The impact of regional institutional quality on firm innovation: evidence from Africa

ABSTRACT

The main aim of this study is to empirically examine the role of regional institutional quality, and compare the relative importance of three different institutional arrangements, namely regulatory quality, rule of law, and corruption, on propensity of firm innovation in African countries. To this end, we used the most recent World Bank Enterprise Survey (WBES) of firms from 19 African countries. Our empirical investigation using multilevel logistic regression has drawn three important conclusions. First, there is no significant relationship between firm-specific factors, such as age, ownership, and exporting, and firm innovation. Nonetheless, firms’ propensity to innovate is significantly determined by the size of the firm, R&D expenditures, on-the-job training (OJT), and capability of managers/owners to adopt new technologies. Second, the innovation performance of firms is significantly affected by regional institutional quality. Third, among different regional institutional arrangements, reducing corruption and maintaining rule of law appeared to matter the most for firm innovation.

KEYWORDS:

• Innovation
• regional institutions
• multilevel logistic regression
• Africa

Acknowledgements

The authors are very grateful to two anonymous referees for their insightful comments and suggestions which helped to improve the paper. We also would like to thank the Editor for helpful comments and suggestions. All remaining errors are of ours.

Disclosure statement

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

Notes

1 See for instance Barasa et al. (2017) and Chadee and Roxas (2013).

2 For instance, (Lu, Png, and Tao 2013) argued that although the de jure property right is the same at the national level, de facto property right varies across Chinese cities. These differences, according to them, affect the performances of firms. See Acemoglu, Johnson, and Robinson (2005) for detailed information regarding de facto and de Jure institutions.

3 See, for instance, Pattit, Raj, and Wilemon (2012) for USA, Wu, Wu, and Zhuo (2015) for China, Blind (2012) for 21 OECD countries, Oluwatobi et al. (2015) for 40 African countries.

4 According to Cooke (2001) a region is defined as a sub-national entity that has some cultural or historical homogeneity but had some power to intervene and support economic development, particularly innovation.

5 Detailed information on the data is available on WBES website https://www.enterprisesurveys.org/.

6 Since the regions are also nested in countries, three-level model could be an alternative. However, we preferred two-level model for two reasons: First, country characteristics, which may influence innovativeness of the studied countries are more or less the same. For instance, economically they all belong to either low or lower-middle income categories. All of them are also characterized with poor political and judicial institutions etc. Second, some of these countries are too small to divide in regions, like Swaziland, Lesotho, and Namibia. Therefore, in these countries all surveyed firms are in the same region which is the country itself.

7 https://www.africa-business.com/features/mobile_phones_africa.html and https://qz.com/africa/1206935/whatsapp-is-the-most-popular-messaging-app-in-africa/

8 ICC is computed by first fitting the empty multilevel logistic model

$\log \left({\displaystyle \frac{p}{1-p}}\right)={\text{β}}_0+u_j$

and then calculating

$\mathrm{I}\mathrm{C}\mathrm{C}={\displaystyle \frac{\mathrm{v}\mathrm{a}\mathrm{r}\left(u_j\right)}{\mathrm{v}\mathrm{a}\mathrm{r}\left(u_j\right)+\left({\pi }^2/3\right)}}$.

ICC value ranges from 0 to 1 and indicates whether multilevel modelling is appropriate. ICC closer to 0 indicates that individuals within countries are no more similar than individuals from other countries. Therefore, there is no variance to explain at the country level and multilevel modelling is of no use. Conversely, if the ICC is close to 1, it indicates that there is no variance to explain at the individual level.

9 Foreign ownership is only significant at 10% significance level.

10 31 industrial sectors were included in the survey.

11 Note that similar robustness check also conducted for disaggregated regional institution indicators-corruption, rule of law, and regulatory quality- and the estimation results are similar to Table 7.