Corruption, ICT and Military Spending in Sub-Saharan Africa

ABSTRACT

It has been argued that large military spending is often a crucial constraint in Sub-Saharan Africa (SSA) countries, and high corruption has enabled these countries to keep such large military spending. One possible remedy is to promote information and communication technology (ICT), which may cause the anti-corruption policies to become more effective in increasing transparency for the public and reducing unnecessary and abusive military spending. This study discusses the nexus among corruption, military spending, and ICT with a panel data of 48 SSA countries from 2003 to 2015. Our analysis reveals that when ICT prevails, military spending is negatively associated with the control of corruption, so that anti-corruption policies with the usage of ICT could reduce the extravagant budget allocation to military spending. Policymakers should associate sound ICT policy with traditional anti-corruption factors to bring more transparency and less corrupt behaviors in the military department. The reduction in the misuse of military allocation enables the country to increase nonmilitary public spending to face sustainable development challenges for the sake of the population.

KEYWORDS:

• Corruption
• ICT
• military spending
• Sub-Saharan Africa

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1. This assertion is confirmed by the United Nations High Commissioner for Refugees (UNHCR) in its 2019 Global Trends Report showing that the number of people running away from peace turbulence was more than 70 million in 2018. This number is the highest level registered by the UNHCR since its foundation in 1950.

2. From the customer’s perspective, the use of ICT encourages private and public agents to provide better quality services through the reduction of cost and waiting time (Breen 2000), improve transparency in jobs opportunities (Brueckner 2005), enhance government commitment (Goodwin 2005) and accountability (Rose 2004; Wong and Welch 2004), and support community development (Hammerman 2005). From the producer’s perspective, ICT improves productivity (Yang and Rho 2007), helps good governance (Basu 2004), and enhances trust and cooperation of citizens through tax payments and sustainability of public goods and services.

3. E-government can be defined as the use of ICT to work more effectively, share information, and deliver better services to the public by the government (Elbahnasawy 2014). A high level of telecommunication infrastructure, like internet and mobile cellular networks, is the prerequisite to adopt e-governance, but this decision is mainly political since the high level of internet and mobile cellular networks may not always associate with e-government (Kanyam, Kostandini, and Ferreira 2017).

4. It is important to note that this study considers the contemporaneous effect of ICT without careful consideration of the dynamic effect. Some scholars argue that it takes some time to first affect traditional anti-corruption factors, such as transparency and governance, and then affect the outcome of interest. Although we admit that our model specification does not incorporate such dynamic aspects, we believe that our model can provide some important implication about ICT and military expenditure.

5. The GMM estimation technique is used mainly for three main reasons: (i) identification; (ii) simultaneity, and (iii) exclusion restrictions (Asongu, Nwachukwu, and Orim 2018). The number of groups (countries or cross-sections) (N equals 48) is large enough compared with the periodicity per cross-section (yearly observation: T equals 13). This identification technique is consistent with a panel data structure by including time and time-invariant specific effects in the specification. The difference estimator technique of GMM (Arellano and Bond 1991) removes the time-invariant effects (country specific effects or simply fixed effects) but does not for other biases (time-specific effects, omitted time and time-invariant specific effects). The regression procedure mitigates the endogeneity bias among and between the studied variables (regressors) and omitted variables (included in the error term). The difference GMM uses the following moment conditions (Anyanwu, Anyanwu, and Cieślic 2021):

$\mathrm{E}[{\mathrm{y}}_{\mathrm{i},\mathrm{t}-\mathrm{s}}\left({\epsilon }_{\mathrm{i},\mathrm{t}}-{\epsilon }_{\mathrm{i},\mathrm{t}-1}\right)$] = 0 for s ≥ 2; t = 3, …, T,

$\mathrm{E}[{\mathrm{X}}_{\mathrm{i},\mathrm{t}-\mathrm{s}}\left({\epsilon }_{\mathrm{i},\mathrm{t}}-{\epsilon }_{\mathrm{i},\mathrm{t}-1}\right)$] = 0 for s ≥ 2; t = 3, …, T.

6. The additional moment conditions are as follows (Anyanwu, Anyanwu, and Cieślic 2021):

$\mathrm{E}[\mathrm{\Delta }{\mathrm{y}}_{\mathrm{i},\mathrm{t}-\mathrm{s}}\left({\epsilon }_{\mathrm{i},\mathrm{t}}+{\alpha }_{\mathrm{i}}\right)$] = 0 for s = 1; t = 3, …, T,

$\mathrm{E}[\mathrm{\Delta }{\mathrm{X}}_{\mathrm{i},\mathrm{t}-\mathrm{s}}\left({\epsilon }_{\mathrm{i},\mathrm{t}}+{\alpha }_{\mathrm{i}}\right)$] = 0 for s = 1; t = 3, …, T.

7. We notice that our model specification may suffer from various endogeneity problems, although system GMM may solve part of them. For example, there are some time-variant unobserved variables that are correlated with ICT and military expenditure. It is generally difficult to find valid IVs to meet all requirements, including external validity. Given the difficulty, the subnational literature on the role of ICT and governance uses IVs for ICT (Guriev, Melnikov, and Zhuravskaya 2021), which could be helpful for further robustness checks in future research.

8. Although the average level of ICT (INTERNET) over sample countries and sample periods is low at 7.7, the level is still diverse over the range from 0.03 to 54.26. Figure 1 shows margin plots for the cases of the high level of ICT (INTERNET = 50) and the low level of ICT (INTERNET = 0), based on model (4). It is easily observed that when ICT prevails in a country (high INTERNET), its military expenditure is negatively associated with the control of corruption.

Marginal plots of the effects of control of corruption on military expenditure.

Note: Shaded regions show the 95% confidence intervals.

Marginal plots of the effects of control of corruption on military expenditure.