With the liberalization of global power markets, modeling of exchange traded electricity contracts has attracted significantly the attention of both academic and industry. In this paper we offer an overview of the most common deseasonalization techniques and modeling approaches in the literature. We extract the deterministic component of EEX Phelix hourly electricity prices and we discuss different financial and time series models for their stochastic component. Additionally we apply Extreme Value Theory (EVT) to investigate the tails of the price changes distribution. Generally our results suggest EVT to be of interest to both risk managers and portfolio managers in the highly volatile electricity markets.
In Basel II the regulators stress the importance of finding realistic volumes models for non-maturing accounts (NMAs), given their cash-flow uncer- tainty due to optionality. Focusing on Swiss savings accounts, we identify their seasonal pattern and we derive their sensitivity to market rates and to relevant macroeconomic factors. We derive a realistic volumes model, that allows for cash-flow forecasting, which is essential for liquidity management in banks.
Retail banks usually apply simple linear regression models for describing the dynamics of the deposit rates of non-maturing accounts (NMA) like savings deposits. Thus, typical patterns like asymmetry or rigidity that banks follow when adjusting their deposit rates are ignored. This is insofar surprising, as the asymmetric deposit rate adjustment affects the pricing of embedded options for NMA. In this work we contribute to the elimination of these inconsistencies. Based on data for deposit rates from a representative sample of Swiss banks we provide a strong evidence for both asymmetric adjustment and rigidity pattern. Our proposed modeling approaches reveal that the strategies of Swiss banks to adjust deposit rates are regime dependent. In times of market stress, Swiss banks are tight to market rates; however, in normal regimes this is not observed.
The paper presents a convergence proof for a broad class of sampling algorithms for multistage stochastic linear programs in which the uncertain parameters occur only in the constraint right-hand sides. This class includes SDDP, AND, ReSa, and CUPPS. We show that, under some independence assumptions on the sampling procedure, the algorithms converge with probability 1.
