Another Financial Crisis Eminent – Big Data and Machine Learning Won’t Help
Even after being hit hard by a financial crisis a decade ago, the financial markets have managed to stay in the loop. To be precise, the stock markets have consistently achieved new peaks while volatility levels are quite low than before.
Technological advancements and innovations in the financial industry have led to key players adopting computational algorithms and big data. These two have been enabled to a greater extent by machine learning. This however, has led to many questions regarding potential financial risks.
The risk faced by key players in the financial markets today is the failure to understand the extent to which technology advancement is or is not reducing financial uncertainty. In the concept of irreducible uncertainty; John Maynard Keynes was able to differentiate between incidents one can reasonably calculate probabilities for and those which remain fundamentally unknown. This, however is not the case with players in finance industry as they have not been able to determine this fine line.
Therefore, the main concerns are the numerous alarming parallels that exist between the contemporary development in machine learning and algorithmic trading and the massive growth of financial engineering before the crisis. Secondly, big data and an advanced computing power is not a guarantee that these players will shake off financial risks.
In the 20th and early 21st century, so much was done by the most intelligent minds to subdue financial uncertainty. Businesses were built by mathematical tools generated by derivative desks, hence increased profits and high investment returns. It was a time of booming entrepreneurship with advanced probabilistic modeling and unrivalled computational power. The New Yorkers so thought that they were on the right track.
We may be at risk of repeating the past: “quant delusion” where many modeling speculations like correlations between asset prices were found to be extremely faulty while foundational basis of quantitative finance collapsed. The quants had recklessly generated a set of possible outcomes and calculated conditional probabilities of incidents; subjective to how they had known the world. These decisions were revealed to be illogical once impossible incidents occurred.
With all these contemporary advancements in the financial markets, we are able to explore enormous possible outcomes. However, let’s not forget to ask ourselves to what degree our calculated conditional probabilities differ from actual probabilities?