Does digitalization improve government effectiveness? Evidence from developing and developed countries

ABSTRACT

This study aims to analyze the effect of digitalization on government effectiveness in developing and developed countries. It uses a panel methodology with data from 138 countries between 2006 and 2016. The results suggest that a government’s use of information and communication technologies (ICT) improves its effectiveness in both developing and developed countries. However, this effect is stronger in developed than in developing countries. Moreover, we find that the effect of aggregate ICT use by individuals, businesses, and government on overall government effectiveness is greater than that of individual use by each individually. The results are robust after several tests (reverse causality, dynamic effect, sensitivity analysis, heterogeneities, and alternative measurements). These results highlight the fact that governments could fully benefit from digitalization by adopting policies that promote access to and use of ICT at all levels of the economy, that is, the government itself, businesses, and individuals.

KEYWORDS:

• Digitalization
• government effectiveness
• developed countries
• developing countries

JEL CLASSIFICATION:

• H1
• H4
• H7
• O3
• O5

Acknowledgement

We are indebted to the Editor and an anonymous referee, in addition to Jean-François BRUN, Damien CUBIZOL, Abdramane CAMARA, Fayçal SAWADOGO, Mahamady OUEDRAOGO, SILUE Tarna and Seydou COULIBALY for their precious comments and suggestions. We are also grateful to the participants at the CERDI PhD students’ seminars for their interesting and useful discussions. This work was supported by the LABEX IDGM+ (ANR-10-LABX-14-01) within the program “Investissements d’Avenir” operated by the French National Research Agency (ANR).

Disclosure statement

The author declares that he has no known competing financial interests or personal relationships that could have influenced the work reported in this paper.

Notes

¹ To analyze the effect of government digitalization on the quality of governance, Dobrolyubova, Klochkova, and Alexandrov (2019) first performed a correlation test and then a Granger causality test. The limitations of this approach are that, first, correlation does not necessarily explain the existence of an effect or causality. This is confirmed by the causality test carried out by the author, which shows an absence of causality between digitalization and the quality of governance. Correlation measures the strength of the link between the variables. Furthermore, Granger causality does not test for a true cause-and-effect relationship; it tests the order of arrival of one variable relative to another in the time series. Thus, Granger causality does not necessarily imply a causal relationship in the real sense. Finally, although these two methods allow the identification of a possible correlation as well as a causal link, it is impossible to determine the extent of the effect of a factor on another. Therefore, despite having the same objective as our study, this study is limited by the methodologies used.

² It should also be noted that Dobrolyubova et al. (2019) also consider the government digitalization, using E-government development, Online Service, and E-participation indexes, along with the proportion of citizens using the Internet to submit completed forms via government websites.

³ A list of countries with income group is provided in Appendix A.

⁴ Dutta and Mia (20072009, 2010,2011); Dutta, López-Claros, and Mia(2008); Dutta and Bilbao-Osorio (2012); Dutta, Geiger and Lanvin (2015) Baller et al. (2016); Bilbao-Osorio et al. (2013, 2014). The report was published with a one-year delay before 2012. However, from 2012 onwards, each report covers the year of its publication. Because of the lack of data for 2011, we have estimated it by using the average of 2010 and 2012.

⁵ Appendix B.1 presents the ICT variables uses for each usage sub-index.

⁶ Appendix B.2 presents definitions and sources of all variables used in this study, while Appendix B.3 gives summary statistics.

⁷ See Torres-Reyna. (2007) for more details about the Hausman test and the time fixed effect test.

⁸ Here, the interpretation is done considering that the variable is taken in logarithm. An increase of the explanatory variable of 1% leads to a variation of the explained variable of $\bar{X}$ units.

⁹ We also consider digitalization and the lag of dependent variables as endogenous in the system GMM.

¹⁰ To capture the size of government, the final consumption of government relative to GDP Is used, while the proportion of women in parliament is used to capture the gender composition of government.

¹¹ For more details on the literature on the effect of natural resources on the quality of governance and institutions, see Busse and Gröning (2013).

¹² For African and European countries, the geographical grouping has been assigned according to the football confederation to which they are affiliated. This allows to easily classify countries that straddle two continents or belong to a region different from their continent. There are two countries in Oceania (namely Australia and New Zealand) that are not considered in this heterogeneity analysis due to the small number of observations. However, the result does not change if they are added to the group of Asian countries.

¹³ When disaggregating South and North American countries, the effect remains positive and non-significant for North American countries (there are 15 countries and islands, including the USA and Canada), while it becomes significant for South American countries. Therefore, it is likely that the non-significant effect comes from North American countries.

¹⁴ We perform an additional geographic analysis based on the WB region groups. The results are presented in the Appendix C

¹⁵ The results are displays in appendix D.1 and D.2, for developed and developing countries, respectively.

¹⁶ A definition of these indicators is given in appendix B.2.

¹⁷ As EGDI is available over two years, we replaced data for gap year by the mean of the year after and the year before.

¹⁸ Several reasons may explain this effect in developing countries. First, most of these countries are in the beginning stages of using these services (e-government, e-services, and e-participation), therefore it takes time for users to adapt and take full advantage of them. Furthermore, Jacolin, Massil Joseph, and Noah (2019) and Brun et al. (2020) explain that experience matters in the use of digital services. Second, the implementation of these services needs to be supported by policies to promote and sustain them, and thus by political will. Third, most developing countries face either technical difficulties and/or user reluctance. Thus, even if the services exist, the expected effects will not be achieved. Finally, Dobrolyubova, Alexandrov, and Yefremov (2017) argue that the insignificant effect of EGID indices can be explained by the fact that in addition to not reflecting the actual level of digitalization, they do not capture all dimensions of digitalization.

Additional information

Funding

The work was supported by the Agence Nationale de la Recherche [LABEX IDGM+ (ANR-10-LABX-14-01)]