The financial development–energy consumption nexus revisited

ABSTRACT

This study investigates the impact of financial development on energy consumption for a panel of 32 high-income countries over the period 1990–2014. Unlike previous studies, this analysis incorporates the bond market alongside the banking sector and the stock market in defining financial development. This study utilizes principal component analysis to construct indices for each of the three financial sectors (banking sector, stock market, and bond market) as well as an overall financial development index. Moreover, unlike previous studies, this study also implements heterogeneity panel estimation techniques. The results of this study reveal the absence of a statistical relationship between the overall financial development index and energy consumption. However, it is observed that an increase in the stock market index yields a slight decline in energy consumption. Policy implications pertaining to the empirical results are also discussed.

KEYWORDS:

• Energy consumption
• financial development
• heterogeneous panel techniques
• principal components

JEL Codes:

• C33
• C38
• O16
• Q43

Notes

¹ There is also a parallel empirical literature on the relationship between financial development and emissions. See Tamazian et al. (2009), Tamazian and Rao (2010), Yuxian and Chen (2010), Jalil and Feridum (2011), Zhang (2011), Al-mulali and Sab (2012a;b), Lee (2013), Omri (2013), Ozturk and Acaravci (2013), Shahbaz (2013), Shahbaz et al. (2013a;b;c;d), Boutabba (2014), Al-mulali et al. (2015a;b), Li et al. (2015), Mugableh (2015), Saidi and Hammami (2015), Salahuddin et al. (2015), Ziaei (2015), Abbasi and Riaz (2016), Dogan and Sher (2016), Dogan and Turkekul (2016), Farhani and Ozturk (2016), Abbasi and Riaz (2016), Rafindadi (2016), and Shahbaz et al. (2016).

² See Thumrongvit et al. (2013) and Fanta and Makina (2016), among others, on the role of the bond market in economic growth.

³ Real oil prices were constructed by dividing the crude oil price by each country’s consumer price index (CPI 2010 = 100).

⁴ Despite the high correlation between West Texas and Brent real oil prices, we are aware that West Texas Intermediate crude oil prices would also be a good measure for the non-European countries. Thus, we estimated the same model using West Texas real oil prices to find no discernable difference in the results. The results using West Texas real oil prices are available upon request.

⁵ See principal component analysis in the next section for a more detailed information on the components of the sub-market indices.

⁶ The weights obtained from the principal component analysis with respect to each indicator and country are not reported, but are available upon request.

⁷ This result for real oil prices parallels the findings by Chang (2015).

⁸ In case of low-income countries, the difficulty rests with obtaining the appropriate bond market data.