Financial depth, internet penetration rates and economic growth: country-panel evidence

ABSTRACT

The article investigates causal relationships between internet penetration rates, financial depth and per capita economic growth in the Next-11 countries. Using panel vector autoregressive (VAR) approaches, our empirical results show that these variables are cointegrated. Moreover, we find bidirectional causality between internet penetration rates and economic growth, and between financial depth and economic growth in the short run.

KEYWORDS:

• Financial depth
• internet penetration rates
• economic growth
• panel VAR
• Next-11 countries

JEL CLASSIFICATION:

• L96
• O32
• O33

Notes

¹ Despite these findings, Indjikian and Siegel (2005) argue that developing nations have certain hurdles to overcome in effectively utilizing ICT for improving economic performance. First, there is a lack of understanding/knowledge of best practices in using ICT. Second, there are deficiencies in ICT skills in the workforces of developing countries.

² See, for instance, Samargandi, Fidrmuc, and Ghosh (2015), Pradhan, Arvin, Norman, et al. (2014), Jedidia, Boujelbène, and Helali (2014) and Wolde-Rufael (2009).

³ See, for example Pradhan, Arvin, Bahmani, et al. (2014) for an explanation of how PCA may be utilized.

⁴ All the monetary variables in our analysis, including gross domestic product, are measured in real US dollars.

⁵ See, for instance, Pradhan, Arvin, Norman, et al. (2014) for a detailed explanation of this procedure.

⁶ The estimation process follows Holtz-Eakin, Newey, and Rosen (1988).

⁷ See, for example Dolado and Lütkepohl (1996) for an explanation of this test.

⁸ This is true for all the three specific measures of internet penetration rates: penetration rates of internet users (INTUSE), penetration rates of internet servers (INTSER) and overall internet penetration rates (INTCOM) – represented as Specifications 1–3 in our tables.

⁹ Lack of short-run causality is determined by inspecting the Wald-statistics.

¹⁰ Lack of long-run causality is evident by the insignificance of lagged ECTs.

¹¹ INTPEN is used for INTUSE, INTSER and INTCOM – as commented earlier.

¹² This is congruent with the findings of Pradhan, Arvin, Bahmani, et al. (2014) and Pradhan, Bele, and Pandey (2013).

¹³ This is consistent with the findings of Pradhan, Arvin, Norman, et al. (2014), Wolde-Rufael (2009), Darrat, Elkhal, and McCallum (2006), Dritsakis and Adamopoulos (2004) and Calderon and Liu (2003).

¹⁴ In these figures internet penetration rates are captured in three ways: penetration rates of internet users, penetration rates of internet servers and overall internet penetration rates (the composite index), as described earlier.