Author Topic: Estimating future returns from stocks (and bonds)  (Read 3867 times)

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 6633
Re: Estimating future returns from stocks (and bonds)
« Reply #1 on: June 29, 2018, 10:18:01 AM »
Thanks for linking to that article, it's interesting.  My bias is that I've read many of Larry Swedroe's books, but I still try to think about what he says for myself.

The estimates shown in that article use a range, which is appropriate.  Stock market returns have been in the predicted range of 7% +/- 5%.  But that might not be useful: I would definitely want stock returns at 12% / year, while I wouldn't like stock returns of 2% / year.
 Even if you pretend this range is a crystal ball (and there is no crystal ball for stock returns, as the article points out), I'm not sure I can make use of that range.

It may be interesting to see Vanguard's white paper on forecasting stock returns.  They wind up preferring the same metric Larry Swedroe mentions (CAPE 10), but the correlation to future returns is surprisingly low (0.4 or so as I recall).
https://personal.vanguard.com/pdf/s338.pdf
« Last Edit: June 29, 2018, 10:20:56 AM by MustacheAndaHalf »

VoteCthulu

  • Bristles
  • ***
  • Posts: 409
Re: Estimating future returns from stocks (and bonds)
« Reply #2 on: June 29, 2018, 10:27:57 AM »
Good info for people deciding on a portfolio allocation with a 10 year time horizon. Somewhat less useful for those of us investing for the next 50 years.

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: Estimating future returns from stocks (and bonds)
« Reply #3 on: June 29, 2018, 02:36:43 PM »
Good info for people deciding on a portfolio allocation with a 10 year time horizon. Somewhat less useful for those of us investing for the next 50 years.

I don't know. Seems pretty relevant to me. Also pretty actionable. E.g., if you expect lower returns over next ten years, that should suggest bigger savings amount.

Also, an extended period of lower returns suggests either a lower withdrawal rate or a variable withdrawal rate. Boglehead Siamond did a really good blog post on this a while back:

https://finpage.blog/2018/02/28/cape-and-safe-withdrawal-rates/

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: Estimating future returns from stocks (and bonds)
« Reply #4 on: June 29, 2018, 02:44:21 PM »
Thanks for linking to that article, it's interesting.  My bias is that I've read many of Larry Swedroe's books, but I still try to think about what he says for myself.

The estimates shown in that article use a range, which is appropriate.  Stock market returns have been in the predicted range of 7% +/- 5%.  But that might not be useful: I would definitely want stock returns at 12% / year, while I wouldn't like stock returns of 2% / year.
 Even if you pretend this range is a crystal ball (and there is no crystal ball for stock returns, as the article points out), I'm not sure I can make use of that range.

It may be interesting to see Vanguard's white paper on forecasting stock returns.  They wind up preferring the same metric Larry Swedroe mentions (CAPE 10), but the correlation to future returns is surprisingly low (0.4 or so as I recall).
https://personal.vanguard.com/pdf/s338.pdf

The Siamond blog post referenced in my message just above this talks about a .4 correlation too. Perhaps the same number or maybe calculated from the same data. It would make sense...

VoteCthulu

  • Bristles
  • ***
  • Posts: 409
Re: Estimating future returns from stocks (and bonds)
« Reply #5 on: June 29, 2018, 04:05:27 PM »
I don't know. Seems pretty relevant to me.
So someone investing for 50 years should follow the article's advice and reallocate equities to bonds multiple times a year? That might work for a shorter time horizon, but seems questionable to me for the long term.

I do agree that valuations should be considered, but other articles (like those from Mad Fiendist and earlyretirementnow) seem much more actionable for long term investors.

Tyson

  • Magnum Stache
  • ******
  • Posts: 3025
  • Age: 52
  • Location: Denver, Colorado
Re: Estimating future returns from stocks (and bonds)
« Reply #6 on: June 30, 2018, 10:49:45 AM »
This is a good article from Larry Swedroe:

http://www.etf.com/sections/index-investor-corner/swedroe-toss-your-expectations?nopaging=1

It's all just speculation and a rather desperate attempt to "beat the market".  These people remind me of "sophisticated" gamblers in Vegas.  They all have their schemes and stats and probabilities.  But the truth is, the house always wins. 

My advice to MMM people - keep it simple.  Stick with plain old boring S&P500 index funds with whatever ratio of bonds lets you sleep at night. 

Remember, when you buy Index funds, you don't have to beat the house.  You ARE the house.

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: Estimating future returns from stocks (and bonds)
« Reply #7 on: June 30, 2018, 07:30:36 PM »
This is a good article from Larry Swedroe:

http://www.etf.com/sections/index-investor-corner/swedroe-toss-your-expectations?nopaging=1

It's all just speculation and a rather desperate attempt to "beat the market".  These people remind me of "sophisticated" gamblers in Vegas.  They all have their schemes and stats and probabilities.  But the truth is, the house always wins. 

My advice to MMM people - keep it simple.  Stick with plain old boring S&P500 index funds with whatever ratio of bonds lets you sleep at night. 

Remember, when you buy Index funds, you don't have to beat the house.  You ARE the house.

I don't think Larry is actively managing or trying to beat the house. I read him to be saying that equity returns and bond returns will probably be lower and we need to plan accordingly:

http://www.etf.com/sections/index-investor-corner/swedroe-4-horsemen-your-portfolio?nopaging=1



SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: Estimating future returns from stocks (and bonds)
« Reply #8 on: June 30, 2018, 07:37:47 PM »
I don't know. Seems pretty relevant to me.
So someone investing for 50 years should follow the article's advice and reallocate equities to bonds multiple times a year? That might work for a shorter time horizon, but seems questionable to me for the long term.

I do agree that valuations should be considered, but other articles (like those from Mad Fiendist and earlyretirementnow) seem much more actionable for long term investors.

I didn't read the article to say that...

I think Larry is really arguing to be conservative in your expectations. Also, Larry is usually (I think) applying modern portfolio theory to his thinking. As a result, he's looking at standard deviations and correlation coefficients and not just expected returns when he constructs a portfolio.

Regarding rebalancing and lowering returns, that's true using your example... but if in a case like that you're not rebalancing, you're also dialing up your risk if you don't rebalance.

Obviously, if you're rebalancing between two asset classes that generate similar returns, you boost your portfolio return by rebalancing as I showed here:

https://evergreensmallbusiness.com/100-stocks-allocation-suffers-two-big-flaws/

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 6633
Re: Estimating future returns from stocks (and bonds)
« Reply #9 on: July 01, 2018, 05:58:24 AM »
I think we have several posters all saying to keep passively holding the US market, but making different claims about what Larry Swedroe thinks.  Mr Swedroe often tells people "his crystal ball is hazy", meaning that he can't predict the future of the stock market, so stay the course.  Whatever people want to think about the article, though, if the result is you remain passively invested, I'd say we mostly agree.

P/E ratio can be used to divide stocks into value vs growth.  And Larry Swedroe's books often bring up the possibility of tilting to value.  So what happens if you switch from S&P 500 to a large-cap value index fund?  That effectively takes your investments into a lower P/E environment.  I wonder if there's studies on avoiding high P/E ratios using that approach.

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: Estimating future returns from stocks (and bonds)
« Reply #10 on: July 02, 2018, 03:06:44 PM »
I think we have several posters all saying to keep passively holding the US market, but making different claims about what Larry Swedroe thinks.  Mr Swedroe often tells people "his crystal ball is hazy", meaning that he can't predict the future of the stock market, so stay the course.  Whatever people want to think about the article, though, if the result is you remain passively invested, I'd say we mostly agree.

P/E ratio can be used to divide stocks into value vs growth.  And Larry Swedroe's books often bring up the possibility of tilting to value.  So what happens if you switch from S&P 500 to a large-cap value index fund?  That effectively takes your investments into a lower P/E environment.  I wonder if there's studies on avoiding high P/E ratios using that approach.

Good points, MustacheAndaHalf... good points. :-)

Telecaster

  • Magnum Stache
  • ******
  • Posts: 3551
  • Location: Seattle, WA
Re: Estimating future returns from stocks (and bonds)
« Reply #11 on: July 02, 2018, 03:25:34 PM »

I don't think Larry is actively managing or trying to beat the house. I read him to be saying that equity returns and bond returns will probably be lower and we need to plan accordingly:


^ Exactly.  One question we see here all the time around here is something like "should I invest in the current market with its high valuations?"  With the implication that the market will crash and money will be lost.

But that's the wrong way to look at it.  Higher than average valuations imply lower than average future returns (over some reasonable time period).   So for planning purposes don't assume you'll get the longer term average from today's date.  Assume you'll get something lower than average. 

PizzaSteve

  • Pencil Stache
  • ****
  • Posts: 501
Re: Estimating future returns from stocks (and bonds)
« Reply #12 on: July 02, 2018, 04:17:31 PM »
Larry is a good resource.  An earlier post prompts me to respond that we should not be hung up on nominal return rates (expecting a specific number) vs real return rates (expecting to earn a premium over the going risk free option, such as t-bills, inflation rate etc) .  Saying 2% is bad, but 11% is good means you need to hit the books a bit.  I would be fine with 2% real returns in a 0% inflation or deflationary market.  11% returns sucked during the peak of the oil/energy price driven inflation era of the 70s.

Keep in mind that these low expected returns reflect market assumptions about risk premiums, actual risk with stocks and bonds (lower now), inflation expectations, etc.  Since it is a closed system and capital is relatively cheap, we investors will not be seeing as many 'great deals.'  However I expect that the lower returns should come with a bit less volatile and pretty consistent returns.

Those seeking risk premiums will likely start to chase small cap, international and maybe REATs or commodity producer focused ETFs and stocks I expect.
« Last Edit: July 02, 2018, 04:22:48 PM by PizzaSteve »

ChpBstrd

  • Walrus Stache
  • *******
  • Posts: 6662
  • Location: A poor and backward Southern state known as minimum wage country
Re: Estimating future returns from stocks (and bonds)
« Reply #13 on: July 02, 2018, 04:23:52 PM »
In terms of what to do with this info, the author offers working longer, cutting spending, or moving to a lower-cost area as the changes to make.

Overall, this study makes a good case for pursuing boring single digit returns from preferred stocks, REITs, mortgages, or bonds. Bonus: if a crash occurs, you'll know it and may be able to pivot back to equities.

PizzaSteve

  • Pencil Stache
  • ****
  • Posts: 501
Re: Estimating future returns from stocks (and bonds)
« Reply #14 on: July 02, 2018, 04:32:04 PM »
In terms of what to do with this info, the author offers working longer, cutting spending, or moving to a lower-cost area as the changes to make.

Overall, this study makes a good case for pursuing boring single digit returns from preferred stocks, REITs, mortgages, or bonds. Bonus: if a crash occurs, you'll know it and may be able to pivot back to equities.
I dont think you are reading the article accurately.  He is saying that using current valuations in your models (such as a rolling Shiller CAPE 10) has proven to be a robust model to predict returns.  He isnt saying be more conservative with the portfolio, though I know he is someone who applies typical stock and bond mix allocations for his clients to reduce volatility.

Having higher expectations does not generate better returns. Stocks are volatile...thats just fact.  100% equities is higher return, but also higher volatility of returns.

Anyway, it is known that a bond portfolio has less risk but also lower expected returns. 

An analogy to his main point would be applying inflation to expense budgeting.  It is like saying you should use current inflation rates in your expense budget, adjusting every year, rather than a flat inflation rate of 5%, just because that is a historic average.

Assuming expected returns is a critical part of your model to predict future actual returns.  How conservative or aggressive you are in your assumptions wont change the actual outcome once you have your allocations, so I am not sure I understand your comment.  Boredom or excitement has nothing to do with it.  An exciting portfolio doesnt always do what a retirement plan needs.  Bogle says investing should be boring and I agree.  Buy, ignore, sell when you need $.

If you dont save enough or get sufficient returns to meet expenses, yes you have to make up the gap somehow.
« Last Edit: July 02, 2018, 04:39:10 PM by PizzaSteve »

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: Estimating future returns from stocks (and bonds)
« Reply #15 on: July 02, 2018, 04:53:59 PM »
In terms of what to do with this info, the author offers working longer, cutting spending, or moving to a lower-cost area as the changes to make.

Overall, this study makes a good case for pursuing boring single digit returns from preferred stocks, REITs, mortgages, or bonds. Bonus: if a crash occurs, you'll know it and may be able to pivot back to equities.
I dont think you are reading the article accurately.  He is saying that using current valuations in your models (such as a rolling Shiller CAPE 10) has proven to be a robust model to predict returns.  He isnt saying be more conservative with the portfolio, though I know he is someone who applies typical stock and bond mix allocations for his clients to reduce volatility.

Having higher expectations does not generate better returns. Stocks are volatile...thats just fact.  100% equities is higher return, but also higher volatility of returns.

Anyway, it is known that a bond portfolio has less risk but also lower expected returns. 

An analogy to his main point would be applying inflation to expense budgeting.  It is like saying you should use current inflation rates in your expense budget, adjusting every year, rather than a flat inflation rate of 5%, just because that is a historic average.

Assuming expected returns is a critical part of your model to predict future actual returns.  How conservative or aggressive you are in your assumptions wont change the actual outcome once you have your allocations, so I am not sure I understand your comment.  Boredom or excitement has nothing to do with it.  An exciting portfolio doesnt always do what a retirement plan needs.  Bogle says investing should be boring and I agree.  Buy, ignore, sell when you need $.

If you dont save enough or get sufficient returns to meet expenses, yes you have to make up the gap somehow.

+1 to this comment and to PizzaSteve's earlier comment in this thread too.

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 6633
Re: Estimating future returns from stocks (and bonds)
« Reply #16 on: July 03, 2018, 09:39:54 AM »
Saying 2% is bad, but 11% is good means you need to hit the books a bit.  I would be fine with 2% real returns in a 0% inflation or deflationary market.  11% returns sucked during the peak of the oil/energy price driven inflation era of the 70s.
You're stripping my comments of their context: the article's prediction that starting now, forward returns will be in the range of 2-12%.  My comments are based on that section of the article, which uses today's inflation (2-3%, I believe), for a range of 0-10% real returns.

hodedofome

  • Handlebar Stache
  • *****
  • Posts: 1463
  • Age: 44
  • Location: Texas
Re: Estimating future returns from stocks (and bonds)
« Reply #17 on: July 05, 2018, 11:29:22 AM »
This model is pretty good, and David Merkel keeps it updated quarterly I believe. http://alephblog.com/2016/04/09/estimating-future-stock-returns/

SeattleCPA

  • Handlebar Stache
  • *****
  • Posts: 2369
  • Age: 64
  • Location: Redmond, WA
    • Evergreen Small Business
Re: Estimating future returns from stocks (and bonds)
« Reply #18 on: July 05, 2018, 12:38:16 PM »
I can't easily tell if that's a nominal return (in which case it looks a lot like the CAPE10 prediction) or a real return.

It sounds like he's talking "nominal" and not real though... (That impression meshes with his statement to save more.)