S and P 500 Graph - Scary?

[willn]

I attached a screen grab of the price of the S and P 500 since 1975. Does this seem pretty wild? Note the last two peaks compared to the current one. One was just before the dot com bubble, one was just before the mortgage crises.  Are there differences that make the high valuations we are seeing more lasting?  Why shouldn't we expect a crash resembling the last two, which is a steep decline for two years or so before a grinding recovery.  At least in the last recovery hiring resumed.

I know this is a black art and the only certainty is what goes up comes down, and it doesn't affect my strategy for continuous monthly investing, but it is a striking data set I think!

[TheDude]

I think  you should look at the PE ration. Back on the previous peak it was nuts. Its currently a higher than I would like but imho this is not 2008 again.

[Reepekg]

Um, looks to me like we are today exactly where you'd expect to be if you did a regression fit of the points to a curve.

Said another way, look how crazy high the market was at the end of 90'! Those two peaks in 87' and 89' before the crashes set in means you'd have to have nerves of steel to be in stocks at the next peak in early 91'... oh wait, in the long view those look like speed bumps now. Nevermind.

[arebelspy]

Um, looks to me like we are today exactly where you'd expect to be if you did a regression fit of the points to a curve.

This is what I saw too.  Take a look at that trend line from the 80s and early 90s, and extrapolate it as an exponential curve to now... we're about where we should be at, not overvalued.

[Integrate]

Um, looks to me like we are today exactly where you'd expect to be if you did a regression fit of the points to a curve.

This is what I saw too.  Take a look at that trend line from the 80s and early 90s, and extrapolate it as an exponential curve to now... we're about where we should be at, not overvalued.

Of course you could also argue a pretty good linear fit for a bit more than half of the current value. That's a silly way to do valuations.

For example if growth had gone from 10% per year in the 70's and 80's to -10% per year in the 2000's, would you still feel comfortable extrapolating a exponential curve? What if inflation had moved to 5000% per year? Then we'd be greatly undervalued.

I'm not trying to make a point either way on the value of the market (if I knew that I'd be retired), but I am simply making a point that extrapolating a line that's convenient ignores all the fundamentals. Indeed, if we are going to argue for an exponential curve we'll all be retiring soon.

[willn]

Of course you could also argue a pretty good linear fit for a bit more than half of the current value. That's a silly way to do valuations.

For example if growth had gone from 10% per year in the 70's and 80's to -10% per year in the 2000's, would you still feel comfortable extrapolating a exponential curve? What if inflation had moved to 5000% per year? Then we'd be greatly undervalued.

I'm not trying to make a point either way on the value of the market (if I knew that I'd be retired), but I am simply making a point that extrapolating a line that's convenient ignores all the fundamentals. Indeed, if we are going to argue for an exponential curve we'll all be retiring soon.

Right.

So we know that there was a bubble on the value of the market due to disruptive technologies in the late 90's.  We know there was a bubble on housing prices in 2007.   My thinking is that with the Fed pumping 80+ Billion/month into the economy, we have a bubble in stock prices now.  Is it supported by steadily increasing employment? A new disruptive technology leading to higher productivity?  More natural resources?   I don't know.  I think its reasonable to assume we won't have another 5 years of increasing stock prices. So, how to play the next five years, or do you bother making any strategic changes based on this?

[matchewed]

Of course you could also argue a pretty good linear fit for a bit more than half of the current value. That's a silly way to do valuations.

For example if growth had gone from 10% per year in the 70's and 80's to -10% per year in the 2000's, would you still feel comfortable extrapolating a exponential curve? What if inflation had moved to 5000% per year? Then we'd be greatly undervalued.

I'm not trying to make a point either way on the value of the market (if I knew that I'd be retired), but I am simply making a point that extrapolating a line that's convenient ignores all the fundamentals. Indeed, if we are going to argue for an exponential curve we'll all be retiring soon.

Right.

So we know that there was a bubble on the value of the market due to disruptive technologies in the late 90's.  We know there was a bubble on housing prices in 2007.   My thinking is that with the Fed pumping 80+ Billion/month into the economy, we have a bubble in stock prices now.  Is it supported by steadily increasing employment? A new disruptive technology leading to higher productivity?  More natural resources?   I don't know.  I think its reasonable to assume we won't have another 5 years of increasing stock prices. So, how to play the next five years, or do you bother making any strategic changes based on this?

Your crystal ball is as cloudy as any other crystal ball. If you want to make a play on speculation be my guest, if you want a safe(r) bet just set up your automatic investments, enjoy the bumpy ride, and get on with the other parts of life. Don't try to anticipate what the market will do in five years (let alone one month), also - http://jlcollinsnh.com/2013/01/04/how-to-be-a-stock-market-guru-and-get-on-msnbc/

[KingCoin]

Indeed, if we are going to argue for an exponential curve we'll all be retiring soon.

Compounding returns are in fact exponential, and it is amazing, and yes, you  can retire on them (though not too soon). This is why most personal finance books highlight the "magic" of compounding returns in the first two chapters and the importance of starting early.

OP, is it your contention that the S&P500 should never go above 1500? That the economic gods become displeased when the S&P500 has the audacity to flirt with 1500? That the number 1500 strikes fear into the heart of investors? That corporations in the S&P500 are only capable of earning enough money to support a valuation of 1500?

If the second crash happened when the S&P500 hit 1400 would you be less worried? If it crashed when the S&P500 hit 1600 would you be more worried? Is the very tenuous "pattern" that two crashed happened at 1500 somehow important or telling? Does the fact that the S&P500 has risen to 1700 mean the curse is broken?

I could go on, but you get the drift. Nothing "wild" here.

To your last comment, the world GDP continues to grow as humans become more and more productive, and industrial technology reaches more and more people. S&P500 companies have an increasing number of affluent customers to sell their products and services to. You really don't need anything crazy or disruptive to support earnings growth.

I think its reasonable to assume we won't have another 5 years of increasing stock prices. So, how to play the next five years, or do you bother making any strategic changes based on this?

So you're basing this opinion almost entirely on the fact that the S&P500 fell when it hit 1500 twice? Can you really argue that the world is in a different place that 1982 or 1990? Every era has its reasons for bullishness and bearishness. Sometimes bearish sentiment is the best reason to get long. Sometimes, when it looks like clear sailing it's the most dangerous time to invest. I just don't see any arguments that corporations can't continue to grow in this environment. Is it not entirely possible that the corporations can continue to grow, even as the fed tapers over the next couple years?

[matchewed]

Slightly relevant Wade Pfau post - http://wpfau.blogspot.com/2013/11/william-bernstein-on-deep-risk-shallow.html

He talks about how different asset classes contain different risks. The S&P 500 doesn't actually contain much long term deep risk aside from deflation and confiscation. What you see in your graph is what is termed shallow risk in the post. Shallow risk is just day to day/year to year churn. Regardless of the causes it is no real cause for fear as it will happen several times in your life but will also recover.

[brewer12345]

*Snore*

Wake me when we either drop 300 points (buying opportunity) or crest 2000.

[Mississippi Mudstache]

Indeed, if we are going to argue for an exponential curve we'll all be retiring soon.

Compounding returns are in fact exponential, and it is amazing, and yes, you  can retire on them (though not too soon). This is why most personal finance books highlight the "magic" of compounding returns in the first two chapters and the importance of starting early.

Yes, if you don't believe that your investment returns will fit an exponential curve, then I would strongly suggest that you don't retire. That's pretty much what we're all counting on. It will not, however, be a smooth curve.

[dadof4]

Here is the same graph on a logarithmic scale. Not so scary anymore, huh?

[El Gringo]

Here is the same graph on a logarithmic scale. Not so scary anymore, huh?

Can you explain the difference?

[matchewed]

Here is the same graph on a logarithmic scale. Not so scary anymore, huh?

Can you explain the difference?

The logarithmic scale takes into account that the S&P chart is not a chart of a linear relationship but of an exponential relationship (overly simple but good enough for this circumstance). The logarithmic scale shows a more realistic portrayal of the performance of the S&P 500 as it "smooths" the exponential growth into a more visible format. This isn't a manipulation of the data but an adjustment of perspective really.

[dadof4]

Thin

Can you explain the difference?

In a linear graph (like in the OP), even if growth is constant, it will look like the curve is getting steeper and steeper. That is because a 7% growth year over year at 400 is much lower in absolute terms than identical 7% growth at 800. A logarithmic graph corrects this perception.

[arebelspy]

Here is the same graph on a logarithmic scale. Not so scary anymore, huh?

I have found myself agreeing with many of your posts over the last few days.

Well done, and thank you.

[AdrianM]

If you are familiar with Harry Dent's work with demographics and peak spending ages.
Would suggest that most of the run up in stock prices can be attributed to boomer spending.

The quick and dirty explantion
http://www.investopedia.com/terms/b/baby-boom-age-wave.asp

The book
http://www.amazon.com/The-Great-Depression-Ahead-Prosper/dp/141658899X

[Integrate]

Indeed, if we are going to argue for an exponential curve we'll all be retiring soon.

Compounding returns are in fact exponential, and it is amazing, and yes, you  can retire on them (though not too soon). This is why most personal finance books highlight the "magic" of compounding returns in the first two chapters and the importance of starting early.

I agree that compounding interest is exponential. I don't agree that means the S&P 500 will necessarily follow an exponential curve. The S&P 500 index has no inherent reason to return 7% annually. It's easy to imagine lots of situations where it doesn't return anything over a 10 year period (the likelihood of those situations is a separate matter). I'm betting this isn't the case, as I'm guessing most of you are, given my investments are in stocks.

My point is simply that looking at the past and saying this curve fits so we're all good for the future is not a very convincing argument. Regression analysis is great for finding correlation, okay at best for finding causation and not very good at all for forecasting.

[matchewed]

Indeed, if we are going to argue for an exponential curve we'll all be retiring soon.

Compounding returns are in fact exponential, and it is amazing, and yes, you  can retire on them (though not too soon). This is why most personal finance books highlight the "magic" of compounding returns in the first two chapters and the importance of starting early.

I agree that compounding interest is exponential. I don't agree that means the S&P 500 will necessarily follow an exponential curve. The S&P 500 index has no inherent reason to return 7% annually. It's easy to imagine lots of situations where it doesn't return anything over a 10 year period (the likelihood of those situations is a separate matter). I'm betting this isn't the case, as I'm guessing most of you are, given my investments are in stocks.

My point is simply that looking at the past and saying this curve fits so we're all good for the future is not a very convincing argument. Regression analysis is great for finding correlation, okay at best for finding causation and not very good at all for forecasting.

Fair enough, that is why everyone states the line of past performance is not indicative of future returns. We are all just guessing, but some guesses are better than others. Guessing that the stock market will go up in the long run is a pretty good guess. It's not only regression analysis but a belief that our economy (country and global) and the companies which comprise it will continue to grow and provide/create goods and services people want/need.

Regardless of that the S&P 500 as it stands now is an exponential graph and compounding returns whether at 1% or 8% still will exhibit exponential behavior, no one is saying it's a smooth ride as it is often said those returns are on average over a period of time but looking at the S&P500 in 60 year time frames it is a (rough) exponential graph.

[deltaecho]

there is an interesting article here: http://www.marketwatch.com/story/you-really-can-time-the-stock-market-2013-11-04
about predicting some long term trends in the market.

[Integrate]

there is an interesting article here: http://www.marketwatch.com/story/you-really-can-time-the-stock-market-2013-11-04
about predicting some long term trends in the market.

Bunch of silly things in this article. For one he seems to be arguing that if you invested ALL YOUR MONEY AT ONCE and then no money at future point in times, there exists times where you could lose money or make a poor return. To that I say, no shit. Talk about cherry picking.

Most investors tend to invest over time. Some of your money may be invested at a poor time, but most of it won't. You'll earn money in the long run.

Let's say I wanted to use his q ratio or Shiller PE ratio as a criteria for investing. Let's say to maximize my return I only invest when these values are below the mean. If I had been doing this since 1990, the only times I would have put money in the stock market were 1990 and 2008. I'm pretty sure I would have missed out on some nice returns over that 23 year time frame. The Shiller PE's correlation is ~.6 with the S&P 500. This means 40% of the time they are not moving together. If one thinks they can predict the future with that, be my guest. I'll sit that one out, though.

Lastly, the author seems to confuse a random distribution and a random walk. He shows a graph of returns and says the "waves are ... clear" and that "these results are not random." He makes the argument that "First, let’s demolish the myth that the stock market produces entirely “random” returns—that some years it’s up, other years it’s down, that over time it just goes up, and no one can predict anything in advance." He presents his chart:


Here is a picture of a random walk:


I'll let you decide if it looks like a random walk or not.

[mm31]

The market can be incredibly distracting. As long as you're in the accumulation phase, I wouldn't worry about it too much