Past returns vs future returns

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
cfa-ish
Posts: 15
Joined: Mon Oct 18, 2021 5:43 pm

Past returns vs future returns

Post by cfa-ish »

It seems axiomatic to some investors that if a stock zooms upwards quickly, it will keep zooming up. And to others, it is equally axiomatic that, since it has drawn returns from the future into the present, its future returns will be less.
I don't think either is a valid rule of thumb for an individual stock. Assuming you are doing fundamental analysis, you have to do your due diligence, form some idea of its future earnings and earnings growth (or X and growth in X where X is somewhat predictable and related to earnings, e.g. X = sales or FCF or whatever makes sense for that company). And then how that X will affect the current price.

But I wonder how the stock market as a whole should be viewed. On one hand, the momentum effect is real, for up to a year. On the other hand, earnings growth is not going to exceed the GDP (US or your country's or the world GDP depending on the definition of "stock market"). On the third hand, in the recent past (last 15 years) S&P 500 earnings growth has been spectacular. On the fourth hand, sales growth has been so-so, and (on the 4.5th hand) interest rates have been unusually low in the last 15 years (at least after the 2008-9 crisis). On the fifth hand, as Lynch said, investors have lost more money anticipating recessions than in actual recessions (paraphrasing slightly). On the sixth hand...

So this is not a theoretical question. Are you doing anything differently because the market is so highly valued, several standard deviations above its historical average? Or still a steady X% VTI, Y% VXUS, (100-X-Y)% BND (or the equivalent in your three-fund portfolio)? Trust an yearly rebalancing to make the decision for you? i.e. if VTI goes too far up you will automatically sell some of it and invest in VXUS + BND and so on?

I don't know what to do, especially if interest rates start going up and all three ETFs start coming down in price. I realize that personal circumstances etc will vary. So just looking for any general guidelines or your own strategies.

I do believe that inflation will tick up, and is not as transitory as the Fed has (publicly) said. No idea if the general interest rates will follow suit (except on CPI-linked securities).
Karamatsu
Posts: 1447
Joined: Mon Oct 27, 2008 2:42 am

Re: Past returns vs future returns

Post by Karamatsu »

I share your uncertainty about the future but alas have no real insights other than rebalancing because (of course) we have no way of knowing what's going to happen or how the various forces will interact. Generally I agree with the "What do you know that the market doesn't?" sentiment, but of course it is true that, for example, someone in their 20's can afford to just let things ride while someone in their 70's cannot. I suppose people who can't afford to let the market do its thing might want to consider hedging, either directly or via one of the leveraged ETF/CEF offerings, but I have no recommendations there.

There is the larger question, though, of X, Y, and (100-(X+Y)). If we think the historical record is reliable, it's easy to choose numbers that will either maximize return or, alternatively, minimize volatility for a chosen return. But as you say, we're at the tail end of bull markets for both US equities and US bonds, while other asset classes haven't contributed all that much. Still... what of it, and what do I know that the market does not? All I can see is sticking to the plan.
Ferdinand2014
Posts: 2390
Joined: Mon Dec 17, 2018 5:49 pm

Re: Past returns vs future returns

Post by Ferdinand2014 »

Growth of a countries GDP may not be the ceiling of the same countries stock market because the stock market is not the economy (it doesn’t represent private companies ranging from the local plumber to the Koch brothers empire, government state, local, etc) and large multinational companies pull revenue and grow earnings from multiple countries. Share buybacks, share dilution, etc all have effects as well that may not directly reflect the GDP of an economy vs the stock market.

No, I am not doing anything different based on my intuition regarding the current value of the stock market. I buy every week and keep enough cash so I do not need to sell unless of my choosing.
“You only find out who is swimming naked when the tide goes out.“ — Warren Buffett
magicrat
Posts: 1191
Joined: Sat Nov 29, 2014 6:04 pm

Re: Past returns vs future returns

Post by magicrat »

No, you can't time the market.
atdharris
Posts: 2091
Joined: Wed Jan 02, 2019 2:18 pm

Re: Past returns vs future returns

Post by atdharris »

I don't try to time the market. I've kept my allocation the same since March 2020 and plan to continue that path. I can't predict what the market will do, but I don't want to be sitting on a lot of cash trying to hit the bottom.
Ramjet
Posts: 1464
Joined: Thu Feb 06, 2020 10:45 am

Re: Past returns vs future returns

Post by Ramjet »

cfa-ish wrote: Mon Oct 18, 2021 6:23 pm Are you doing anything differently because the market is so highly valued...
It was a factor in me moving to global market cap weight in stocks

If I had better funds in retirement accounts I would probably have a SCV tilt too
Topic Author
cfa-ish
Posts: 15
Joined: Mon Oct 18, 2021 5:43 pm

Re: Past returns vs future returns

Post by cfa-ish »

magicrat wrote: Tue Oct 19, 2021 8:27 am No, you can't time the market.
Thanks everyone for your inputs.

Specifically about timing the market, does that include increasing your DCA when the market is cheap and decreasing it when market is expensive? Irrespective of the time periods. Meaning, I am not predicting that the market will go down in the next 3, 6 or 12 months, ergo don't put money in now. But I am wondering if it will go down eventually due to a combination of high valuations and rising interest rates. Logically, what else can happen?

The blind DCA does have a huge attraction that I won't ever kick myself in the future for not being clairvoyant. Downside is to have to live with poor or negative returns that all other indexers are also getting (and most active funds, I bet).
strakert
Posts: 85
Joined: Sat Oct 02, 2021 12:06 pm

Re: Past returns vs future returns

Post by strakert »

cfa-ish wrote: Tue Oct 19, 2021 1:17 pm
magicrat wrote: Tue Oct 19, 2021 8:27 am No, you can't time the market.
Thanks everyone for your inputs.

Specifically about timing the market, does that include increasing your DCA when the market is cheap and decreasing it when market is expensive? Irrespective of the time periods. Meaning, I am not predicting that the market will go down in the next 3, 6 or 12 months, ergo don't put money in now. But I am wondering if it will go down eventually due to a combination of high valuations and rising interest rates. Logically, what else can happen?

The blind DCA does have a huge attraction that I won't ever kick myself in the future for not being clairvoyant. Downside is to have to live with poor or negative returns that all other indexers are also getting (and most active funds, I bet).
You may want to search for "Value Averaging". William Bernstein has talked about this in the past. This blog post may be a good starting point to investigate further. Note the intro:
A reader asks: "I have a fairly large sum of money to invest, but I’m nervous about current stock market levels. What are my options to put this money to work in the markets?"
I’ve actually been getting this question or some variation of it from quite a few people for the past few years.
And then note that the post is from 2015.

And yes, this is also market timing. The risk is that the dip you're looking to buy may not appear for several years and you may sit on the sidelines and watch the market run up. Even if the market does dip, it may not go below the level it was at when you decided to get out and wait. E.g. Alan Greenspan made his famous "irrational exuberance" speech in December 1996. The tech bubble didn't pop until March 2000. Even then, it didn't fall below its December 1996 level.
Karamatsu
Posts: 1447
Joined: Mon Oct 27, 2008 2:42 am

Re: Past returns vs future returns

Post by Karamatsu »

The blind DCA does have a huge attraction that I won't ever kick myself in the future for not being clairvoyant. Downside is to have to live with poor or negative returns that all other indexers are also getting (and most active funds, I bet).
Oh, sorry. I misunderstood what you were asking. If the question is one of entering the market, then the desire to DCA just means that your asset allocation is too aggressive. Since the AA should be designed to produce acceptable results no matter what the markets do (within some expected distribution, of course -- there are always "unknown unknowns"), then as long as today's market conditions fall within that distribution, you should be confident about investing all of your capital right now. If you're not confident enough about your AA to do that, it's a signal that you need to dial down the risk (usually by reducing the percentage allocated to equities) until you are willing to lump-sum in.

Situations like these are extremely valuable for understanding your actual risk tolerance, as opposed to what we like to imagine.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: Past returns vs future returns

Post by HomerJ »

cfa-ish wrote: Tue Oct 19, 2021 1:17 pm
magicrat wrote: Tue Oct 19, 2021 8:27 am No, you can't time the market.
Thanks everyone for your inputs.

Specifically about timing the market, does that include increasing your DCA when the market is cheap and decreasing it when market is expensive? Irrespective of the time periods. Meaning, I am not predicting that the market will go down in the next 3, 6 or 12 months, ergo don't put money in now. But I am wondering if it will go down eventually due to a combination of high valuations and rising interest rates. Logically, what else can happen?
Sure, the market will go down in the future. But what if it goes up 20% first before going down 10%. Or up 100% before crashing 40%? Would waiting to invest been a smart move or a dumb move? I mean, you were right that a crash will happen someday...

But people were saying that 5 years ago, and 3 years ago, and anyone who "decreased their DCA" back then is still waiting.

Here's the important thing to remember... The long-term nominal 9%-10% annual return of the stock market INCLUDES the crashes.

In the past, you didn't have to avoid the crashes to make good long-term returns and become wealthy.

Even investing at the very top hasn't hurt that badly over the long run.

If you were just blindly DCA investing back in 2000, you would have bought in at the highest valuations in U.S. history.

That money has grown 7.39% nominal a year since then. That was the WORST time to invest, and it still turned out pretty good. $10,000 has turned into $47,000. And of course, one was DCAing in 1999, and 1998, and 1997, and 1996, and 2001, and 2002, and 2003.

And all THOSE years turned out even better...

We mostly believe in buy and hold here for the long run, and don't try to predict what the market will do in the short-run. Money you need in the next 5 or 10 years, probably shouldn't be in the market. Money that can sit there long-run, just invest it now and don't worry about the short-term swings.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
000
Posts: 8211
Joined: Thu Jul 23, 2020 12:04 am

Re: Past returns vs future returns

Post by 000 »

I'm bullish on gold and miners both as a diversifier to the usual stock/bond portfolio and as standalone asset classes.
Topic Author
cfa-ish
Posts: 15
Joined: Mon Oct 18, 2021 5:43 pm

Re: Past returns vs future returns

Post by cfa-ish »

Karamatsu wrote: Tue Oct 19, 2021 7:55 pm
The blind DCA does have a huge attraction that I won't ever kick myself in the future for not being clairvoyant. Downside is to have to live with poor or negative returns that all other indexers are also getting (and most active funds, I bet).
Oh, sorry. I misunderstood what you were asking. If the question is one of entering the market, then the desire to DCA just means that your asset allocation is too aggressive. Since the AA should be designed to produce acceptable results no matter what the markets do (within some expected distribution, of course -- there are always "unknown unknowns"), then as long as today's market conditions fall within that distribution, you should be confident about investing all of your capital right now. If you're not confident enough about your AA to do that, it's a signal that you need to dial down the risk (usually by reducing the percentage allocated to equities) until you are willing to lump-sum in.

Situations like these are extremely valuable for understanding your actual risk tolerance, as opposed to what we like to imagine.
I am not sure if this is anything to do with my AA or risk tolerance. I am OK having 100% equities, but I'd like to have a lot of them instead of only a few.

Analogy: I am perfectly willing to buy and eat a hamburger every single day. Especially if that is an In-N-Out burger (California chain).
Doesn't mean I want to pay $15 for the burger when all indications are the burger is not worth more than $5 and the chain will have to adjust its prices. I don't know when, but the cash is not burning a hole in my pocket. I realize I am giving up the pleasure of eating the burger "for a while" where I don't know how long that will be.
tibbitts
Posts: 23716
Joined: Tue Feb 27, 2007 5:50 pm

Re: Past returns vs future returns

Post by tibbitts »

cfa-ish wrote: Wed Oct 20, 2021 11:58 am Analogy: I am perfectly willing to buy and eat a hamburger every single day. Especially if that is an In-N-Out burger (California chain).
Doesn't mean I want to pay $15 for the burger when all indications are the burger is not worth more than $5 and the chain will have to adjust its prices. I don't know when, but the cash is not burning a hole in my pocket. I realize I am giving up the pleasure of eating the burger "for a while" where I don't know how long that will be.
Except that currently there actually is a hole in your cash pocket in the form of inflation, and it's like money is falling out of it constantly.
lws
Posts: 831
Joined: Tue Apr 25, 2017 6:12 pm

Re: Past returns vs future returns

Post by lws »

Sticking to the plan.
Doing nothing differently.
Accepting what the market provides.
Taking the pain and the gain as they come.
strakert
Posts: 85
Joined: Sat Oct 02, 2021 12:06 pm

Re: Past returns vs future returns

Post by strakert »

cfa-ish wrote: Wed Oct 20, 2021 11:58 am
Karamatsu wrote: Tue Oct 19, 2021 7:55 pm
The blind DCA does have a huge attraction that I won't ever kick myself in the future for not being clairvoyant. Downside is to have to live with poor or negative returns that all other indexers are also getting (and most active funds, I bet).
Oh, sorry. I misunderstood what you were asking. If the question is one of entering the market, then the desire to DCA just means that your asset allocation is too aggressive. Since the AA should be designed to produce acceptable results no matter what the markets do (within some expected distribution, of course -- there are always "unknown unknowns"), then as long as today's market conditions fall within that distribution, you should be confident about investing all of your capital right now. If you're not confident enough about your AA to do that, it's a signal that you need to dial down the risk (usually by reducing the percentage allocated to equities) until you are willing to lump-sum in.

Situations like these are extremely valuable for understanding your actual risk tolerance, as opposed to what we like to imagine.
I am not sure if this is anything to do with my AA or risk tolerance. I am OK having 100% equities, but I'd like to have a lot of them instead of only a few.

Analogy: I am perfectly willing to buy and eat a hamburger every single day. Especially if that is an In-N-Out burger (California chain).
Doesn't mean I want to pay $15 for the burger when all indications are the burger is not worth more than $5 and the chain will have to adjust its prices. I don't know when, but the cash is not burning a hole in my pocket. I realize I am giving up the pleasure of eating the burger "for a while" where I don't know how long that will be.
Whether due to fear or greed or virtue (patience, intelligence, knowledge), the end result is the same. Market timing has been shown to underperform buy and hold.

Do you think all the active investors prefer to buy a lot of equities when they're expensive instead of cheap? Do you think they aren't aware of the high valuations, low expected returns or inflation uncertainty you're seeing? Remember that the market price is set by active investors. So if the price is high, it's because two well informed (much more so than you and I) and opinionated (enough to bet millions of dollars on it) market participants agreed to exchange securities at that price, both expecting to profit from the transaction - one expecting it to go higher and one expecting it to go lower in the future. If you want to side with one of them, that's an active bet. You know what the studies say about that.

Consider that the suggestions of "nobody knows nothing" and "take what the markets give" are not coming from a position of ignorance, laziness, or risk aversion. They're coming from academics and practitioners who have thought long and deeply about the markets, and from individuals who have learned these lessons the hard way after "market timing" in various ways.
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: Past returns vs future returns

Post by alex_686 »

cfa-ish, how much of the CFA curriculum have you read? If you have you should know that...

Determining current expected returns are hard. Returns are based on a host of economic factors and that these factors and their relationships are not stable.

You can see that in historical data. Returns, volatility, and correlations are time varying. While they may be mean reverting within a secular period they are not between different periods. The stock market is pretty close to a random walk. Specifically that means that historic price data contains little informational value on tomorrow's return.

If modeling current expected returns are hard it is a nightmare to predict what future expected returns. To simplify, to determine the current value of a stock we need to estimate cashflows, interest rates, and the Equity Risk Premium for the next 10 years. To calculate what future returns will be you will need to do the same calculations for years 11 to 20.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
Topic Author
cfa-ish
Posts: 15
Joined: Mon Oct 18, 2021 5:43 pm

Re: Past returns vs future returns

Post by cfa-ish »

Fair enough, you guys are changing my mind.
I will still hold my nose, purely out of vanity, but I will buy the stinky fish :D

Of course, I will be happy to be proven wrong and enjoy returns in the next 10, 20, 30 years as they have been for the last 10, 20, 30 years.

Alex - I passed all CFA exams 3/3 in 1.5 years (December, June, June). So I have read it all. Now how much I have believed in some of it is a different question.
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: Past returns vs future returns

Post by alex_686 »

cfa-ish wrote: Wed Oct 20, 2021 1:16 pm Alex - I passed all CFA exams 3/3 in 1.5 years (December, June, June). So I have read it all. Now how much I have believed in some of it is a different question.
So lets work through this - what parts do you take issue with?

1: There is a "intrinsic, fundamental, correct" P/E ratio that you calculate. The CFA lays out about 1/2 a dozen different approaches to calculate this. This is different than the market P/E ratio.

2. If your calculated P/E ratio is above the market's P/E ratio then the market is undervalued. If below, then overvalued.

3. The P/E ratio is a relative measure. i.e., you can't use itself to determine if the market is over/under valued. If the PE ratio is higher than the historical average this does not mean it is overvalued. It could be because interest rates are historically low.

So, let us stop at point 3.
cfa-ish wrote: Wed Oct 20, 2021 1:16 pm I will still hold my nose, purely out of vanity, but I will buy the stinky fish
OK, why do you think the fish is stinky?

I can make a solid argument that today's historically high PE ratios are correct. In particular I will point to low interest rates on long term government bonds (very solid evidence here) and a low Equity Risk Premium (indirect evidence here).

But this is just my opinion. What is yours and why? How confident are you of it?

Now, this is all in the CFA curriculum and ties about to your question. I don't take the CFA curriculum as dogma but this is all bread-and-butter investment theory.

Next, you are on Boglehead's which strongly believes in the wisdom of the crowd. Some believe this so strongly they say you can't beat or time the market. I am not exactly in the crowd. Rather I think it is darn hard to beat the crowd and that there are better and easier things to do with skill and time - like research individual stocks.

But then again I work in risk management so I tend to be risk adverse.
Last edited by alex_686 on Wed Oct 20, 2021 1:52 pm, edited 1 time in total.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
strakert
Posts: 85
Joined: Sat Oct 02, 2021 12:06 pm

Re: Past returns vs future returns

Post by strakert »

cfa-ish wrote: Wed Oct 20, 2021 1:16 pm Fair enough, you guys are changing my mind.
I will still hold my nose, purely out of vanity, but I will buy the stinky fish :D

Of course, I will be happy to be proven wrong and enjoy returns in the next 10, 20, 30 years as they have been for the last 10, 20, 30 years.
Diversification may be the only free lunch in investing, but that doesn't mean you'll enjoy the stinky fish being served :)

Good luck! And remember that holding is the hard part. When the inevitable crash and extended drawdown comes and everyone says "I told you so!", don't go "I knew it was coming, should have trusted my instincts" and time the markets.
garlandwhizzer
Posts: 3565
Joined: Fri Aug 06, 2010 3:42 pm

Re: Past returns vs future returns

Post by garlandwhizzer »

I think it is likely that the great returns we've enjoyed in US stocks and bonds over the past 20 years will not continue to the same extent for the next 20. The question really is does this change your investment strategy? Presumably includes an appropriate mix of widely diversified risk assets with less volatile assets like quality fixed income sufficient to allow you to tolerate a deep bear market without panic selling. Granted, the equity and bond markets are richly priced now relative to underlying fundamentals. Granted also, economic growth, inflation, interest rates etc., going forward are all worrisome question marks, perhaps more so than what we've had for the past 20 years. As always, risks loom ahead of us.

If you up know up front how all of this will play in the future you can position your portfolio now to take full advantage of that knowledge. The problem is that no one knows that with actionable certainty in my opinion. I believe a balanced portfolio attuned carefully to your own risk tolerance, financial circumstances, and long term financial goals is now, as it has always been, the best approach for most of us. Making wholesale portfolio changes based on your predictions of the future, changing an already well thought out portfolio is unlikely to beneficial. The thing to concentrate on IMO is not looking for a short cut, but doing the hard work of finding the right portfolio mix for you. This requires balancing conflicting inputs of risk and reward that fits you both objectively and emotionally. That is not easy. Once you have that right, stick with it with only minor changes from time to time based on changing circumstances. Your investing strategy should not change every time the market and economic winds change.

Garland Whizzer
Topic Author
cfa-ish
Posts: 15
Joined: Mon Oct 18, 2021 5:43 pm

Re: Past returns vs future returns

Post by cfa-ish »

alex_686 wrote: Wed Oct 20, 2021 1:40 pm
cfa-ish wrote: Wed Oct 20, 2021 1:16 pm Alex - I passed all CFA exams 3/3 in 1.5 years (December, June, June). So I have read it all. Now how much I have believed in some of it is a different question.
So lets work through this - what parts do you take issue with?

1: There is a "intrinsic, fundamental, correct" P/E ratio that you calculate. The CFA lays out about 1/2 a dozen different approaches to calculate this. This is different than the market P/E ratio.

2. If your calculated P/E ratio is above the market's P/E ratio then the market is undervalued. If below, then overvalued.

3. The P/E ratio is a relative measure. i.e., you can't use itself to determine if the market is over/under valued. If the PE ratio is higher than the historical average this does not mean it is overvalued. It could be because interest rates are historically low.

So, let us stop at point 3.
cfa-ish wrote: Wed Oct 20, 2021 1:16 pm I will still hold my nose, purely out of vanity, but I will buy the stinky fish
OK, why do you think the fish is stinky?

I can make a solid argument that today's historically high PE ratios are correct. In particular I will point to low interest rates on long term government bonds (very solid evidence here) and a low Equity Risk Premium (indirect evidence here).

But this is just my opinion. What is yours and why? How confident are you of it?

Now, this is all in the CFA curriculum and ties about to your question. I don't take the CFA curriculum as dogma but this is all bread-and-butter investment theory.
I realize that a high CAPE is somewhat justified by a combination of low interest rates as I mentioned in a previous post. However, one thing I fervently believe is that nobody knows what interest rates are going to do, not even the Fed. So relying on them is a mistake.
Low ERP reflects investor sentiment, not a law of nature or anything close. Give them a high real risk-free rate and they will flock to bonds from stocks.
Both of these reflect the recent past. No reason to assume they will continue. So it's a matter of opinion as you said.
And - no reason to assume they will stop either. Hence my change of mind. Out of "nobody knows nothing" rather than any conviction that the good times will keep rolling.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: Past returns vs future returns

Post by HomerJ »

Just FYI - Almost no one here is sure the "good times will keep rolling".

What we have is two camps of "Bad times are super likely soon" and "No one knows"

We've had posts about "Bad times are super likely soon" for the past 10 years, so the position of "No one knows" seems more correct to me.
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

Re: Past returns vs future returns

Post by secondopinion »

cfa-ish wrote: Wed Oct 20, 2021 1:59 pm I realize that a high CAPE is somewhat justified by a combination of low interest rates as I mentioned in a previous post. However, one thing I fervently believe is that nobody knows what interest rates are going to do, not even the Fed. So relying on them is a mistake.
Low ERP reflects investor sentiment, not a law of nature or anything close. Give them a high real risk-free rate and they will flock to bonds from stocks.
Both of these reflect the recent past. No reason to assume they will continue. So it's a matter of opinion as you said.
And - no reason to assume they will stop either. Hence my change of mind. Out of "nobody knows nothing" rather than any conviction that the good times will keep rolling.
I do not rely on rates being what they are forever; that does not mean they will do up, however. I would like real returns to be more sane again for bonds, but I cannot count on that.

How is the high real risk-free rate going to show up out of nowhere? It be the same assuming an even lower real risk-free rate going to show up out of nowhere. Granted, they would flock to the better real rate; but it would be arbitraged out for some time.

Assuming good times will continue is the problem; that is not the same as "nobody knows nothing". The former stacks even more onto stocks; the latter just stays put with their allocation.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: Past returns vs future returns

Post by alex_686 »

cfa-ish wrote: Wed Oct 20, 2021 1:59 pmHowever, one thing I fervently believe is that nobody knows what interest rates are going to do, not even the Fed. So relying on them is a mistake.
This one is firmly in my wheel house. We in fact have a pretty decent idea of where interest rates are going. The models and their predictions are well calibrated. I am not saying you can't make money here. But to do so you are going to have to make a accurate large bet on a a probability event.
cfa-ish wrote: Wed Oct 20, 2021 1:59 pmLow ERP reflects investor sentiment, not a law of nature or anything close. Give them a high real risk-free rate and they will flock to bonds from stocks.
Sort of. Supply and demand is the bigger picture. It is also dependent on the supply of savings (which seems to be high) and the demand for capital (which seems to be low). The investor is faced with the poor choice of investing in safe assets (government bonds) which have negative real rates or risky assets. Even pessimistic investors are being herded into equites since it is the least bad option.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
BJJ_GUY
Posts: 882
Joined: Wed Mar 13, 2019 7:45 am

Re: Past returns vs future returns

Post by BJJ_GUY »

alex_686 wrote: Wed Oct 20, 2021 12:35 pm cfa-ish, how much of the CFA curriculum have you read? If you have you should know that...

Determining current expected returns are hard. Returns are based on a host of economic factors and that these factors and their relationships are not stable.

You can see that in historical data. Returns, volatility, and correlations are time varying. While they may be mean reverting within a secular period they are not between different periods. The stock market is pretty close to a random walk. Specifically that means that historic price data contains little informational value on tomorrow's return.

If modeling current expected returns are hard it is a nightmare to predict what future expected returns. To simplify, to determine the current value of a stock we need to estimate cashflows, interest rates, and the Equity Risk Premium for the next 10 years. To calculate what future returns will be you will need to do the same calculations for years 11 to 20.
While price data may have little value on subsequent tape, and returns/vol/correlations are time varying, that does not mean that valuations are not mean reverting within and across time periods. While average multiples may have trended higher over time, this doesn't change the fact that over long timer periods (roughly 10-12 years), the multiples are indicative of directional movements.

What I am not saying is anything about predicting what a future return will be at a specific date.

I am saying that fundamentals link price paid today for cash flows (and terminal value) long into the future. A dollar paid for a stock trading at 5x compared to a dollar paid for one trading at 40x are just not the same thing, so why would anyone expect the same economic benefits from two drastically different prices (assuming all else equal to prove the point)?

I'm totally fine with people saying that it's still really tough to put this information into practice. This is often the case because human behavior gets in the way and we aren't good at missing out on near-term results. The so-called evidence of bad results from market timing tend to be from two things: 1.) Market timing efforts are in deed foolish, but value based investment decisions are not the same thing, and 2.) To reduce exposure to that which is expensive and continues to rip, human behavior/lack of descipline tends to mean reversing course at the wrong time.

Most of us shouldn't be making big swings in our allocation. And at the same time, valuations are real and do help long-term directionality of the market
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: Past returns vs future returns

Post by alex_686 »

BJJ_GUY wrote: Wed Oct 20, 2021 2:49 pm
alex_686 wrote: Wed Oct 20, 2021 12:35 pm cfa-ish, how much of the CFA curriculum have you read? If you have you should know that...

Determining current expected returns are hard. Returns are based on a host of economic factors and that these factors and their relationships are not stable.

You can see that in historical data. Returns, volatility, and correlations are time varying. While they may be mean reverting within a secular period they are not between different periods. The stock market is pretty close to a random walk. Specifically that means that historic price data contains little informational value on tomorrow's return.

If modeling current expected returns are hard it is a nightmare to predict what future expected returns. To simplify, to determine the current value of a stock we need to estimate cashflows, interest rates, and the Equity Risk Premium for the next 10 years. To calculate what future returns will be you will need to do the same calculations for years 11 to 20.
While price data may have little value on subsequent tape, and returns/vol/correlations are time varying, that does not mean that valuations are not mean reverting within and across time periods. While average multiples may have trended higher over time, this doesn't change the fact that over long timer periods (roughly 10-12 years), the multiples are indicative of directional movements.

What I am not saying is anything about predicting what a future return will be at a specific date.

I am saying that fundamentals link price paid today for cash flows (and terminal value) long into the future. A dollar paid for a stock trading at 5x compared to a dollar paid for one trading at 40x are just not the same thing, so why would anyone expect the same economic benefits from two drastically different prices (assuming all else equal to prove the point)?

I'm totally fine with people saying that it's still really tough to put this information into practice. This is often the case because human behavior gets in the way and we aren't good at missing out on near-term results. The so-called evidence of bad results from market timing tend to be from two things: 1.) Market timing efforts are in deed foolish, but value based investment decisions are not the same thing, and 2.) To reduce exposure to that which is expensive and continues to rip, human behavior/lack of descipline tends to mean reversing course at the wrong time.

Most of us shouldn't be making big swings in our allocation. And at the same time, valuations are real and do help long-term directionality of the market
So let us unpack this a bit.

You say that "returns/vol/correlations are time varying, that does not mean that valuations are not mean reverting within and across time periods". If I am understanding the double negatives in this sentence, you are saying that valuations like P/E are mean reverting.

First, I am not squaring your circle. Valuations drive returns. How can we have mean reverting valuations but not mean reverting returns, etc.?

Second, lets make sure you are on the same page on what "mean reverting" means.

Let us take 100 years of monthly P/E data so we have 1,2000 data points. Lets say the average P/E ratio 15. We are currently at P/E 40. If the market was mean reverting then there would be a greater chance of a downward movement than a upwards movement. Or if we are at 5 then a greater chance of a upward movement than a downwards movement.

This does not happen. There is a mean reverting tendency within a secular period but not with prior periods.

There are some periods where economics dictate P/E ratios of 5, others 40. Well, maybe. 5 and 40 are a bit extreme. But even at more common numbers of 10 and 20.

Third, I am confused when you say that you can't predict the future yet there is a difference in investing at P/E of 5 or 40. Which is it? Should we make predictions or not? A stock's value is its future cashflows to equity and their value. Yes, predicting the future is hard, complex, and full of uncertainty. But if not that then investing and gambling are the same. And if we don't base our predictions on valuations then what?

Which takes us back to the OP's question. Does past historical data tell us what future returns will be? And broadly the answer is no, it does not.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
UpperNwGuy
Posts: 9475
Joined: Sun Oct 08, 2017 7:16 pm

Re: Past returns vs future returns

Post by UpperNwGuy »

cfa-ish wrote: Mon Oct 18, 2021 6:23 pm
Are you doing anything differently because the market is so highly valued, several standard deviations above its historical average? Or still a steady X% VTI, Y% VXUS, (100-X-Y)% BND (or the equivalent in your three-fund portfolio)? Trust an yearly rebalancing to make the decision for you? i.e. if VTI goes too far up you will automatically sell some of it and invest in VXUS + BND and so on?
I am simply staying the course. I don't have different investment strategies or tactics as the market moves up and down (or interest rates move up or down).
BJJ_GUY
Posts: 882
Joined: Wed Mar 13, 2019 7:45 am

Re: Past returns vs future returns

Post by BJJ_GUY »

alex_686 wrote: Thu Oct 21, 2021 2:58 pm
BJJ_GUY wrote: Wed Oct 20, 2021 2:49 pm
alex_686 wrote: Wed Oct 20, 2021 12:35 pm cfa-ish, how much of the CFA curriculum have you read? If you have you should know that...

Determining current expected returns are hard. Returns are based on a host of economic factors and that these factors and their relationships are not stable.

You can see that in historical data. Returns, volatility, and correlations are time varying. While they may be mean reverting within a secular period they are not between different periods. The stock market is pretty close to a random walk. Specifically that means that historic price data contains little informational value on tomorrow's return.

If modeling current expected returns are hard it is a nightmare to predict what future expected returns. To simplify, to determine the current value of a stock we need to estimate cashflows, interest rates, and the Equity Risk Premium for the next 10 years. To calculate what future returns will be you will need to do the same calculations for years 11 to 20.
While price data may have little value on subsequent tape, and returns/vol/correlations are time varying, that does not mean that valuations are not mean reverting within and across time periods. While average multiples may have trended higher over time, this doesn't change the fact that over long timer periods (roughly 10-12 years), the multiples are indicative of directional movements.

What I am not saying is anything about predicting what a future return will be at a specific date.

I am saying that fundamentals link price paid today for cash flows (and terminal value) long into the future. A dollar paid for a stock trading at 5x compared to a dollar paid for one trading at 40x are just not the same thing, so why would anyone expect the same economic benefits from two drastically different prices (assuming all else equal to prove the point)?

I'm totally fine with people saying that it's still really tough to put this information into practice. This is often the case because human behavior gets in the way and we aren't good at missing out on near-term results. The so-called evidence of bad results from market timing tend to be from two things: 1.) Market timing efforts are in deed foolish, but value based investment decisions are not the same thing, and 2.) To reduce exposure to that which is expensive and continues to rip, human behavior/lack of descipline tends to mean reversing course at the wrong time.

Most of us shouldn't be making big swings in our allocation. And at the same time, valuations are real and do help long-term directionality of the market
So let us unpack this a bit.

You say that "returns/vol/correlations are time varying, that does not mean that valuations are not mean reverting within and across time periods". If I am understanding the double negatives in this sentence, you are saying that valuations like P/E are mean reverting.

First, I am not squaring your circle. Valuations drive returns. How can we have mean reverting valuations but not mean reverting returns, etc.?

Second, lets make sure you are on the same page on what "mean reverting" means.

Let us take 100 years of monthly P/E data so we have 1,2000 data points. Lets say the average P/E ratio 15. We are currently at P/E 40. If the market was mean reverting then there would be a greater chance of a downward movement than a upwards movement. Or if we are at 5 then a greater chance of a upward movement than a downwards movement.

This does not happen. There is a mean reverting tendency within a secular period but not with prior periods.

There are some periods where economics dictate P/E ratios of 5, others 40. Well, maybe. 5 and 40 are a bit extreme. But even at more common numbers of 10 and 20.

Third, I am confused when you say that you can't predict the future yet there is a difference in investing at P/E of 5 or 40. Which is it? Should we make predictions or not? A stock's value is its future cashflows to equity and their value. Yes, predicting the future is hard, complex, and full of uncertainty. But if not that then investing and gambling are the same. And if we don't base our predictions on valuations then what?

Which takes us back to the OP's question. Does past historical data tell us what future returns will be? And broadly the answer is no, it does not.
We're kind of talking past each other here. Three primary reasons I immediately see that may be the reason you aren't following the distinction I'm attempting to highlight. And it's probably because I'm not articulating it well...

1.) When you say that 'historic price data contains little informational value on tomorrow's return' -- I agree with this. We aren't event talking about returns over time at this point, so I was immediately agreeing, though I think that point has no relevance in this discussion.

2.) I agree with you in general that returns, volatility, and correlations are difficult to predict. This becomes even more challenging when you put an explicit time period as a strict parameter. So without any additional information, it would be a fool's errand to make any sort of assumptions for these three variables based only on historical values for only those three variables. (Basically, I'm saying it's not all that helpful to use portfolio visualizer to run a 20 year back-test and then use those results for the forward looking expectations... A bit of an assumption, but from the stuff you post, I'm pretty sure we're aligned with this point.)

3.) Now if you take valuations, this is different because we're adding new variables into the analysis. I prefer to use other models that have a more reliable/less cyclical denominator, so P/E isn't great. Rather than go into a totally different discussion, I'll use P/E in writing, but it isn't a great ratio. Point is, you are absolutely correct that the only way to square the circle is to understand that a valuation ratio on one hand of an equation = returns. This is simple arithmetic, which I believe was the point you were actually making as well. What I'm trying to say is that when the valuations are higher or lower than the historical average we can determine what future return would be needed for the valuation to move back somewhere close to the historical average... I'm very specifically avoiding using the word "predict" because the connotation is typically associated with a strict time frame for that return to be realized, and that isn't how it works. The return that solves for the valuation multiple (and of course accounting for some other variables, but again not for this conversation), doesn't predict precisely when.

But yes, If I buy something at 40 or 50x earnings/revenue (whatever), I am pretty confident in saying that I have a very strong likelihood thatI am going to lose money on the asset in the long run (lets say something like within 12 years, but not perfect every time). Something trading at 40x can always go to 60x, but that doesn't change the fact that you are still buying the same cash flows into the future. The multiple is tethered to reality, so I am confident that really stretched valuations have quite a good deal of information about the direction of future returns.
User avatar
abuss368
Posts: 27850
Joined: Mon Aug 03, 2009 2:33 pm
Location: Where the water is warm, the drinks are cold, and I don't know the names of the players!
Contact:

Re: Past returns vs future returns

Post by abuss368 »

Ferdinand2014 wrote: Tue Oct 19, 2021 4:52 am
No, I am not doing anything different based on my intuition regarding the current value of the stock market. I buy every week and keep enough cash so I do not need to sell unless of my choosing.
Hi Jon -

Your strategy is so simple and does not have a lot of moving parts. That keeps the risk of tinkering and making unfortunate moves at bay. I like your S&P 500 and Treasuries portfolio.

I recently read Rick Ferri mention most investors only need 4 -5 funds tops. It feels as if Rick has simplified as well over the years.

Thanks.
Tony
John C. Bogle: “Simplicity is the master key to financial success."
alex_686
Posts: 13320
Joined: Mon Feb 09, 2015 1:39 pm

Re: Past returns vs future returns

Post by alex_686 »

BJJ_GUY wrote: Thu Oct 21, 2021 3:29 pm 3.) Now if you take valuations, this is different because we're adding new variables into the analysis. I prefer to use other models that have a more reliable/less cyclical denominator, so P/E isn't great. Rather than go into a totally different discussion, I'll use P/E in writing, but it isn't a great ratio. Point is, you are absolutely correct that the only way to square the circle is to understand that a valuation ratio on one hand of an equation = returns. This is simple arithmetic, which I believe was the point you were actually making as well. What I'm trying to say is that when the valuations are higher or lower than the historical average we can determine what future return would be needed for the valuation to move back somewhere close to the historical average... I'm very specifically avoiding using the word "predict" because the connotation is typically associated with a strict time frame for that return to be realized, and that isn't how it works. The return that solves for the valuation multiple (and of course accounting for some other variables, but again not for this conversation), doesn't predict precisely when.
I would be curious what metrics you think are better than earnings and P/E. I will gladly conceded that there are issues and nuances with them, however I think that have more informational value than other values.

But let me pick apart the sentence that I underlined.

I would say that if valuations are moving in your favor you will make gains, if valuations move against you you lose value.

You however are choosing "historical average", which I strongly believe is a arbitrary and meaningless mark. If there is no regression to the mean then the mean has no value. And I don't believe there is a regression to the mean. I have dug into the theory and the historical data And you can dig into the theory or the historical data and it is just not there. What goes up does not have to come down. Historically high valuations are just as likely to go higher, go sidewise or go down as historically low values are. There is no information here to data mine.

For a bit more context I work with a team were we have to predict the future. We do a fairly decent job. But we have to use Jump Diffusion Models and other esoterica mathematical tools which don't require a regression to the mean.
Former brokerage operations & mutual fund accountant. I hate risk, which is why I study and embrace it.
luckyducky99
Posts: 415
Joined: Sun Dec 15, 2019 6:47 pm

Re: Past returns vs future returns

Post by luckyducky99 »

alex_686 wrote: Fri Oct 22, 2021 11:22 am
BJJ_GUY wrote: Thu Oct 21, 2021 3:29 pm What I'm trying to say is that when the valuations are higher or lower than the historical average we can determine what future return would be needed for the valuation to move back somewhere close to the historical average...

You however are choosing "historical average", which I strongly believe is a arbitrary and meaningless mark. If there is no regression to the mean then the mean has no value. And I don't believe there is a regression to the mean.
I buy this. Does anyone really think that monetary policy and accounting practices and values of physical vs intangible assents of 100 years ago or 50 years ago represent a meaningful model of the current economic situation such that we should expect things to trend that way? Historical averages based on those features just aren't meaningful today.
User avatar
nisiprius
Advisory Board
Posts: 52211
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Past returns vs future returns

Post by nisiprius »

In 1937, Cowles and Jones published this chart in Econometrica:

Image

It showed that "sequences" (runs of stock price changes going in the same direction) occurred more often than chance over periods of 20 minutes to two years, while "reversals" predominated over periods of 3 to 10 years. Nowadays we call these phenomena "momentum" and "mean reversion."

They wrote that
This evidence of structure in stock prices suggests alluring possibilities in the way of forecasting. In fact, many professional speculators, including in particular exponents of the so-called "Dow Theory" widely publicized by popular financial journals, have adopted systems based in the main on the principle that it is advantageous to swim with the tide.
They then spent about eight pages describing tests of, essentially, a market timing strategy, and concluded that
This type of forecasting could not be employed by speculators with any assurance of consistent or large profits. On the other hand, the significant excess of sequences over reversals for all units from 20 minutes up to 6 months, with the exception of units of 2 weeks and 3 weeks mentioned previously, represents conclusive evidence of structure in stock prices.
And that's pretty much where we are today. Momentum exists. Mean reversion exists. It is hard to believe that they can't be exploited to beat the market, but the evidence--particularly from the actual records of tactical asset allocation mutual funds--is that it is not easy to do.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
Zeno
Posts: 1042
Joined: Wed Sep 12, 2018 10:44 am

Re: Past returns vs future returns

Post by Zeno »

“I have approximate answers, and possible beliefs, and different degrees of certainty about different things, but I'm not absolutely sure of anything. There are many things I don't know anything about. It doesn't frighten me.”

Richard Feynman
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

Re: Past returns vs future returns

Post by secondopinion »

nisiprius wrote: Sun Oct 24, 2021 7:13 pm In 1937, Cowles and Jones published this chart in Econometrica:

Image

It showed that "sequences" (runs of stock price changes going in the same direction) occurred more often than chance over periods of 20 minutes to two years, while "reversals" predominated over periods of 3 to 10 years. Nowadays we call these phenomena "momentum" and "mean reversion."

They wrote that
This evidence of structure in stock prices suggests alluring possibilities in the way of forecasting. In fact, many professional speculators, including in particular exponents of the so-called "Dow Theory" widely publicized by popular financial journals, have adopted systems based in the main on the principle that it is advantageous to swim with the tide.
They then spent about eight pages describing tests of, essentially, a market timing strategy, and concluded that
This type of forecasting could not be employed by speculators with any assurance of consistent or large profits. On the other hand, the significant excess of sequences over reversals for all units from 20 minutes up to 6 months, with the exception of units of 2 weeks and 3 weeks mentioned previously, represents conclusive evidence of structure in stock prices.
And that's pretty much where we are today. Momentum exists. Mean reversion exists. It is hard to believe that they can't be exploited to beat the market, but the evidence--particularly from the actual records of tactical asset allocation mutual funds--is that it is not easy to do.
When people have to wait 8 or so years for a reversal, almost no investor has that kind of patience. And we should invest it for the time being.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
User avatar
HomerJ
Posts: 21281
Joined: Fri Jun 06, 2008 12:50 pm

Re: Past returns vs future returns

Post by HomerJ »

WyomingFIRE wrote: Sun Oct 24, 2021 7:19 pm “I have approximate answers, and possible beliefs, and different degrees of certainty about different things, but I'm not absolutely sure of anything. There are many things I don't know anything about. It doesn't frighten me.”

Richard Feynman
Wow, great quote.

I saw a rock once that had a carving in it stating "Don't believe everything you think".

:beer
"The best tools available to us are shovels, not scalpels. Don't get carried away." - vanBogle59
secondopinion
Posts: 6011
Joined: Wed Dec 02, 2020 12:18 pm

Re: Past returns vs future returns

Post by secondopinion »

HomerJ wrote: Mon Oct 25, 2021 10:38 am
WyomingFIRE wrote: Sun Oct 24, 2021 7:19 pm “I have approximate answers, and possible beliefs, and different degrees of certainty about different things, but I'm not absolutely sure of anything. There are many things I don't know anything about. It doesn't frighten me.”

Richard Feynman
Wow, great quote.

I saw a rock once that had a carving in it stating "Don't believe everything you think".

:beer
How about the Hamilton's etch about quaternions? Another great one.
Passive investing: not about making big bucks but making profits. Active investing: not about beating the market but meeting goals. Speculation: not about timing the market but taking profitable risks.
BJJ_GUY
Posts: 882
Joined: Wed Mar 13, 2019 7:45 am

Re: Past returns vs future returns

Post by BJJ_GUY »

alex_686 wrote: Fri Oct 22, 2021 11:22 am
BJJ_GUY wrote: Thu Oct 21, 2021 3:29 pm 3.) Now if you take valuations, this is different because we're adding new variables into the analysis. I prefer to use other models that have a more reliable/less cyclical denominator, so P/E isn't great. Rather than go into a totally different discussion, I'll use P/E in writing, but it isn't a great ratio. Point is, you are absolutely correct that the only way to square the circle is to understand that a valuation ratio on one hand of an equation = returns. This is simple arithmetic, which I believe was the point you were actually making as well. What I'm trying to say is that when the valuations are higher or lower than the historical average we can determine what future return would be needed for the valuation to move back somewhere close to the historical average... I'm very specifically avoiding using the word "predict" because the connotation is typically associated with a strict time frame for that return to be realized, and that isn't how it works. The return that solves for the valuation multiple (and of course accounting for some other variables, but again not for this conversation), doesn't predict precisely when.
I would be curious what metrics you think are better than earnings and P/E. I will gladly conceded that there are issues and nuances with them, however I think that have more informational value than other values.

But let me pick apart the sentence that I underlined.

I would say that if valuations are moving in your favor you will make gains, if valuations move against you you lose value.

You however are choosing "historical average", which I strongly believe is a arbitrary and meaningless mark. If there is no regression to the mean then the mean has no value. And I don't believe there is a regression to the mean. I have dug into the theory and the historical data And you can dig into the theory or the historical data and it is just not there. What goes up does not have to come down. Historically high valuations are just as likely to go higher, go sidewise or go down as historically low values are. There is no information here to data mine.

For a bit more context I work with a team were we have to predict the future. We do a fairly decent job. But we have to use Jump Diffusion Models and other esoterica mathematical tools which don't require a regression to the mean.
I would be curious what metrics you think are better than earnings and P/E. I will gladly conceded that there are issues and nuances with them, however I think that have more informational value than other values.
I think there are a couple of valuation methods that are fundamentally more sound, and also statistically stronger than P/E or CAPE. The following three are superior:
1.) US Non-financial market cap / GVA (adjusted to estimate foreign revenues)
2.) Price-to-revenue
3.) Market cap / GDP

With the valuation the goal is for the denominator to be an estimate of the long-term cash flows derived from equity ownership. Earnings is too cyclical, even when smoothed out over 10 years. A version of CAPE that is adjusted for profit-margins is an upgrade, however, CAPE adjusted for profit-margins (compared to historical average) is basically just getting back to price-to-revenue.
Historically high valuations are just as likely to go higher, go sidewise or go down as historically low values are. There is no information here to data mine.
I'm not sure what you mean here. Or, rather, I'm not sure why we can have such different information. Over what time period are you basing this statement? If you plot any of the three valuation ratios I listed above against the subsequent 12 year returns, there is >0.91 correlation.
Post Reply