Modelling Electricity Swaps with Stochastic Forward Premium Models
提出一种电力互换定价新模型，用随机远期溢价和MNIG分布刻画风险，在欧洲能源交易所数据上优于四种竞争模型，并更准确地估计VaR尾部风险。
We present a new model for pricing electricity swaps. Two general factors affect contracts but unique risk elements affect each contract. General factors are average swap prices and deterministic trend-seasonal components, and unique elements are forward premiums. Innovations follow MNIG distributions. We estimate the model with data from the European Energy Exchange. The model outperforms four competitors, both in in-sample valuation and in out-of-sample forecasting, and in fitting the term structure of volatilities by market segments. Competitor models are (i) diffusion spot prices, (ii) jump-diffusion spot prices with time dependent volatility, (iii) HJM-based and (iv) Levy multifactor model with NIG distributions. Value-at-Risk measures based on normality strongly underestimate tail risk but our model gives estimates that are more exact.Keywords: Electricity swaps; Stochastic forward premium; Multivariate Normal Inverse Gaussian distribution; Levy processes