Stochastic Control in Commodity & Energy Markets : Model Uncertainty, Algorithmic Trading, and Future Directions

I will provide an overview of cutting edge stochastic control problems related to commodity and energy markets. One of the common threads will be incorporating model uncertainty into valuation and trading of financial instruments. I will show how model uncertainty can be cast as a robust optimal control problem and demonstrate how derivative valuation is modified when agents account for it. As a second application, I will discuss how agents optimize their trading actions in an interconnected electricity market. When agents account for the price impact of their trades, I will show that they induce cointegration in prices. Moreover, when agents trade using market orders they incur costs due to the existence of a bid-ask spread and by orders walking through layers of the limit order book. Hence, as a third application, I will show how agents can act instead as liquidity providers and optimize the placement of their limit orders to benefit from the bid-ask spread. Finally, I will describe some open problems related to partial information, and multiple agents using mean-field games

The role of mathematical models in the energy industry

The future of energy trading in the UK and Europe: a single energy market?

Measuring returns to those who invest in energy through futures

It is commonly asserted that maintaining an ongoing position in energy futures subjects the investor to a gain or loss referred to as a ‘roll yield’, which refers to the difference in price across futures contracts at the time positions are closed and opened.   In fact, the roll yield as a gain or loss to an investor is mythical.  Futures investors earn or pay the price change that occurs while they hold a given contract, not any difference in prices across contracts.   However, the roll yield does contain useful and relevant information