Forecasting the joint distribution of Australian electricity prices using dynamic vine copulae

We consider the problem of modelling and forecasting the distribution of a vector of prices from interconnected electricity markets using a flexible class of drawable vine copula models, where we allow the dependence parameters of the constituting bivariate copulae to be time-varying. We undertake in-sample and out-of-sample tests using daily electricity prices, and evidence that our model provides accurate forecasts of the underlying distribution and outperforms a set of competing models in their abilities to forecast one-day-ahead conditional quantiles of a portfolio of electricity prices. Our study is conducted in the Australian National Electricity Market (NEM), which is the most efficient power auction in the world. Electricity prices exhibit highly stylised features such as extreme price spikes, price dependency between regional markets, correlation asymmetry and non-linear dependency. The developed approach can be used as a risk management tool in the electricity retail industry, which plays an integral role in the apparatus of modern energy markets. Electricity retailers are responsible for the efficient distribution of electricity, while being exposed to market risk with extreme magnitudes.