Read on Twitter
/1
Assume machine learning and algorithms taking over the finance industry with complete robotic rationality, following problems would still exist:
- Customers have emotions, demand explanations and follow narratives. If the AI model suggests buying a Russian Steel company and
Assume machine learning and algorithms taking over the finance industry with complete robotic rationality, following problems would still exist:
- Customers have emotions, demand explanations and follow narratives. If the AI model suggests buying a Russian Steel company and
/2
shorting a fast-growing US cloud company, the average person will not understand and trust the system
- Even with similar sector allocation, tracking error would scare customers as the model performance will deviate strongly from the benchmark.
- Scalability is limited for
shorting a fast-growing US cloud company, the average person will not understand and trust the system
- Even with similar sector allocation, tracking error would scare customers as the model performance will deviate strongly from the benchmark.
- Scalability is limited for
/3
"rational" security selection, as it demands the deviation from Market cap weighting (increasing trading and market impact costs) and higher concentration in the case of factor models. Most providers are more interested in AUM than Alpha
- An algorithm still faces career risk
"rational" security selection, as it demands the deviation from Market cap weighting (increasing trading and market impact costs) and higher concentration in the case of factor models. Most providers are more interested in AUM than Alpha
- An algorithm still faces career risk
/4
as it will compete with benchmarks and other algos over highest possible AUM*Benefit/ImplementationCost. The best long-term algo will still be shut down by management if it doesn& #39;t attract assets or is too expensive to operate
- The wealth management market will always be
as it will compete with benchmarks and other algos over highest possible AUM*Benefit/ImplementationCost. The best long-term algo will still be shut down by management if it doesn& #39;t attract assets or is too expensive to operate
- The wealth management market will always be
/5
distorted to a certain degree due to marketing, media, news etc. E.g. Smart Beta sells the promise of Value outperformance by "tilting towards the factor" with minimal implementation cost and >0.9 correlation to the benchmark. If customers already buy the narrative with a
distorted to a certain degree due to marketing, media, news etc. E.g. Smart Beta sells the promise of Value outperformance by "tilting towards the factor" with minimal implementation cost and >0.9 correlation to the benchmark. If customers already buy the narrative with a
/6
halfhearted process why should I bother improving the procedure? Also frauds, hypes and entertainment will always result in investment topic macrotrends (bubbles) which attract a lot of AUM despite little prospect of alpha. Gambling and incentives are always part of the game
halfhearted process why should I bother improving the procedure? Also frauds, hypes and entertainment will always result in investment topic macrotrends (bubbles) which attract a lot of AUM despite little prospect of alpha. Gambling and incentives are always part of the game
/7
Aware retail investors and smaller investment shops can exploit the barriers of BIG MONEY. Using this edge is a privilege. Can it fail? Of course! We are still dealing with a highly complex system controlled by chaos and randomness. However, subtracting the inevitable chains
Aware retail investors and smaller investment shops can exploit the barriers of BIG MONEY. Using this edge is a privilege. Can it fail? Of course! We are still dealing with a highly complex system controlled by chaos and randomness. However, subtracting the inevitable chains
/8
of big funds, f.e. by global exposure, equal weighting etc. are first steps to make a difference. Random equal-weighted portfolios outperform managed market-cap weighted portfolios on average. So be aware why they are still dealing with it.
of big funds, f.e. by global exposure, equal weighting etc. are first steps to make a difference. Random equal-weighted portfolios outperform managed market-cap weighted portfolios on average. So be aware why they are still dealing with it.
/9
So should the average investor sell all his funds and go full-retard? Hell, no! But there are steps to go down the rabbit hole until the personal frontier of comfort/risk appetite and willpower is reached:
Level1: Globally diversified MC-weighted asset allocation with
So should the average investor sell all his funds and go full-retard? Hell, no! But there are steps to go down the rabbit hole until the personal frontier of comfort/risk appetite and willpower is reached:
Level1: Globally diversified MC-weighted asset allocation with
/10
cost-effective ETF& #39;s (Play the game but don& #39;t try to beat it and play it at low cost)
Level 2: Differ by using equal-weighting and factor tilts available at low costs e.g. PE10 based allocation to US, Europe, Emerging Markets (Embrace randomness, exploit human behavior)
cost-effective ETF& #39;s (Play the game but don& #39;t try to beat it and play it at low cost)
Level 2: Differ by using equal-weighting and factor tilts available at low costs e.g. PE10 based allocation to US, Europe, Emerging Markets (Embrace randomness, exploit human behavior)
/11
Level 3: Systematic, equal-weighted, evidence-based security selection with intermediate-term rebalancing (Embrace higher concentration in e.g. 20-50 names, high tacking error, fishing at the extremes of factor spreads, long-term mean reversion with short-term pain)
Level 3: Systematic, equal-weighted, evidence-based security selection with intermediate-term rebalancing (Embrace higher concentration in e.g. 20-50 names, high tacking error, fishing at the extremes of factor spreads, long-term mean reversion with short-term pain)
/12
Level 4: Discretionary (Buffet-style) investing with deep-dive research and high-conviction investments / Short-term Trading / Shorting (not advised, hardest games in town, for true maniacs only)
I am at Level 3 and plan to stay there. (End of Thread I guess...)
Level 4: Discretionary (Buffet-style) investing with deep-dive research and high-conviction investments / Short-term Trading / Shorting (not advised, hardest games in town, for true maniacs only)
I am at Level 3 and plan to stay there. (End of Thread I guess...)
