge.worldbank.org/handle/10986/32495 License: CC BY 3.0 IGO. Pék, É., Fertő, I., & Alobid, M. (2019). Evalua�ng the Effect of Farmers’ Par�cipa�on in Irriga�on Management on Farm Produc�vity and Profitability in the Mubuku Irriga�on Scheme, Uganda. Water, 11(11), 2413. Porumbescu, G. A. (2016). Linking public sector social media and e-government website use to trust in government. Government Information Quarterly, 33(2), 291-304. 21 Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. Simon and schuster. Ragnedda, M. (2018). Conceptualizing Digital Capital. Telematics and Informatics, 35 (8). pp. 2366-2375. ISSN 0736-5853. Roos, H., Ank, M., & Albert, M. (2022). Online democra�c par�cipa�on during COVID-19: Assessing implica�ons for inclusivity of ci�zen engagement. Information Polity, 28(4):1-16. Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observa�onal studies for causal effects. Biometrika, 70(1), 41-55. Sawadogo, P. N. (2020). Can fiscal rules improve financial market access for developing countries?. Journal of Macroeconomics. Volume 65, September 2020, 103214. Schlozman, K. L., Verba, S., & Brady, H. E. (2012). The Unheavenly Chorus: Unequal Poli�cal Voice and the Broken Promise of American Democracy. Princeton University Press. Senshaw, D., & Twinomurinzi, H. (2022). Innova�ng with government digital pla�orms in low-income countries: the dynamic capabili�es of Woredas in Ethiopia. Journal of Science and Technology Policy Management, 13(4), 812-836. Serdült, U., Hofmann, G., Benli-Trichet, M., Vayenas, C., Villeneuve, J-P., Picco-Schwendener, A., & Colosante, L. (2023). Introducing the DigiPart-Index on the Na�onal Level. Mapping digital poli�cal par�cipa�on for Austria, France, Canada, Germany, Italy, Japan and Switzerland. A work in progress paper for the annual congress of the Swiss Poli�cal Science Associa�on 2023. Shim, D. C., & Eom, T. H. (2008). E-government and an�-corrup�on: Empirical analysis of interna�onal data. Intl Journal of Public Administration, 31(3), 298-316. Sianesi, B. (2004). “An Evalua�on of the Swedish System of Ac�ve Labor Market Programs in the 1990s”. Review of Economics and Statistics, 86(1): 133–55. Spada, P., Mellon, J., Peixoto, T., & Sjoberg, F. M. (2015). Effects of the Internet on Par�cipa�on. Sun, Y., & Yan W. (2020). The Power of Data From the Global South: Environmental CivicTech and Data Ac�vism in China. Interna�onal Journal of Communica�on 14(20): 2144–2162. Tolbert, C. J., & Mossberger, K. (2006). The effects of e-government on trust and confidence in government. Public administration review, 66(3), 354-369. Vaccari, C., Chadwick, A., & O'Loughlin, B. (2015). Dual screening the poli�cal: Media events, social media, and ci�zen engagement. Journal of communication, 65(6), 1041-1061. Vadlamanna�, K. C., & Cooray, A. (2016). Do freedom of informa�on laws improve bureaucra�c efficiency? An empirical inves�ga�on. Oxford Economic Papers, 68(4), 968-993. Waddington H., Sonnenfeld A., Fine� J., Gaarder M., John. D., & Stevenson J. (2019). Ci�zen engagement in public services in low- and middle-income countries: A mixed-methods systema�c review of par�cipa�on, inclusion, transparency and accountability (PITA) ini�a�ves. Campbell Systema�c Reviews. 2019;15: e1025. htps://doi.org/10.1002/cl2.1025 . Wandaogo, A. A. (2022). Does digitaliza�on improve government effec�veness? Evidence from developing and developed countries. Applied Economics, 54(33), 3840-3860. Welch, E. W., Hinnant, C. C., & Moon, M. J. (2005). Linking ci�zen sa�sfac�on with e-government and trust in government. Journal of public administration research and theory, 15(3), 371-391. Wooldridge, J. M. (2002). Econometric analysis of cross sec�on and panel data mit press. Cambridge, MA, 108. World Bank. (2013). World Bank Group Strategy. Washington, DC. © World Bank. htps://openknowledge.worldbank.org/handle/10986/16095 License: CC BY-NC-ND 3.0 IGO. World Bank. (2015). e-Government. World Bank Group Brief. htps://www.worldbank.org/en/topic/digitaldevelopment/brief/egovernment#:~:text=What%20is% 20e%2DGovernment%3F,and%20other%20arms%20of%20government. World Bank. (2016). Evalua�ng Digital Ci�zen Engagement: A Prac�cal Guide. World Bank, Washington, DC, February 2016. 22 World Bank. (2022). CivicTech: Transparency, Engagement, and Collabora�on for Beter Governance. Washington, DC: World Bank. © World Bank. htps://openknowledge.worldbank.org/handle/10986/38169 License: CC BY 3.0 IGO. Yamouri N. (2016). Ci�zen Engagement in fragile and conflict-affected situa�ons: Really?. World Bank Blog published on Development for Peace. htps://blogs.worldbank.org/dev4peace/ci�zen-engagement- fragile-and-conflict-affected-situa�ons-really. 23 Annex A: Overview of variables, definitions, and sources Definition Description Source Digital platform to participate in Are there national platforms that allow citizens to participate in policy decision- GovTech Maturity decision-making I30(2020 and making? It takes the value 1 if such a platform exists and 0 otherwise. Index (GTMI) 2022 2022 data) Are there government platforms that allow citizens to provide feedback (e.g., A digital platform to participate compliments, complaints, suggestions, info requests) on service delivery? This in the enhancement of the GovTech Maturity captures the availability of a platform that allows citizens to provide feedback, make quality of public service delivery Index (GTMI) 2022 suggestions, collect information, or complain about the delivery of public services. It I31(2020 and 2022 data) also takes 1 if such a platform exists and 0 otherwise. This variable indicates whether at least one of I30 or I31 exists for the country. It Authors’ computation GovTech (2020 and 2022 data) takes 1 if at least I30 or I31 exists and 0 otherwise. based on GTMI 2022 It measures the level of civil society participation in policymaking and political processes, with a focus on major Civil Society Organizations (CSOs) being routinely consulted by policymakers, the level of people's involvement in CSOs, the extent to Varieties of Civil Society Participation which women participate, and the process for nominating legislative candidates Democracy (V-Dem) within party organizations. The CSPI ranges from 0 to 1 and can be transformed into three different ordinal versions with three (_3C), four (_4C), and five (_5C) levels. It measures how wide and how independent are public deliberations, when Varieties of Engaged society important policy changes are being considered. Democracy (V-Dem) The Polity2 index is a revised and combined version of the POLITY score index, which Democracy (Polity 2) captures the spectrum of the authority of the political regime on a scale from -10 Polity V (hereditary monarchy) to 10 (consolidated democracy). Presented in constant 2017 international USD, it refers to gross domestic product that is converted to international dollars using purchasing power parity rates, which GDP per capita PPP in 2017 equalize the purchasing power of different currencies. The calculation includes the World Development International USD constant sum of gross value added by all resident producers in the country plus any product Indicators (WDI) taxes and minus any subsidies not included in the value of the products, without deductions for depreciation or depletion of natural resources. Fixed-broadband subscriptions refers to fixed subscriptions to high-speed access to The International Fixed broadband subscriptions the public Internet (a TCP/IP connection); at downstream speeds equal to; or greater Telecommunication per 100 inhabitants than; 256 kbit/s. The variable represents the ratio of Fixed broadband subscribers to Union (ITU) population, multiplied by 100. It measures the extent to which public sector employees grant favors in exchange for bribes, kickbacks, or other material inducements, and how often do they steal, Varieties of Public sector corruption embezzle, or misappropriate public funds or other state resources for personal or Democracy (V-Dem) family use. Population aged between 15 and World Development Percentage of the population aged between 15 and 44 years. 44 years Indicators (WDI) 24 Definition Description Source It indicates how political power is distributed in a country based on various social groups such as caste, ethnicity, language, race, region, and religion. It is contextual Equal distribution of political and cross-cutting, and individuals can belong to multiple groups. It ranges from 0 to Varieties of power 4, with 0 indicating political power monopolized by a minority group and 4 indicating Democracy (V-Dem) equal political power among all social groups or the irrelevance of social group characteristics to politics. The HCI calculates the contributions of health and education to worker productivity. The final index score ranges from zero to one and measures the productivity as a World Development Human capital index future worker of child born today relative to the benchmark of full health and Indicators (WDI) complete education. It refers to the conflict intensity level in the dyad per calendar year. Two different UCDP/PRIO Armed Conflict (intensity level) intensity levels are coded: minor armed conflicts and wars. Conflict Dataset UCDP/PRIO Armed Conflict (cumulative intensity) It refers to the intensity of the conflict, taking into consideration the conflict history. Conflict Dataset This variable was developed by the authors through the World Bank 2020 List of The World Bank FY20 Fragile and Conflict-affected Situations. A value of 2 corresponds to a high-intensity List of Fragile and Conflict-affected situations conflict; 1 represents a medium-intensity conflict; and 0 indicates the absence of Conflict-affected conflict. Situations This variable was developed by the authors through the World Bank 2020 List of The World Bank FY20 Fragile and Conflict-affected Situations. A value of 3 corresponds to a high-intensity Fragile and Conflict-affected List of Fragile and conflict; 2 represents a medium-intensity conflict; 1 indicates the absence of conflict situations Conflict-affected but high institutional and social fragility; and 0 indicates the absence of both conflict Situations and high institutional and social fragility. Government Effectiveness captures perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility Worldwide Government effectiveness of the government's commitment to such policies. Estimate gives the country's score Governance Indicators on the aggregate indicator, in units of a standard normal distribution, i.e., ranging from approximately -2.5 to 2.5. World Bank country It takes 1 if the country is a developing country (or not a high-income country) and 0 classification (WDI) Developing countries’ dummy otherwise. and authors’ computation Authors’ computation based on GTMI 2022 It takes 1 if the country is an African country having implemented a GovTech platform GovTech Africa and World Bank and 0 if it is an African or a non-African country without GovTech platform. country classification (WDI) 25 Annex B: List of countries included in the estimates Adopter countries Non-Adopter countries Afghanistan Kyrgyz Republic Algeria Namibia Albania Lithuania Azerbaijan Nepal Bahrain Angola Luxembourg Barbados Nicaragua Argentina Belarus Niger Malaysia Armenia Malta Benin Nigeria Bhutan Australia Mauritius Bosnia Panama Austria Mexico Botswana Papua New Guinea Bangladesh Moldova Bulgaria Paraguay Belgium Bolivia Mongolia Burkina Faso Poland Burundi Brazil Montenegro Cabo Verde Romania Cameroon Central African Republic Cambodia Morocco Chad Russian Federation Canada Netherlands China Rwanda New Zealand Senegal Chile North Macedonia Comoros Seychelles Congo, Dem. Rep. Colombia Norway Congo, Rep. Sierra Leone Côte d’Ivoire Oman Costa Rica Slovak Republic Czechia Djibouti Croatia Equatorial Guinea Solomon Islands Eritrea Somalia Cuba Pakistan Eswatini South Sudan Cyprus Peru Ethiopia Sri Lanka Sudan Suriname Syrian Arab Republic Denmark Philippines Gabon São Tomé and Príncipe Dominican Republic Portugal Gambia, The Taiwan, China Ecuador Qatar Guatemala Tajikistan Guinea Egypt, Arab Rep. Saudi Arabia Guinea-Bissau Timor-Leste El Salvador Serbia Guyana Togo Trinidad and Tobago Estonia Singapore Haiti Turkmenistan Ukraine Venezuela RB Hungary West Bank and Gaza Fiji Slovenia Iceland Yemen, Rep. Finland South Africa Iran, Islamic Rep. Zimbabwe France Spain Iraq 26 Ireland Korea, Dem. People’s Georgia Sweden Rep. Germany Switzerland Kenya Ghana Tanzania Kuwait Greece Thailand Lao PDR Honduras Hong Kong SAR, China Tunisia Latvia India Türkiye Lebanon Indonesia Uganda Lesotho Liberia Israel United Arab Emirates Libya Italy United Kingdom Madagascar Malawi Jamaica United States Maldives Japan Uruguay Mali Uzbekistan Jordan Vanuatu Mauritania Kazakhstan Viet Nam Mozambique Korea, Rep. Kosovo Zambia Myanmar 27