Market Return

[Jane]

Everytime I use some financial calculators, they always ask me to enter estimated future return. And everytime, my knee-jerk reaction is always a frustrated "How do I know?!". Of course my second reaction is to enter 4-6% after inflation. That usually gets me thinking of the conventional wisdom of "8-10% market return" (before inflation).

My question is: where did this number come from and how did it get justified to estimate future return of equity investment?

I read many times about some studies that showed that the return of equities from 1800 - present was a consistent 8-10% before inflation. And from these studies, people in finance/financial media concluded that 8-10% is "reasonable" to predict future return on investment.

The trouble is for me, I think this kind of thinking is flawed.

First, I wonder how the "studies" got their numbers? Survivor bias is mentioned many times. To me, "survivor bias" means we should realize that we are talking about a specific case study and thus instead of concluding 8-10% return in general, we should conclude "if we invest in winning companies and hold this investment for decade(s) then history shows that we may get 8-10% return". But how easy it is to always buy winning companies/funds and hold them for decades? Choosing/holding such companies/funds is only easy in retrospect.

Second, these "studies" surveyed mostly US companies in US market. This "8-10% return" resulted during a period which US grew from a new nation to world's economical superpower. Again, that is a very specific time in history which makes this a "special case study". Will US grow from superpower to superpower^3 and will it affect the market return similarly? Should we invest in countries poised to be the next superpower to duplicate historical studies? (And which country anyway?)

Third, history can and has at times, repeated itself in similar circumstances. So is the market today is similar with the markets of the past? Is the economy today similar with the economy of the past? If they are 2 different circumstances then how can we conclude that history will repeat itself? When did history decide then it owes us 8-10% return when we can't produce a similar circumstance where it can repeat itself?

Can anyone put me at ease and show me (or point some books for me) that this magic number can be used reasonably to estimate future return? At work, if I say "We will have X flow since the pump will produce y pumping capacity", my boss will then ask me "How do you know the pump has y capacity?". I will certainly get in trouble if I say, "Well operators said that in the past 10 years the pump has been producing y". That answer would not be acceptable. My boss can say, "well, it could be a different pump, pump could be old etc. Find out what pump they intend to use and what exactly its capacity". Ok, it's long-winded but my point is that we do not use historical data in our work to estimate future production, we use present data that we know will still be applicable in the future (ie. you don't replace pump every year). I am having trouble at the thoughts that I am using "historical" data in my financial planning.

My naive concern is that you can't predict future return/or it's more of "vodoo magic" than exact science.

What do you think?

Jane

 

[Bob_Smith]

Jane, read this:

http://www.vanguard.com/bogle_site/sp20030605.html

 

[Hyperborea]

First, I wonder how the "studies" got their numbers? Survivor bias is mentioned many times.

Survivor bias can be discounted on things like S&P500 which takes us back to 1927 (and back to 1880 IIRC from the Cowles commission numbers). The reason for this is that usually before a company would implode it would fall from the S&P500. If it imploded while on the S&P500 the results are included. So, if you are investing in say an S&P500 index fund you can rely on that historical number as historically valid.

Second, these "studies" surveyed mostly US companies in US market. This "8-10% return" resulted during a period which US grew from a new nation to world's economical superpower. Again, that is a very specific time in history which makes this a "special case study". Will US grow from superpower to superpower^3 and will it affect the market return similarly?

Shhhh! You keep bringing up the pink elephant in the corner and the Americans will get upset. I've mentioned this a number of times here and on other sites (TMF for example) and the Americans regularly dismiss this because nothing will happen to stop the US from remaining the world superpower. The same thing that happened to every other country won't happen to them.

Personally, I would think that it is prudent to diversify across the world market. I've been doing this with my investments pretty much since I seriously started investing. You might want to look at this paper for a discussion of international return data - http://www.gsm.uci.edu/~jorion/papers/century.pdf

Should we invest in countries poised to be the next superpower to duplicate historical studies? (And which country anyway?)

The best advice that I've seen here is to invest in a world market portfolio and you will capture most of the return as if you had "magically" guessed the correct answer to who will have the best returns.

My naive concern is that you can't predict future return/or it's more of "vodoo magic" than exact science.

What do you think?

There is definitely a large amount of "voodoo" to it. The best that we can do it seems is to diversify, use what correlations have the most evidence and some reasonable logic behind them (to differentiate them as much from the random correlations), have some fallbacks (be able to work longer if returns aren't up to snuff, skills to return to working if FIRE fails, etc.), and hope for the best.

 

[haha]

My naive concern is that you can't predict future return/or it's more of "vodoo magic" than exact science.
What do you think?
Jane

Jane, speaking only for myself, I think you are not naive, you are right to question these numbers. You have pointed to the undeniable fact that the Emporer is naked.

The way of the world seems to be that things go along most of the time without discontinuous change.

Then suddenly that is no longer true.

My personal response to this given current valuations is to stay light in equities, and of the equities I do own to prefer yield and likely stability over other considerations. Also I keep fixed income very short duration and with little or no credit risk.

I also have overweighted energy at 17%, and gold equities at 7%. Don't know for sure whether this is a good plan or not. So far the energy has been spectacular, the gold well above cost of capital but perhaps not yet an adequate return considering risk. I also plan to invest between 1 and 1.5% per year in index put options.

I guess it seems to me that our assumptions are so important, that fine tuning with calculators etc. could be deceptive. Except when they show that something clearly won't work.

Mikey

 

[ESRBob]

Jane, read this:

http://www.vanguard.com/bogle_site/sp20030605.html

Bob,
Nice link,
thanks for this. Didn't know about this Bogle research info inside the vanguard site...

ESRBob

 

[unclemick]

One of my favorites among many Bogle articles.

 

[Bob_Smith]

Bob,
Nice link,
thanks for this.  Didn't know about this Bogle research info inside the vanguard site...

ESRBob

ESRBob and unclemick, I re-read that periodically. It helps me refocus.

 

[Nords]

Re: Market Return (whatever that is)

My question is: where did this number come from and how did it get justified to estimate future return of equity investment?

The trouble is for me, I think this kind of thinking is flawed.

Yup, you're right. Past results still aren't good guarantees of future performance, whether they pass the survivor bias test or not.

When you come up with a better method, let me know!

My naive concern is that you can't predict future return/or it's more of "voodoo magic" than exact science.

What do you think?

Jane

Even if you can make some sort of macro prediction, chaos is still working against you. For a still gloomier summary of your disturbing thoughts, try Bernstein's "Retirement Calculator from Hell" series!

http://www.efficientfrontier.com/ef/998/hell.htm
http://www.efficientfrontier.com/ef/101/hell101.htm
http://www.efficientfrontier.com/ef/901/hell3.htm
http://www.efficientfrontier.com/ef/103/hell4.htm
http://www.efficientfrontier.com/ef/403/hell5.htm

Unless you elect to live in Ted Kaczynski's old place and eat MREs for the rest of your life, then projecting old data is as good as it gets. Subtract a "safety factor" from your returns (right, "How the @#$% do I know what safety factor to use?!?) depending on your degree of paranoia/pessimism.

Or try a library copy of Bud Hebeler's planning book on his iterative closed-loop negative feedback approach: http://www.analyzenow.com/retirement_plan_book.htm.

 

[John Galt]

Hey Nords, good post! I was thinking of moving to
Ted Kaczinski's "old place" but the authorities
disassembled it so that is no longer an option. Bummer!

John Galt

 

[Jane]

Thanks for all the links I will check them out (slowly though - here in Canada it's not Thanksgiving holiday).

As Nords said, it's a disturbing thought but in principal I think that it is the most important concept to grasp in investing. What's more disturbing is that many investors based their picks on "historical performance" without understanding that: history *does not* owe us anything!

It's distrubing that 3 financial advisors I interviewed few years back mouthed off the "historical number" as if it was equivalent to say, Gravity. Maybe it was why I decided against FA altogether.

Like Hyper I diversified (as best as I could). I hold index funds for Canada, US and International. I still have some no-load mutual funds in my pension plan because I don't have access to index funds there. And I am crossing my fingers hoping for the best.

Jane

 

[ats5g]

Hey Jane,

And the next time someone tries to tell you that stocks are "safe long term investments", show them this Jorion paper :

The Long-Term Risks of Global Stock Markets

- Alec

 

[ats5g]

opps, hit post too soon.

This Journal of Financial Planning article maps out how people arrive at these numbers, and what may be logical/reasonable given the current prices of stocks:

What Do Past Stock Market Returns Tell Us About the Future?

It's also thought provoking to consider the question, "Did any investor actually get these historical rates of return?" Given the lack of tax efficient market index funds, heavy loads and expenses, I would highly doubt it.

- Alec

 

[hocus]

"My naive concern is that you can't predict future return/or it's more of "vodoo magic" than exact science."

You brought some good questions to the table, Jane. This is an important thread, in my view.

That said, I agree with the point made by Nords. If an aspiring early retiree is not enthused with the idea of using historical data to get a fix on how her income streams are likely to hold up over the long run, what the heck is she going to use instead? You need a number to plug into your plan. At some point, your vague thoughts about how your investments might perform must be translated into a number. Making use of historical data permits you to translate the words that your vauge thoughts are made up of into the numbers you need to put a plan down on paper.

It makes sense to use calculations performed by others to come up with starting-point numbers. But it is critical to understand the assumptions in which those calculations are rooted and to perform your own assessment as to whether the numbers that result from the calculations make real-world sense. If not, you need to make adjustments to them before incorporating them into your plan.

Developing a deep understanding of the assumptions behind the numbers you use is critical. Accepting numbers when you don't fully grasp the assumptions on which they are based is dangerous. It is probably better to make use of no numbers at all than to do that. There's an old saying that "It's not the things you don't know that really hurt you, it's the things you know for certain that just ain't so." When you accept a number without understanding the assumptions on which it is based, you run the risk of coming to know for certain something that ain't so.

 

[Roger_R]

If an aspiring early retiree is not enthused with the idea of using historical data to get a fix on how her income streams are likely to hold up over the long run, what the heck is she going to use instead?

I afraid I am a bit of contrarian in my own thoughts on this. I'm not entirely negative about using historical data, but take it with a larger grain of salt than many. If you want an idea of how long your income stream is going to last, without using historical data, you could include in your considerations your total savings, divided by your remain life expectancy. Consider a total in vestment in inflation protected securities and add a small bonus for real rate of return.

 

[Bob_Smith]

If you want an idea of how long your income stream is going to last, without using historical data, you could include in your considerations your total savings, divided by your remain life expectancy.

Roger, I'm not sure I follow how that formula would tell you how long your income will last. If you divide savings by life expectancy, wouldn't the result provide a rough idea of what you can spend? In other words, a 40 year lifespan would always produce a 2.5% withdrawal rate, right? Or am I missing something?

 

[Roger_R]

If you divide savings by life expectancy, wouldn't the result provide a rough idea of what you can spend? In other words, a 40 year lifespan would always produce a 2.5% withdrawal rate, right? Or am I missing something?

Yes.  That's a simple approach without using historic returns.  Throw in a real rate of return of one or two percent if you'd like.  I realize that this tends to over simplify and relies on the fact that you are entirely invested in inflation protected securities.  

If you are thinking perhaps that I am somewhat out of my gourd, it is a little similar to a method mentioned by the Coffee House invester in one of their "portfolio ponderings" articles a few weeks ago.  Not to say they may be a bit daft, too .

 

[unclemick]

And then there is 'old school' - take the market return of 'your portfolio' - div/interest and let the principle ride. That's about 3% for us nowadays.

 

[Hyperborea]

If you want an idea of how long your income stream is going to last, without using historical data, you could include in your considerations your total savings, divided by your remain life expectancy.  Consider a total in vestment in inflation protected securities and add a small bonus for real rate of return.

So, you're suggesting that somebody retiring at 40 with maybe an estimate of 60 years of life left who wants a $40K / year income needs to have $2.4 Million? Should they deposit this all at their local bank? Or buy gold coins and bury them in mason jars in the back yard?

 

[Roger_R]

The original issue was if there was a method to calculate income stream without historic returns or if you have little faith in them. That is what I presented.

Obviously if you put your savings in the bank or buried it in your back yard inflation would eat you alive if you had 60 years of retirement.

 

[hocus]

There is a tension at the core of these discussions that can never be overcome because it is inherent to the SWR project. Most early retirees have two important investing goals that are in conflict with each other: (1) most want to participate in the long-term growth potential offered by stocks; and (2) most want predictable income streams. TIPS provide predictable income streams, but they lack the growth potential possessed by stocks. Stocks offer growth potential, but they offer volatile returns and thus do not guarantee predictable income streams.

One of the core purposes of SWR analysis is to resolve this conflict in a satisfactory way. The SWR analyst says, "What if we were to look at the worst-case scenario for stock returns, the worst returns sequence that we have seen in history? If we were to calculate the income stream that would work in that case, would that be good enough for you? Would that be predictable enough for you to justify going with stocks over something like TIPS?"

The SWR is NOT the expected return of your stock investment. It is the highest take-out number that will work in the event that a worst-case scenario pops up. The idea is get a number that is comparable to the number that would apply for an investment class that is highly predictable, like TIPS.

If the SWR for a stock index is higher than the SWR for an asset class with a lower growth potential, it's hard for me to see how you justify going with the alternative asset class. If the high-growth class provides the safety you need, you generally are going to go with the high-growth class. It's only when the SWR for stocks is lower than the SWR for the alternative class that you need to devote a lot of thought to the trade-off between safety and growth.

Just about everyone understands the nature of this trade-off. It comes up in some form in just about all investing decisions. The magic of SWR analysis is that it permits you to put numbers to the trade-offs. It doesn't answer all of the questions for you. But it informs your decisions in a powerful way by giving you something hard and objective (numbers) to toss around so that you no longer need to rely on just subjective impressions.

 

[Roger_R]

*****, that's such a good explantion I think I'll print it out.  Thanks.  The issues that I have are that we are using performance of an economy making model T's and tube radios, to predict one that is quickly becoming a service economy.  Performance in years when less than 5% of the American households own mutual funds, predicting one where now about 50% of house holds have some market influence.  Are we using an apple to predict an orange?

There is also the assumption that the markets will never assume some sort of statistical indepenence.  When ever issues come up here like peak oil, China economic dominance, terrorist threats, etc., the answer is always, we got throught worse, we'll get through this. Or "not another disaster prediction". As if there is no chance any of these things could swing the reversion to the mean. I was lunching with a friend two weeks ago after the 60 minutes program featuring a former CIA official predicting almost certain nuclear attemps on our cities.  My friend thinks this is probable, but is almost 100% equities in his retirement funding.  This seems like a conflict of beliefs to me.

The SWR theory and use are great references that are a good base to consider.  I am not a full doubter and indeed it's as a reference for my own future retirement.  However, as the gospel to assure my well-being for the next 40years, I have reservations. When we say a 95% chance of success, I'm thinking more like a 75% chance considering the events around the numbers.  In the end, they are only numbers predicting other numbers.  The historical events surrounding these numbers as well as the current and future global econonomics are equally as important as the numbers themselves.

 

[hocus]

When we say a 95% chance of success, I'm thinking more like a 75% chance considering the events around the numbers.

I think you are right to see it this way. The 95 percent number presumes a future no worse than the worst that we have seen in the past. Nuclear war or hyperinflation or mass spread of disease could put us in a circumstance where the future for this particular country is worse than anything we have in our historical record. If that happens, the 95 percent number is out the window.

To get the true odds of a withdrawal rate not working, you need to add the percentage chance of a future worse than the worst we have seen in the past to the percentage chance of failure of your portfolio presuming that we do not see anything worse in the future than the worst that we have seen in the past.

 

[Hyperborea]

Obviously if you put your savings in the bank or buried it in your back yard inflation would eat you alive if you had 60 years of retirement.

I understand that it's obvious - that's why I asked it since you seemed to be ignoring this in your suggestion to basically save up all you need for retirement and put it somewhere that you left unspecified.  In another thread on this board you were also suggesting bank accounts because they had a (marginally) longer history of usage by "common people" than stocks.  If history of usage is important then precious metals and goats win by thousands of years.  Where were you suggesting to put this very large sum of money that has a "long enough" history of usage by "common people"?

There's another factor in your "plan" that you've left unspecified. What you also haven't explained is how our potential early retiree is going to come up with 2.5 times as much retirement stash unless it's to continue working until retirement is no longer early.

 

[hocus]

RogerR:

I think you have made some good points on this thread. I much appreciate your participation.

Here's a link to a thread from the SWR Research Group board in which JWR1945 looks at some numbers for TIPS. It might help you develop some data-based strategies that are in tune with the investing concerns you have put forward here.

http://www.nofeeboards.com/boards/viewtopic.php?t=2598

JWR1945: "A 100% TIPS portfolio at currently available rates are truly safe at a withdrawal rate of 4.0% more than thirty years. This finding is highly significant considering the hazards in today’s stock market."

 

[wabmester]

The 95 percent number presumes a future no worse than the worst that we have seen in the past. Nuclear war or hyperinflation or mass spread of disease could put us in a circumstance where the future for this particular country is worse than anything we have in our historical record. If that happens, the 95 percent number is out the window.

This appears to be a common misperception.   It won't take a disaster for the future return to be worse than the past (perhaps much worse).   There are several ways the future can be worse than the past worst-case:

1) We can have an economic disaster on par with the Great Depression.   This is the least likely scenario, but there's still a 30% chance somebody would see something like this in their lifetime.

2) We can have less RTM to the upside.   There is no law that says we'll have a max of 3 bad years in a row followed by a bunch of great years in the stock market.   If we ever have a longer bear market, or less of a bull following a bear, than we have historically, then future returns will likely be worse than past returns.

3) The economy can grow slower than it has in the past.   This is extremely likely, and may result in a lower long-term return than we've had historically.

4) The rate of equity dilution can increase.   Bernstein has shown that long-term market returns closely track economic growth less the rate of equity dilution.   Equity dilution occurs when companies issue new shares, which they routinely do when they grant stock options and acquire new companies.   This is a subtle effect that is likely to further dampen future returns.

In other words, chances are very good that the future won't be as rosie as the past.