The Shortcomings of CAPE

Aswath Damodaran’s, noted value investor and professor of finance at NYU, latest blog post centers on the shortcomings of CAPE. Since there’s been some talk about CAPE’s historically high levels lately on this board, I thought I would share it here:

Robert Shiller has been a force in finance, forcing us to look at the consequences of investor behavior and chronicling the consequences of “irrational exuberance”. His work with Karl Case in developing a real estate index that is now widely followed has introduced discipline and accountability into real estate investing and his historical data series on stocks, which he so generously shares with us, is invaluable. You can almost see the “but” coming and I will not disappoint you. Of all of his creations, I find CAPE to be not only the least compelling but also potentially the most dangerous, in terms of how often it can lead investors astray. So, at the risk of angering those of you who are CAPE followers, here is my case against putting too much faith in this measure, with much of it representing updates of what my post from two years ago.

Damodaran goes on to provide a detailed explanation as to why he believes that: 1) The CAPE is not that informative; 2) The CAPE is not that predictive; 3) Investing is relative, not absolute; and 4) Its cash flow, not earnings that drives stocks. I know this isn’t necessarily a value-focused board, but it’s still a great read. You can read the entire post at http://aswathdamodaran.blogspot.com/2016/08/superman-and-sto…

Matt
MasterCard (MA), PayPal (PYPL), and Verizon (VZ) Ticker Guide
See all my holdings at http://my.fool.com/profile/CMFCochrane/info.aspx

6 Likes

Neil deGrasse Tyson, the Director of the Hayden Planetarium in New York City, is one of my favorite scientists. One comment he made about growth curves sticks in my mind. If you plot growth on a linear scale chart you will always get a hockey stick curve. Presenters love to scare audiences with “unsustainable” growth by showing them a liner chart.

The “Superman” chart in Damodaran’s article is a linear chart and it shows the typical hockey stick curve. If the chart were longer the curve would be even more noticeable.

http://softwaretimes.com/pics/superman-chart.png

Compare that to the Klein S&P 500 chart which has both a semi-logarithmic* chart at the top followed by a linear chart:

http://invest.kleinnet.com/bmw1/stats40/^GSPC.html

What the semi-log chart does is to paint steady growth as a straight line with the slope indicating the rate of growth, the higher the rate, the steeper the line. One of the first things I do is to check which kind of chart the presenter is using. Some, like a TED talk I saw recently, clearly indicated that the presenter didn’t know about log charts. The scare mongers invariably use linear charts with big hockey stick curves.

Now take a look at the Klein S&P 500 chart again, the semi-log chart is not scary at all, it says the index is undervalued! A 50% growth from a 100 base is 50. A 50% growth from a 1000 base is 500. The semi-log chart makes them the same size because they are both 50% growth rate.

This is quite apart from the data going into the CAPE-Shiller index which is what Damodaran’s article discusses.

Denny Schlesinger

  • Semi-log means a log Y axis and a linear X axis.
11 Likes

Presenters love to scare audiences with “unsustainable” growth by showing them a liner chart.
Now take a look at the Klein S&P 500 chart again, the semi-log chart is not scary at all, it says the index is undervalued!

No one who has done any research; for instance, comparing valuations with trailing and forward earnings, and historical norms, believes the index is undervalued.

Your comparison shows—inadvertently perhaps—just how wrong a semi-log chart can be. If a linear chart scares people, then, certainly in this instance, a semi-log chart lulls them into a comfy sense of well being and optimism…

kelbon

7 Likes

Your comparison shows just how wrong a semi-log chart can be. If a linear chart scares people, then, certainly in this instance, a semi-log chart lulls them into a comfy sense of well being and optimism…

Kelbon, perhaps you simply don’t understand what a semi-log scale is:

It simply shows gains and losses proportionally. For example, on a linear scale a 10% gain from $20 would be $2, but ten years later, when the index is at $200 a 10% gain of $20 would look huge by comparison, even though it’s the same 10%.

On a semi-log scale, 10% always looks like 10% and the size of the movement is the same.

That’s all there is to it, and of course semi-log scales ALWAYS give a more accurate view.

Sasu

15 Likes

Sasu? Still need to get a cup of coffee??

I really just wanted to say that I wish I could have recommended your semi-log paper posting twice. It is such common sense and an example of how people start to build biases against that which they don’t understand.
Great posting.

2 Likes

Kelbon, perhaps you simply don’t understand what a semi-log scale is[?]

I see why you might think so. But, I do understand what a semi-log scale is.

My focus was in response to:
Now take a look at the Klein S&P 500 chart again, the semi-log chart is not scary at all, it says the index is undervalued!

My point was not whether the chart is linear or semi-log, but rather what the chart purports to show; that is, the index is undervalued. It doesn’t matter what kind of chart if the data is meaningless.

Klein price-CAGR charts are a vestige of the BMW Method, the premise of which was that price growth reverts to a mean and this can be used like a divining rod to locate good investments. In this case, the S&P “is undervalued!”

While most other proponents of the BMW Method have long faded away, Denny persists in believing that price-CAGR charts show if an investment, or an index in this instance, is overprice, fairly priced, or undervalued. Which is all fine, but I suggest taking it with a grain of salt.

Prices don’t revert to a mean, in and of themselves, and these charts, though perhaps of historical interest, are not far removed from reading tea leaves as a signpost to the future.

kelbon

16 Likes

Now take a look at the Klein S&P 500 chart again, the semi-log chart is not scary at all, it says the index is undervalued!

That’s using the 40-year chart. If you use the 30-year chart, it says the index is fairly valued! If you use the 20-year chart, it says it’s overvalued!

40 year: http://invest.kleinnet.com/bmw1/stats40/%5EGSPC.html
30 year: http://invest.kleinnet.com/bmw1/stats30/%5EGSPC.html
20 year: http://invest.kleinnet.com/bmw1/stats20/%5EGSPC.html

Okay, you say, well the 40-year chart has more data points, more history. The longer the history the more accurate the results, right? Based on forty years of data the index is definitely undervalued!

The Klein charts that Denny refers to don’t go back more than forty years for the index. However, you can download the data yourself and do the same calculations. If you do that, you’ll find the 50-year and 60-year charts say the the index is OVERVALUED!

What’s going on here!?

…and of course semi-log scales ALWAYS give a more accurate view.

As investors we have to be careful about confusing accuracy with insight. The problem with the charts Denny is using is that they are based on comparison to a mean that shifts depending on the time slice you pick. In other words, the building’s foundation is built on sand. As a result, although they appear to be more accurate, they have limited or nil predictive value. In fact, as Kelbon rightly pointed out, they can lull you into a false sense of security.

Ears

11 Likes

Klein price-CAGR charts are a vestige of the BMW Method, the premise of which was that price growth reverts to a mean and this can be used like a divining rod to locate good investments. In this case, the S&P “is undervalued!”

While most other proponents of the BMW Method have long faded away, Denny persists in believing that price-CAGR charts show if an investment, or an index in this instance, is overprice, fairly priced, or undervalued. Which is all fine, but I suggest taking it with a grain of salt.

Kelbon has been on a search and destroy mission against the BMW Method for at least a decade. You’ll have to ask him why he hates it so much. It could have something to do with the BMW Conferences where he was not invited to attend for free.

Fool BuildMWell, the inventor of the Method was constantly looking to “push the envelope.” He started by looking at the historic DJI index and noticed that time and again it would reach a bottom and bounce back. He joined these low points which, to him, indicated the point below which the index would not go. He called this line the “low CAGR.” The first “push of the envelope” was to look at the individual 30 Dow Jones stocks. Jim claimed to have made good money using the method and kept pushing the envelope and encouraged others to do the same. One day he pushed too far.

Jim did his charts by hand. Mike Klein (Fool mklein9) pushed the envelope by creating computer generated charts on a weekly basis. There are some differences but Mike’s -2RMS line is close enough to Jim’s low CAGR line to make them comparable. Mike did some back testing and showed that RF was a better predictor of outcomes than RMS. Since the charts are generated by a computer program they keep churning out on a weekly basis. There is a small group of people who still use them.

The magic of markets is that they squeeze out any unnecessary profits. This is what makes truly free markets so effective in promoting economic growth. I say “truly free” because there are many ways to encumber markets as with monopolies, cartels, social programs, trade agreements, and other ways in which people try to twist the market in their favor. The stock market is no different. Any time you find a way to beat the market the market will try to find a way to stop you. Don’t ask me how it happens but it happens. That’s why bubbles form and burst.

In the case of the BMW Method I know exactly why it imploded, Jim pushed the envelope one step too far. Jim became convinced that the Method was infallible, a sure sign of Hubris. When financial entities started to drop in 2008 Jim started to buy at low CARG but many didn’t bounce back, instead some went bankrupt. By 2008 I was familiar with Nassim Nicholas Taleb’s work on Black Swans and his opinion that the financial industry was “bad black swan prone” and hence to be avoided. I made a couple of posts to that effect but they were ignored in the heat of battle.

While I was working at IBM in the late 1960s I had the opportunity to learn about their IMPACT inventory management system. For this discussion one concept is relevant. Items in inventory had to be classified as cyclical or not cyclical. This was done by creating consumption charts to see if one could detect some kind of seasonality. But it was not enough to see a hump or a dip at certain times, one had to find a logical explanation to eliminate chance occurrences. This led me, in time, to coin the phrase “It is not enough to know what a chart is saying, one has to find out why it is saying whatever it is saying.”

While most other proponents of the BMW Method have long faded away, Denny persists in believing that price-CAGR charts show if an investment, or an index in this instance, is overprice, fairly priced, or undervalued. Which is all fine, but I suggest taking it with a grain of salt.

Prices don’t revert to a mean, in and of themselves, and these charts, though perhaps of historical interest, are not far removed from reading tea leaves as a signpost to the future.

“It is not enough to know what a chart is saying, one has to find out why it is saying whatever it is saying” is the grain of salt that kelbon recommends. I have said it on multiple occasions on multiple Foolish discussion boards. It ruins his argument so he ignores it.

Is the BMW Method a failure? Not if applied where it works. Taleb forewarned that the financial industry was not a good place to use it. My own take is that it works on companies that have a long history of successfully doing whatever it is they do, companies like 3M. BTW, 3M’s original mining operations were a failure but their glue based products have been a huge success. For me one of the shortcomings of the BMW method was the long wait until a good company’s price crashed for no good reason, my attention span is way too short to play this game and I didn’t have an alert system to keep me informed. Yet the concept of CAGR is very powerful, my quest was how to use it without having to wait for these extreme drops in price. The BMW Method was instrumental in developing my current market strategy.

So yes, look at charts with a pinch of salt: “Find out why the chart is saying whatever it is saying.” Don’t throw it out with the bathwater.

Denny Schlesinger

8 Likes

What’s going on here!?

That you are dealing with a complex system, not with an exact science like astrophysics which allowed Edmond Halley to predict in the 18th century the return of the comet in 76 years. In finance you can’t predict the price from one day to the next. In finance the best we can do is to apply probability, the law of large numbers, and a few other mathematical tricks to try to beat the market.

Once you get used to that idea you can start dealing with the uncertainty. If you treat charts or DCF calculations as exact representations of reality you are in for a big surprise.

Denny Schlesinger

2 Likes

My dissertation adviser was W.W. Howells, who probably did more than anyone else to bring sophisticated mathematical analysis tools to biological anthropology. He drummed into his students that one had nothing from an analysis until on had made biological sense of the numerical result. This seems very similar in principle to Denny’s principle of needing to understand the result.

What I am not sure about is whether the domains are comparable. If one is studying fossil hominids, for example, something actually happened and it is the goal of one’s analysis to figure out what that was. And, one is focused on the past, i.e., what happened, not what will happen.

If one is studying the stock price of companies in the market, though, there is a huge amount of idiosyncratic effect where what happened to any one specific company was dependent on the very specific characteristics and context and history of that company. This makes it difficult to assign causality with any confidence. Moreover, one is not really interested in the past … the whole point is predicting the future. In such a context, it is all too easy for a pretty fit to past behavior to have little or no predictive value for the future.

Kelbon has been on a search and destroy mission against the BMW Method for at least a decade. You’ll have to ask him why he hates it so much.

If you define a search and destroy mission as disagreeing with a methodology and backed that disagreement with evidence and logic, then guilty as charged. Honestly, the only reason it concerned me was, at one time, there was so much hype about its infallibility people were being duped into betting real money, and losing real money, on a pipe dream. The chief proponent of this misguided “method” slinked away into the night long ago, but Denny is the old warrior holed up in his hideaway in the mountains, firing an occasional salvo, oblivious that the war is over and almost everyone else has moved on.

[…]It could have something to do with the BMW Conferences where he was not invited to attend for free.

This is laughable, and pathetic.

“It is not enough to know what a chart is saying, one has to find out why it is saying whatever it is saying” is the grain of salt that kelbon recommends. I have said it on multiple occasions on multiple Foolish discussion boards. It ruins his argument so he ignores it.

I completely agree with the notion that data should be seen in context. There’s very smart people who hone in on some calculation, or historical chart, or another and make their case on it. For example, there’s much talk from some quarters how overvalued the stock market is, but they ignore the historically low interest rate environment as if this has no bearing on the price of stocks at all; as if it doesn’t exist.

I have no problem with people incorporating whatever voodoo or methodology into their decision making process. My only complaint is when someone trots out a price-CAGR chart and makes a definitive statement like, “This stock is undervalued, or this index is undervalued, because this chart says so!” This is something Denny does now and again, and sometimes I call him out on it, which he doesn’t like. I think this—where it all started—qualifies: Now take a look at the Klein S&P 500 chart again, the semi-log chart is not scary at all, it says the index is undervalued!

Denny occasionally talks the talk, but he rarely walks the walk.

kelbon

5 Likes

<The Klein charts that Denny refers to don’t go back more than forty years for the index. However, you can download the data yourself and do the same calculations. If you do that, you’ll find the 50-year and 60-year charts say the the index is OVERVALUED!

What’s going on here!?

What’s going on here is if you contort yourself enough you can create the appearance of a theory fitting the facts. Any methodology that has to be bent, stretched, redefined, and contorted to fit facts isn’t worthwhile.

kelbon

1 Like

I have made plenty of money doing just that but I’m a trial lawyer.

Htownrich

3 Likes

I have made plenty of money doing just that but I’m a trial lawyer.

That’s funny; unless you really are a trial lawyer, then it’s not.

1 Like