Expense Ratios Correlation with Returns

Have a question about your personal investments? No matter how simple or complex, you can ask it here.
Post Reply
Njm8845
Posts: 69
Joined: Fri Aug 05, 2016 1:28 am

Expense Ratios Correlation with Returns

Post by Njm8845 » Mon Jul 09, 2018 3:48 pm

Some years ago, the SEC produced an official notice that expense ratios hurt returns. Link below.

The gist is that expense ratios are a better predictor than past returns of future returns.

Does anyone know how this is calculated exactly? I believe that a significant (negative) correlation is present across the entire universe of funds. However, is that correlation still present when you shrink the sample to include only funds where the expense ratio is less than, say, .5%?

It seems like dogma on here to ask for expense ratios, and then rule out the higher ones, citing the logic in the link below. However, I don't know if that logic holds for expense ratios that are relatively small to begin with. I doubt there is much of a predictive value in comparing expense ratios of .2% and .1%.

https://www.sec.gov/reportspubs/investo ... nthtm.html

bloom2708
Posts: 4648
Joined: Wed Apr 02, 2014 2:08 pm
Location: Fargo, ND

Re: Expense Ratios Correlation with Returns

Post by bloom2708 » Mon Jul 09, 2018 3:57 pm

If your fund returns 4%, then you can calculate how much (as a percentage) is lost to the expense ratio.

4% return, .5% ER = 12.5% of your earning out the door. The bad part is you paid 12.5% of your earnings. The worse part is the dollars left your account and now can't grow with the rest of your dollars.

If at Vanguard, there is almost no reason to have an index fund over .2%. Many are less than .1%.

What you can't control (except by changing jobs) is your 401k/403b expense ratios. There you have to work with what you have. Many times the pre-tax savings, match, profit sharing are still good incentives with higher expense ratio funds. The benefits outweigh the higher expense ratio.

The industry average is ~1.05%. It adds up. Really adds up.

Run some scenarios with this calculator. Use .1% 1% and 2% and see the differences with your numbers over 30 years.

http://www.dinkytown.net/java/CompareFees.html

Don't let anyone tell you .5% doesn't matter. Or 1% is just a little bit.
Where to spend your time: | 1. You completely control <--spend your time here! | 2. You partially control <--spend very little time here! | 3. You have no control <--spend no time here

pkcrafter
Posts: 13013
Joined: Sun Mar 04, 2007 12:19 pm
Location: CA
Contact:

Re: Expense Ratios Correlation with Returns

Post by pkcrafter » Mon Jul 09, 2018 4:21 pm

Njm8845 wrote:
Mon Jul 09, 2018 3:48 pm
Some years ago, the SEC produced an official notice that expense ratios hurt returns. Link below.

The gist is that expense ratios are a better predictor than past returns of future returns.

Does anyone know how this is calculated exactly? I believe that a significant (negative) correlation is present across the entire universe of funds. However, is that correlation still present when you shrink the sample to include only funds where the expense ratio is less than, say, .5%?

It seems like dogma on here to ask for expense ratios, and then rule out the higher ones, citing the logic in the link below. However, I don't know if that logic holds for expense ratios that are relatively small to begin with. I doubt there is much of a predictive value in comparing expense ratios of .2% and .1%.

https://www.sec.gov/reportspubs/investo ... nthtm.html
The low cost logic is weaker with small portfolios. As portfolio size increases, ERs (expense rations) and other fees become a larger drag. Also, length of time is an important factor.

Fees compound just like returns.

Paul
When times are good, investors tend to forget about risk and focus on opportunity. When times are bad, investors tend to forget about opportunity and focus on risk.

Njm8845
Posts: 69
Joined: Fri Aug 05, 2016 1:28 am

Re: Expense Ratios Correlation with Returns

Post by Njm8845 » Tue Jul 10, 2018 6:16 am

bloom2708 wrote:
Mon Jul 09, 2018 3:57 pm
If your fund returns 4%, then you can calculate how much (as a percentage) is lost to the expense ratio.

4% return, .5% ER = 12.5% of your earning out the door. The bad part is you paid 12.5% of your earnings. The worse part is the dollars left your account and now can't grow with the rest of your dollars.

If at Vanguard, there is almost no reason to have an index fund over .2%. Many are less than .1%.

What you can't control (except by changing jobs) is your 401k/403b expense ratios. There you have to work with what you have. Many times the pre-tax savings, match, profit sharing are still good incentives with higher expense ratio funds. The benefits outweigh the higher expense ratio.

The industry average is ~1.05%. It adds up. Really adds up.

Run some scenarios with this calculator. Use .1% 1% and 2% and see the differences with your numbers over 30 years.

http://www.dinkytown.net/java/CompareFees.html

Don't let anyone tell you .5% doesn't matter. Or 1% is just a little bit.
I guess I didn't state my point well enough.

I think 1% ER is a lot. I don't own any funds that are even half of that. All else being equal, I'd go with the lower cost fund. However, things are rarely equal. See below theoretical. Listed is 10 year rolling returns and the expense ratio.

Fund A: 8% RR, .05% ER
Fund B: 9% RR, .15% ER
Fund C: 9% RR, .90% ER

All are in the same asset class. Which fund would you pick?

Conventional wisdom is that you should take the low cost one. Why? Because ER has an inverse relationship with future returns. As expense ratios go up, the expected future return goes down. I believe this to be true over the entire universe of funds.

However, I'd like to know if this correlation still holds true over a small range of expense ratios. In other words, I doubt that the difference between a .2% ER and .1% ER has much predictive value. Of course if they return the same, go with the lower cost fund. But what if they don't return the same?

User avatar
JPH
Posts: 785
Joined: Mon Jun 27, 2011 8:56 pm

Re: Expense Ratios Correlation with Returns

Post by JPH » Tue Jul 10, 2018 6:31 am

If you restrict the range of either variable, then the correlation will decrease.
While the moments do summersaults into eternity | Cling to their coattails and beg them to stay - Townes Van Zandt

User avatar
vineviz
Posts: 1275
Joined: Tue May 15, 2018 1:55 pm

Re: Expense Ratios Correlation with Returns

Post by vineviz » Tue Jul 10, 2018 8:50 am

Njm8845 wrote:
Mon Jul 09, 2018 3:48 pm
Some years ago, the SEC produced an official notice that expense ratios hurt returns. Link below.

The gist is that expense ratios are a better predictor than past returns of future returns.

Does anyone know how this is calculated exactly? I believe that a significant (negative) correlation is present across the entire universe of funds. However, is that correlation still present when you shrink the sample to include only funds where the expense ratio is less than, say, .5%?

It seems like dogma on here to ask for expense ratios, and then rule out the higher ones, citing the logic in the link below. However, I don't know if that logic holds for expense ratios that are relatively small to begin with. I doubt there is much of a predictive value in comparing expense ratios of .2% and .1%.
Obviously, the smaller the spread in expense ratios the more that other factors come to dominate the explanation of returns. Controlling for those other factors is challenging, and I don't know an any academic study that has looked at it with the level of detail you're suggesting (though I think it is possibly worthwhile).

I performed a quick analysis of U.S. large cap ETFs, limiting the universe to funds that already had low ER (less than 0.35%) and that broadly tracked the market (i.e. excluded leveraged funds, sector funds, etc). I included turnover and a few basic portfolio characteristics (market beta, average holding size, value/growth tilt) but I didn't apply all the rigor a formal study would apply. Nonetheless, it seems to me that - holding other things else equal - for every 5 bps change in ER you can expect a 1.2 bps change in annualized return. That's pretty tiny and, given how variable stock returns are, almost surely not statistically significant.

Small changes in turnover have a much bigger impact on returns: the difference between 10% turnover and 15% turnover causes a 5 bps change in annualized return.

Other differences I didn't look at but that I would expect to have a bigger impact on returns than a small change in ER include investment efficiency (e.g. average amount of uninvested cash), AUM, and capital gains management procedures (for taxable accounts).

I've definitely been guilty of over-focusing on ER in the past, but the more I think about it the more confident I am that once the ER gets below 10 or 15 bps that it probably isn't very predictive of realized return. In this expense neighborhood, other factors (many of which are unknowable) will be more important.
"Far more money has been lost by investors preparing for corrections than has been lost in corrections themselves." ~~ Peter Lynch

Njm8845
Posts: 69
Joined: Fri Aug 05, 2016 1:28 am

Re: Expense Ratios Correlation with Returns

Post by Njm8845 » Wed Jul 11, 2018 6:40 am

vineviz wrote:
Tue Jul 10, 2018 8:50 am
Njm8845 wrote:
Mon Jul 09, 2018 3:48 pm
Some years ago, the SEC produced an official notice that expense ratios hurt returns. Link below.

The gist is that expense ratios are a better predictor than past returns of future returns.

Does anyone know how this is calculated exactly? I believe that a significant (negative) correlation is present across the entire universe of funds. However, is that correlation still present when you shrink the sample to include only funds where the expense ratio is less than, say, .5%?

It seems like dogma on here to ask for expense ratios, and then rule out the higher ones, citing the logic in the link below. However, I don't know if that logic holds for expense ratios that are relatively small to begin with. I doubt there is much of a predictive value in comparing expense ratios of .2% and .1%.
Obviously, the smaller the spread in expense ratios the more that other factors come to dominate the explanation of returns. Controlling for those other factors is challenging, and I don't know an any academic study that has looked at it with the level of detail you're suggesting (though I think it is possibly worthwhile).

I performed a quick analysis of U.S. large cap ETFs, limiting the universe to funds that already had low ER (less than 0.35%) and that broadly tracked the market (i.e. excluded leveraged funds, sector funds, etc). I included turnover and a few basic portfolio characteristics (market beta, average holding size, value/growth tilt) but I didn't apply all the rigor a formal study would apply. Nonetheless, it seems to me that - holding other things else equal - for every 5 bps change in ER you can expect a 1.2 bps change in annualized return. That's pretty tiny and, given how variable stock returns are, almost surely not statistically significant.

Small changes in turnover have a much bigger impact on returns: the difference between 10% turnover and 15% turnover causes a 5 bps change in annualized return.

Other differences I didn't look at but that I would expect to have a bigger impact on returns than a small change in ER include investment efficiency (e.g. average amount of uninvested cash), AUM, and capital gains management procedures (for taxable accounts).

I've definitely been guilty of over-focusing on ER in the past, but the more I think about it the more confident I am that once the ER gets below 10 or 15 bps that it probably isn't very predictive of realized return. In this expense neighborhood, other factors (many of which are unknowable) will be more important.
Good post. Your quick analysis agrees with my thought experiment. May I ask how exactly you did the analysis? Where'd you get the data from?

Post Reply