Since the introduction of the first weather derivative in the United-States in 1997, a significant number of work was directed towards the pricing of this product and the modelling of the daily average temperature which characterizes most of the traded weather instruments. The weather derivatives were created to enable companies to hedge against climate risks. They respond more to a need to cover seasonal variations which may cause loss of profits for companies than to a coverage need in property damage. Despite the abundance of work on the topic, no consensus has emerged so far about the methodology for evaluating weather derivatives. The major problems of these instruments are on one hand, they are based on an meteorological index that is not traded on financial market which does not allow the use of traditional pricing methods and on the other hand, it is difficult to get round this obstacle by susbtituting the underlying for a linked exchanged security since the weather index is weakly correlated with prices of other financial assets. To further the question of evaluation, we propose in this paper to, firstly, shed light on the difficulties of implementing the three major pricing approaches suggested in the literature for the weather derivatives (actuarial, arbitrage-free and consumption-based methods) and, secondly, to compute the prices of a weather contract by the three methodologies for comparison.
