Modelling the relationship between future energy intraday volatility and trading volume with wavelet

Although the energy and stock markets are both characterized by volatility and liquidity, and there has been substantial research to explore the relationship between volatility and trading volume (TV) in stock markets, few researchers have investigated this relationship in energy markets. Moreover, studies that have explored this association within energy markets did not describe its nature or impetus. To redress this oversight, we investigate this relationship using intraday data from the oil and gas markets – the most liquid energy markets in the world. In this way, the current article extends the previous studies through the use of a frequency approach to propose an original analysis of the relationship between volume and volatility. More specifically, we employ a continuous wavelet transform to identify the lead–lag phase between volatility and volume. This framework supplants usual time series modelling, as it uses a measure of coherence for different frequencies and time-scales to capture further changes and time variation in the volume–volatility relationship. Our results provide supportive evidence for the well-known positive relationship between realized volatility and TV, thereby supporting the mixture distribution hypothesis. In particular, our results show that volume causes volatility only during ‘turbulent times’, while volatility causes volume during ‘good times’. Furthermore, there is no relationship between volume and volatility in the long term, due to the absence of noise traders and liquidity traders in the long run. These findings are helpful for investors and policymakers as they contribute to better forecast the TV and price volatility during turbulent and calm periods and over several investment horizons.