Model Risk and Model Choice in the Case of Barrier Options

We analyze model risk for the pricing of barrier options. In contrast to existing literature, this paper is based on an empirical data set of over 40,000 bonus certificates to analyze the real market extent of model risk for traded barrier options instead of purely synthetic options. For this purpose a local volatility model, the Heston model and the Bates model are applied. Furthermore, we add to the literature on the behavior of issuers of retail derivatives in terms of model choice. We find evidence that the majority of the issuers prefer stochastic volatility over local volatility models, while they do not use the even more realistic Bates model which incorporates jumps in the underlying