Refine
Year of publication
- 2019 (2) (remove)
Document Type
- Journal article (2)
Language
- English (2) (remove)
Has full text
- yes (2) (remove)
Is part of the Bibliography
- yes (2)
Institute
Publisher
- MDPI (2) (remove)
This paper studies option pricing based on a reverse engineering (RE) approach. We utilize artificial intelligence in order to numerically compute the prices of options. The data consist of more than 5000 call- and put-options from the German stock market. First, we find that option pricing under reverse engineering obtains a smaller root mean square error to market prices. Second, we show that the reverse engineering model is reliant on training data. In general, the novel idea of reverse engineering is a rewarding direction for future research. It circumvents the limitations of finance theory, among others strong assumptions and numerical approximations under the Black–Scholes model.
This paper develops a new methodology in order to study the role of dynamic expectations. Neither reference-point theories nor feedback models are sufficient to describe human expectations in a dynamic market environment. We use an interdisciplinary approach and demonstrate that expectations of non-learning agents are time-invariant and isotropic. On the contrary, learning enhances expectations. We uncover the “yardstick of expectations” in order to assess the impact of market developments on expectations. For the first time in the literature, we reveal new insights about the motion of dynamic expectations. Finally, the model is suitable for an AI approach and has major implications on the behaviour of market participants.