Human vs. Machine: Disposition Effect among Algorithmic and Human Day Traders

In finance, one of the puzzles is presence of disposition effect puzzle. We often see how investors book profits relatively quickly but sit on losses. This is a puzzle as one would expect investors to wait for higher profits but limit losses quickly.

Karolis Liaudinskas of Norway Central Bank studies whether algorithmic trading also suffers from disposition effect:

This paper studies whether and why algorithmic traders exhibit one of the most broadly documented behavioral puzzles – the disposition effect. We use trade data from the NASDAQ Copenhagen Stock Exchange merged with the weather data.

We find that on average, the disposition effect for human traders is substantial and increases significantly on colder days, while for similarly-trading algorithms, it is insignificant and insensitive to the weather.

This provides causal evidence of the link between human psychology and the disposition effect and suggests that algorithms can reduce psychology-related human errors. Considering the ongoing AI adoption, this may have broad implications.

Implications?

Given the ongoing ubiquitous adoption of AI, this may have broad implications for economic theory, financial markets, the real economy,
and, potentially, the future of human behavior.

For economic theory, our results suggest that decisions automated by algorithms are more consistent with rational economic models than on-thespot decisions made by humans.

Hence, as humans are replaced by AI, rational economic models, e.g., those based on Bayesian updating of beliefs, the Expected Utility theory (von Neumann and Morgenstern, 1947) or Subjective Expected Utility (Savage, 1954), might become more accurate in explaining the real world. Similarly, as human traders are replaced by algorithmic traders (e.g., Kirilenko and Lo, 2013), financial markets might become easier to explain with rational models.

For the real economy, industries that require more “rational” decision-making might replace humans with algorithms faster, affecting unemployment, productivity and economic growth. Finally, people surrounded by automated decision-making (e.g. self-driving cars, virtual assistants, etc.) that is more “rational”, may either learn to behave more “rationally” or their “rationality” may atrophy due to their reliance on machines.

 

Advertisement

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.


%d bloggers like this: