Thank you so much, Dario. This is very thought-provoking.
An obvious question perhaps: what prevents you from exploiting this signal? It looks like there is a type of ETF — a large illiquid one — where something you can specify a priori as “good performance” is followed systematically by poor performance. Why not short such funds? Or is the classification of “good performance” only possible a posteriori?
What do the features you’ve found suggest about broader market phenomena? Are all stocks self-inflated to an extent? Does that mean they may one day self-deflate, e.g. when demographic changes lead to net outflows?
You have indeed raised two very insightful and important aspects of my results. I will reply briefly here and would be happy to discuss them in more detail during the panel discussion.
Regarding your question, “Why not short such funds?”: while, from an algorithmic point of view, it seems possible to identify a direction in which prices appear “inflated” (with a certain level of confidence), it is much more difficult to determine when a bubble might burst. Therefore, the risk of holding a short position that becomes extremely costly is real and significant.
Moreover, algorithms generally prefer to follow profitable trends rather than bet against them, so it is quite difficult to design strategies that systematically take the opposite side.
Regarding your question, “Are all stocks self-inflated to some extent?”, I believe the answer is likely yes. However, this effect is very difficult to measure, and even harder to disentangle from what reflects genuine fundamental value. In the ETF analysis I presented, I focused only on rebalancing trades, rather than active trades, which helps to address the causality issue to some extent, distinguishing between price discovery and price impact.
Thank you so much, Dario. This is very thought-provoking.
An obvious question perhaps: what prevents you from exploiting this signal? It looks like there is a type of ETF — a large illiquid one — where something you can specify a priori as “good performance” is followed systematically by poor performance. Why not short such funds? Or is the classification of “good performance” only possible a posteriori?
What do the features you’ve found suggest about broader market phenomena? Are all stocks self-inflated to an extent? Does that mean they may one day self-deflate, e.g. when demographic changes lead to net outflows?
Hello Ole,
You have indeed raised two very insightful and important aspects of my results. I will reply briefly here and would be happy to discuss them in more detail during the panel discussion.
Regarding your question, “Why not short such funds?”: while, from an algorithmic point of view, it seems possible to identify a direction in which prices appear “inflated” (with a certain level of confidence), it is much more difficult to determine when a bubble might burst. Therefore, the risk of holding a short position that becomes extremely costly is real and significant.
Moreover, algorithms generally prefer to follow profitable trends rather than bet against them, so it is quite difficult to design strategies that systematically take the opposite side.
Regarding your question, “Are all stocks self-inflated to some extent?”, I believe the answer is likely yes. However, this effect is very difficult to measure, and even harder to disentangle from what reflects genuine fundamental value. In the ETF analysis I presented, I focused only on rebalancing trades, rather than active trades, which helps to address the causality issue to some extent, distinguishing between price discovery and price impact.