On the interactions among environmental degradation, energy consumption and economic growth: a time–frequency analysis using wavelets

ABSTRACT

This paper aims to investigate the interactions between CO₂ emissions, economic growth and energy consumption in India for 1970–2017 within the time–frequency domain using wavelet techniques. The results show that the interaction between energy consumption and CO₂ emissions is strong and varies significantly across time–frequency bands, indicating that the energy consumption leads to environmental degradation. The negative effect of economic growth on CO₂ emissions is only statistically significant in the short run. In addition, the findings reveal a positive association between economic growth and energy consumption across time–frequency bands, demonstrating energy consumption increases economic growth, which in turn heightens the CO₂ emissions. There should be improved energy use and reduced CO₂ emissions.

KEYWORDS:

• Emissions
• Kuznets
• wavelet
• energy
• economy
• causality

Disclosure statement

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

Notes

1. Fourier analysis is mainly used for stationary time-series data; it applies the spectral tool to investigate the linkages between economic series. But, this method reflects only the frequencies and ignores the time information based on the time-series data. There are other Fourier techniques like the Windowed Fourier Transform analysis (WFT), which address the above-mentioned deficiencies, but as mentioned by Kaiser (2011), a potential drawback with the application of the WFT in time series analysis is that it suffers from the lack of accuracy and efficiency of simultaneous time–frequency localisation.

2. Wavelet coherence is used as a tool to identify whether the co-movement between each pair of economic variables is contagion or interdependence. Strong wavelet coherence at high frequencies indicates contagion. In contrast, strong wavelet coherence at low frequencies is identified as interdependence (Bodart & Candelon, 2009; Yang, Cai, Zhang, & Hamori, 2016).