Data source is yahoo finance, tools are python and matplotlib
[Constituent Weights] (as of last monthly rebalance)
Sanmina Corporation: 6.67%
Aflac Incorporated: 6.67%
The Walt Disney Company: 6.67%
Capital One Financial Corporation: 6.67%
Westlake Corporation: 6.67%
Freeport-McMoRan Inc.: 6.67%
Blackstone Inc.: 6.67%
Enterprise Products Partners L.P.: 6.67%
Air Products and Chemicals, Inc.: 6.67%
Watsco, Inc.: 6.67%
Skechers U.S.A., Inc.: 6.67%
Penske Automotive Group, Inc.: 6.67%
Berkshire Hathaway Inc.: 6.67%
BlackRock, Inc.: 6.67%
LVMH Moët Hennessy Louis Vuitton SE: 6.67%
The effect seems most pronounced at exactly the start of the chart. Are we sure we haven’t simply selected a date range that most produces this outcome?
I would theorize that you’re flipping the causation here. Being a CEO over 70 doesn’t mean you’ll outperform. Being a CEO over 70 who *doesn’t* outperform means you’ll get put out to pasture by the board.
correlation is not causation.
Do the returns not not index. If you started this plot in the middle of 2021 instead of 2020 it might be the S&P500 outperforming grey matter.
How much of this outperformance survives if you disregard Berkshire Hathaway?
That time frame is not statistically significant.
Hot take: if you plot a graph like this of several time series and you arbitrarily pick t=0 as the point at which they’re all equal, then you must also plot four other graphs where you pick different dates on which they’re equal.
That way you get more intellectual rigor in determining whether the stories the graphs tell are true, or just a visual artifact of your arbitrary starting point.
Instead of using absolute age as a proxy, use the CEO’s length of service directly.
“Hello, Correlation? Hi, my name is Causation, have you got a minute?”
To add to the other comments talking about causality, companies with an older CEO are (probably) more likely to be the founders themselves, making these companies generally older and more “established” in recovering against market downturns.
The graph starts after the 2020 Covid market crash, so established companies with more experience and stability were always expected to recover better than current S&P 500s.
The Boeing CEO is 68 so this is basically cheating.
Hmm, looks like they outperformed in the first half of 2021 and then performed pretty representative of S&P500 from then.
The timeline is too short to come to a conclusion (it’s only 5 years). It would be great to see a chart over the last decades.
The markets have nothing been but irrational since COVID.
Boomers are winning, we are losing.
How old they were when they became CEOs? maybe the successful ones managed to stay
Is this the Warren Buffet ETF?
Where can we find the information about “Grey Matter ETF”? I would like to check which companies are included.
Nuh-uh: correlation is not causation.
For those asking this is across a longer period of time (as far back as I can get the data) – note that we’re normalising the “grey matter” index and S&P 500 to 100, and the index level shows total return growth over time, meaning it accounts for both price changes and dividends.
1. The age of 70 is arbitrary. There is no significant reason to use that age range.
2. The chosen time frame is arbitrary and heavily influenced by macroeconomic factors such as COVID and the war in Ukraine.
3. It’s unclear whether this completely accounts for *every* company with a leader over 70. It’s also unclear whether it includes leaders who turned 70, turned 80, died, or were appointed during this time frame.
4. The comparison is lacking. Yes, the selected group outperforms the S&P, but do they outperform other age ranges? Is this consistent over other time periods? What about other stock exchanges?
5. As mentioned elsewhere, this is rife with unmitigated bias: selection, survivorship, and more.
6. This could be a poster child for correlation is not causation. Even if the aforementioned problems were accounted for, there are far too many factors at play for this to mean anything.
21 comments
Data source is yahoo finance, tools are python and matplotlib
[Constituent Weights] (as of last monthly rebalance)
Sanmina Corporation: 6.67%
Aflac Incorporated: 6.67%
The Walt Disney Company: 6.67%
Capital One Financial Corporation: 6.67%
Westlake Corporation: 6.67%
Freeport-McMoRan Inc.: 6.67%
Blackstone Inc.: 6.67%
Enterprise Products Partners L.P.: 6.67%
Air Products and Chemicals, Inc.: 6.67%
Watsco, Inc.: 6.67%
Skechers U.S.A., Inc.: 6.67%
Penske Automotive Group, Inc.: 6.67%
Berkshire Hathaway Inc.: 6.67%
BlackRock, Inc.: 6.67%
LVMH Moët Hennessy Louis Vuitton SE: 6.67%
The effect seems most pronounced at exactly the start of the chart. Are we sure we haven’t simply selected a date range that most produces this outcome?
I would theorize that you’re flipping the causation here. Being a CEO over 70 doesn’t mean you’ll outperform. Being a CEO over 70 who *doesn’t* outperform means you’ll get put out to pasture by the board.
correlation is not causation.
Do the returns not not index. If you started this plot in the middle of 2021 instead of 2020 it might be the S&P500 outperforming grey matter.
How much of this outperformance survives if you disregard Berkshire Hathaway?
That time frame is not statistically significant.
Hot take: if you plot a graph like this of several time series and you arbitrarily pick t=0 as the point at which they’re all equal, then you must also plot four other graphs where you pick different dates on which they’re equal.
That way you get more intellectual rigor in determining whether the stories the graphs tell are true, or just a visual artifact of your arbitrary starting point.
Instead of using absolute age as a proxy, use the CEO’s length of service directly.
“Hello, Correlation? Hi, my name is Causation, have you got a minute?”
To add to the other comments talking about causality, companies with an older CEO are (probably) more likely to be the founders themselves, making these companies generally older and more “established” in recovering against market downturns.
The graph starts after the 2020 Covid market crash, so established companies with more experience and stability were always expected to recover better than current S&P 500s.
The Boeing CEO is 68 so this is basically cheating.
Hmm, looks like they outperformed in the first half of 2021 and then performed pretty representative of S&P500 from then.
The timeline is too short to come to a conclusion (it’s only 5 years). It would be great to see a chart over the last decades.
The markets have nothing been but irrational since COVID.
Boomers are winning, we are losing.
How old they were when they became CEOs? maybe the successful ones managed to stay
Is this the Warren Buffet ETF?
Where can we find the information about “Grey Matter ETF”? I would like to check which companies are included.
Nuh-uh: correlation is not causation.
For those asking this is across a longer period of time (as far back as I can get the data) – note that we’re normalising the “grey matter” index and S&P 500 to 100, and the index level shows total return growth over time, meaning it accounts for both price changes and dividends.
https://preview.redd.it/rh0drthl65kf1.png?width=1248&format=png&auto=webp&s=c45f793e3aff23761352b80b4b5c13924383270a
This is bad data science.
1. The age of 70 is arbitrary. There is no significant reason to use that age range.
2. The chosen time frame is arbitrary and heavily influenced by macroeconomic factors such as COVID and the war in Ukraine.
3. It’s unclear whether this completely accounts for *every* company with a leader over 70. It’s also unclear whether it includes leaders who turned 70, turned 80, died, or were appointed during this time frame.
4. The comparison is lacking. Yes, the selected group outperforms the S&P, but do they outperform other age ranges? Is this consistent over other time periods? What about other stock exchanges?
5. As mentioned elsewhere, this is rife with unmitigated bias: selection, survivorship, and more.
6. This could be a poster child for correlation is not causation. Even if the aforementioned problems were accounted for, there are far too many factors at play for this to mean anything.
Comments are closed.