Examining the role of E-government in controlling corruption: A longitudinal study

Abstract

Acknowledging the global relevance of controlling corruption, our study contributes to the existing body of literature by exploring the relationship between E-Government (EGOV) and Corruption Control at the EGOV subindices level. Toward this, we draw upon theoretical insights from Resource-Based View (RBV), Media Literacy Theory, Agency Theory, Information Quality, and Transaction Cost Theory to derive a set of testable hypotheses. Using a panel dataset comprising 102 countries, we employ a Linear Mixed Effects (LME) estimation method to thoroughly investigate and uncover the relative effectiveness of EGOV subindices in controlling corruption. Accordingly, we draw implications for theory, praxis and future research.

Introduction

Extant literature unambiguously recognizes corruption as a significant barrier hindering the socioeconomic development of a country. Corruption obstructs the smooth functioning of government machinery and equitable distribution of public goods and services, thereby resulting in inefficiencies such as irresponsible government spending and diversion of public funds [1], [2], [3]. It also involves the misuse of power by public officials and/or governing authority [4], which leads to the manifestation of the principal-agent problem arising out of information asymmetry, and nonalignment of incentives [5]. According to the World Economic Forum [6,7], the global annual cost of corruption amounts to about 5% of global GDP, which is approximately equal to USD 3.6 trillion. The ambit of negative externalities caused by corruption extends beyond the social, economic, and political domains to encompass sustainability issues like aggravating environmental degradation [8]. In the wake of its adverse impact of such high magnitude, wide focus, and immense scale, controlling corruption at the national level becomes imperative for governments worldwide.

E-Government (EGOV), which uses Information and Communication Technologies (ICTs) to improve public service delivery, has been widely touted as a useful anticorruption tool [9] in the extant literature [10], [11], [12], [13]. The ever-growing body of research [5,8,[14], [15], [16], [17], [18]] on the “E-Government – Corruption Control” (EGOV-COR) relationship stands testimony to this claim. However, the relevant literature is somewhat limited in explaining the EGOV-COR relationship in its entirety due to its partisan emphasis on cross-sectional analyses covering only a selected set of countries [15,17]. Moreover, these studies primarily use datasets that are either time-averaged over a few years or correspond to a particular year under consideration [15,17]. Such restricted analyses lead to several drawbacks. Firstly, both EGOV and corruption levels of a country are dependent on temporal changes; the results obtained from the analysis for a specific year under consideration often may not hold after a couple of years down the line. Such snapshots portray only a period piece in the relationship, which is not much useful for policymakers in taking long-term decisions [19]. Secondly, the use of cross-sectional datasets limits researchers’ ability to model lagged effects, which are extremely crucial considering that policy interventions in the form of EGOV usually take time to exhibit their intended effects. Thirdly, the empirical frameworks analyzed in the existing studies are not suitable to model the within-country factors exhaustively, leading to challenges such as model misspecification and biases in parameter estimates [19]. Therefore, in our opinion, longitudinal studies, which may provide meaningful insights into the validity of the continued effectiveness of EGOV as an anticorruption tool by incorporating the temporal variations in the associated factors, are very much needed in enriching the deeper understanding of the EGOV-COR connection. Belanger and Carter's call for conducting more longitudinal studies in the EGOV research arena also bolsters our claim [20]. To fill this gap, we propose to use a panel dataset of 102 countries (Appendix 1) over a period of about a decade (2008–2019) to conduct a Linear Mixed Effects (LME) analysis with a view to assessing the temporal behavior of the EGOV-COR relationship over a span of nearly 12 years. LME is particularly applicable in our study as it models lagged effects and accounts for both unobserved as well as observed country heterogeneities [21] arising due to within-country factors, such as social, cultural, political, and other country-specific idiosyncrasies [22].

Interestingly, while performing the above analysis, we notice a counterintuitive trend for several countries [23]. In some of these nations, quite surprisingly, corruption rises despite the increase in the EGOV development. In some other cases, even after maintaining a high level of EGOV development over the years, they are experiencing an uptrend in their corruption levels. For instance, Australia and Canada have witnessed a rise in corruption during the period of our analysis. This phenomenon, however, cannot be considered universal because South Korea and Estonia, which exhibit similar EGOV development as that of Australia and Canada, have been able to reduce corruption over the considered period. Unfortunately, the existing empirical literature provides no convincing explanation for this apparent anomaly. Our apprehension is that, since these studies have investigated the EGOV-COR relationship at the composite index level (viz. EGOV Development Index (EGDI) level), they have missed out on the individual impacts of each of the three dimensions (also known as subindices) of EGDI, and hence could not explain if the overall impact is due to equal contribution from all the three subindices or there are dissimilarities in the nature of contributions by individual subindices. Although some studies have employed one of three subindices of EGDI (mostly Online Services Index (OSI) in isolation), they have missed out on the individual impacts by the remaining two subindices of EGDI on Corruption Control. The three subindices of EGDI are (1) Human Capital Index (HCI) – which captures the country-level human capabilities, (2) Telecommunications Infrastructure Index (TII) – which captures the status of telecom infrastructure in a country, and (3) OSI – which determines the scope and quality of online services (OS) [24]. We have observed in our analysis that these subindices differ considerably in their anticorruption effectiveness, which may also explain the existence of widely varying corruption levels in countries with almost identical EGDI. For instance, while Uruguay and Kazakhstan had almost similar EGDI (0.724 and 0.725, respectively) in 2016, their corruption levels differed significantly.¹ Therefore, further investigation is needed to develop a more nuanced understanding of EGOV's effectiveness on Corruption Control at the level of its three subindices. To the best of our knowledge, the EGOV-COR relationship is relatively unexplored down to this level. So, we plan to go inside EGDI to extend the depth of our study with an objective to address this lacuna in extant research. Also, extant studies have not investigated the variations in the impact of the subindices across different time-lags, except one study that has assessed their impact for a one-year lag [5]. Our study further attempts to address this research gap in the existing body of EGOV-COR literature. Fig. 1 represents a conceptual framework for our study and the theoretical perspectives used therein. As indicated in the figure, we have drawn insights from Resource-Based View (RBV), Agency Theory, Theory of Media Literacy, Information Quality stream of literature, and Transaction Cost Theory to conceptualize the link between EGOV and Corruption Control. Specifically, we invoke these theories to explicate the roles of EGOV, HCI, TII, and OSI in the effective management and consumption of information within the broader public service delivery system. We will explain these theories and their relevance to the current research in finer details in the upcoming sections. Further, we account for country heterogeneities by using LME, which incorporates both fixed effects as well as random effects for each country. Fixed effects are adept at controlling time-invariant stable characteristics of countries by focusing only on within-country variations, while random effects are helpful in capturing between-country heterogeneities. We firmly believe that such in-depth analyses, at the end of the day, will certainly help the policymakers in allocating EGOV resources more effectively along HCI, TII, and OSI dimensions, thereby achieving better anticorruption outcomes in the long term.

The rest of the paper is organized as follows: Section 2 contains a review of relevant literature followed by hypothesis formulation, and Section 3 elaborates on the research framework, data, and methodology. Section 4 describes the main results and findings, and Section 5 presents the discussion, including implications, limitations, and future work. Section 6, finally, concludes the paper.