A comparison of regime-switching temperature modeling approaches for applications in weather derivatives

• Weather derivatives are used to hedge weather related financial risks.
• Three regime-switching models are proposed to capture the temperature dynamics.
• A model with a Brownian motion and a mean-reverting process performs the best.
• This model is then applied to price weather options based on heating degree days.

A comparison of regime-switching approaches to modeling the stochastic behavior of temperature with an aim to the valuation of temperature-based weather options is presented. Four models are developed. Three of these are two-state Markov regime-switching models and the other is a single-regime model. The regime-switching models are generated from a combination of different underlying processes for the stochastic component of temperature. In Model 1, one regime is governed by a mean-reverting process and the other by a Brownian motion. In Model 2, each regime is governed by a Brownian motion. In Model 3, each regime is governed by a mean-reverting process in which the mean and speed of the mean-reversion remain the same, but only the volatility switches between the states. Model 4 is a single-regime model, where the temperature dynamics are governed by a single mean-reverting process. All four models are utilized to determine the expected heating degree days (HDD) and cooling degree days (CDD), which play a crucial role in the valuation of weather options. A four-year temperature dataset from Toronto, Canada, is used for the analysis. Results demonstrate that Model 1 captures the temperature dynamics more accurately than the other three models. Model 1 is then used to price the monthly call options based on a range of strike HDD.