Data and computers on steroids have partnered to transform finance and reengineer its future. Past conventions have defined the role of data to be a complement to financial theories, providing a testing ground and an estimator of future prices, whether of assets, stocks, or derivatives.
Theories of finance (such as the Arrow-Debreu framework, Martingale pricing, risk neutral pricing, etc.), while mathematically and theoretically stimulating, also embed a variety of risks and real financial misconceptions. For example, risk is defined by predictable (future states) events while in real finance, uncertainty primes; prices exist only in the present and so on.
Further, while conventional finance is an ex-ante approach to the future, data science is an inverse approach that seeks ex-post to estimate causes or models that explain the data so collected and improve their state of knowledge and know-how by learning through a feedback process. This approach is often structured by terms such as ‘deep learning’, ‘machine learning’, and ‘artificial intelligence’.
Thus, one approach is defined by hypothetical theories, while the other is an analytic data and inverse approach that implies hypothetical models for the purpose of learning and/or deciding.
The purpose of this paper is to elaborate on the fundamental elements that are contributing to the transformation of finance and raise its risk consequences.