Here are some highlights:

- The “rational models” constructed in economics and finance are increasingly disconnected from real-world behavior, as has been shown by research in behavioral finance. As Fabozzi and Sergio Focardi argued in 2012, “economics in its current form does not describe empirical reality but an idealized rational economic world.”
- The problem with relying on “rational models” as a bedrock theory is that new findings inconsistent with these models are often dismissed. Fabozzi recalls that a paper of his, co-authored with the then-chairman of Merrill Lynch, was rejected by a journal, with a reviewer saying that the ideas in the manuscript made no sense because they are “inconsistent with the capital asset pricing model (CAPM).”
- Progress in economics and finance is held back by an over-reliance on calculus-based theories. As Fabozzi and Marcos Lopez de Prado explained in an article in the Journal of Portfolio Management, calculus models were adopted into economics due to what some have called “physics envy,” and have not been very effective in describing real-world phenomena, particularly in today’s complicated, dynamic environment.
- In general, researchers either rely on non-empirical calculus-based theories or on “paleo-statistical tools” that were mostly designed in the pre-computer era. As a result, econometrics has “lost the train of innovation.”
- There is no place for major crises in the “pseudo-rational” world of current theory. The assumption that market forces will quickly correct parameters that deviate from their proper value has proved to be inadequate.
- Economics needs to be rebuilt as an empirical science. In particular, it needs to move more aggressively to adopt machine learning models, which can be quite successful if implemented intelligently.
- University economics and finance curricula are divided between programs that focus on sophisticated mathematics (e.g., stochastic calculus) and those that do not. But both positions are untenable. Modern real-world quants rely on both sophisticated mathematics and rigorous data science, and both should be taught. However, as MIT mathematician Gilbert Strang has emphasized, it is important to present the “mathematics that is most useful to the most students.”

Readers may also be interested in a related commentary by Marcos Lopez de Prado, entitled “How universities are failing finance students.”