Zen Lessons in Market Analysis

Weekly Market Comment Archive

Investment Research & Insight Page

Zen Lessons in Market Analysis John P. Hussman, Ph.D. All rights reserved and actively enforced. Reprint Policy

“The best way of preparing for the future is to take good care of the present, because we know that if the present is made up of the past, then the future will be made up of the present. All we need to be responsible for is the present moment. Only the present is within our reach. To care for the present is to care for the future.”

Thich Nhat Hanh

This week's comment is dedicated to my dear friend Thich Nhat Hanh, a Vietnamese Buddhist monk who was born on October 11, 1926, having been born previously in January of that same year, and twice again about 25 years earlier, not to mention countless other times through his ancestors, teachers, and other non-Thich Nhat Hanh elements. Thay (the Vietnamese word for “teacher”) would simplify this by saying that today is his eighty-third “continuation day,” because to say it is his birthday is not very accurate.

If the quote at the top of this page looks somewhat familiar to our long-term shareholders, it may be because the practice of tending to the present moment – responding to prevailing conditions rather than relying on forecasts – is central to our investment discipline.

Focusing on the present moment doesn't imply ignoring the past or failing to consider the future. It's clear, for example, that we put a great deal of attention on estimating future cash flows and discounting them appropriately in order to evaluate whether various investments are priced to deliver satisfactory long-term returns. We certainly devote our attention to macroeconomic pressures and latent risks that threaten to become full-blown crises later. Still, we rarely make near term forecasts. Nor do we answer surveys like “where do you think the S&P 500 will be at year-end?” – a question that falls entirely outside of our way of thinking – like asking Columbus what sort of trees he thinks are planted along the edge of the Earth. The reason we avoid forecasts, very simply, is that they are not required, and that they can be a hindrance.

Expectations

One of the major debates among investors is between buy-and-hold investing and market timing. Think of the market as a big hat that has both red and green marbles in it, red corresponding to declines, and green corresponding to advances. The buy-and-hold investor essentially believes that it is impossible to predict which color marble will be drawn next, but that on average the marbles will be green. So the buy-and-hold approach simply holds on, regardless of prevailing conditions. The market return expected by a buy-and-hold investor is the “unconditional expected return” – something that has historically been about 10% annually. Let's call this E[R]

In contrast, a forecaster does believe that the next draw can be predicted given some information “X”. As that information varies, forecasters will decide to buy or sell. But forecasters typically do something extra. Generally speaking, forecasters are not content with dealing with the present moment, and instead are prone to making bold forecasts about the next month, quarter, year, or even an entire stream of future returns (bull markets and bear markets).

The problem with this, in our view, is that it implicitly assumes that the information set “X” will remain constant. Worse, the size of the forecasts is generally far too large to be rational. A good forecast is most often a humble one.

Robert Hall of Stanford University (also the chair of the NBER Business Cycle Dating Committee that officially dates the beginning and end of recessions) calls this the Iron Law of Econometrics – the variance of a proper forecasting approach will always be smaller than the variance of the actual data. The reason is that if actual returns are equal to expected returns plus a random error,

R = E[R] + e

then a proper forecast is one where the errors are independent of (not correlated with) the expected returns. That means that the variance of actual returns – call it V(R) – must be equal to the variance of your expected returns V(ER) plus the variance of the error terms V(e). As long as there is any forecast error at all, an efficient forecast will always be one where your expected returns are less variable than what actually takes place. Forecasters hate this, because they like to make big, flamboyant predictions about a whole string of events, rather than focusing on the present moment.

Consider that hat full of marbles again. Suppose you are told that 80% of the marbles are green, and that 10 marbles will be drawn (with replacement). If someone asks your forecast, it's very likely that you'll be comfortable predicting that 8 of the marbles will probably be green.

Now suppose the first marble is drawn, and suddenly, someone switches the hat, right in front of you. What happens to your confidence in your forecast? Well, it should collapse, because suddenly you're facing a new X. If the information set X can change, then it is not reasonable to make forecasts that assume that it will be constant over the forecast horizon.

So if we don't want to assume that market returns are simply constant at 10% regardless of valuations or other conditions, and we also don't want to make inefficient forecasts, what is the alternative?

For us, it is to focus on the present moment. We focus on "conditional expected returns" - the return we can expect, given the particular information set X that we have in hand. This is generally written E[R | X]. But unlike forecasters, we recognize that the predictable component of market behavior for any given period is so small, relative to random noise, that making specific forecasts is futile. We take our information set one X at a time, and we rely on discipline and the law of large numbers to mute the impact of that random noise over the long-term.

Specifically, we can go back over history and use observable conditions such as valuations, market action, overbought/oversold status, macroeconomic factors, and so on to separate history into various “bins.” Each bin represents a combination of observable conditions occurring together (what I've called "X"). Then we can ask, for every observation in the bin, what was the market return over a short subsequent period like a week or a month. Each bin then can be associated with a particular expected return and risk profile. Our basic practice is to align our investment position with the set of conditions that we observe at each moment, and to shift our position as the evidence shifts.

Rather than treating the next week, month, quarter or year as a horizon that demands a specific “forecast,” we simply treat each realization as part of a “repeated game,” and rely on the law of large numbers – that is, the idea that if we follow our discipline period after period after period, over time our inevitable errors will average out, and our long-term results will be largely what we expect. The best way to take good care of the future is to take good care of the present moment.

But isn't E[R | X] a forecast?

One might object that by aligning our investment position with the average return/risk profile associated with a given set of conditions, we must, by definition, be forecasting. This is true in the sense that we do have some expectation that market returns under a given set of conditions will be satisfactory or unsatisfactory, given the risks involved. But we differ from “forecasters” in recognizing that the expected return E[R | X] for any short period of time is overwhelmed several times over by the conditional error term “u”. It is only over many, many repetitions that the error terms dampen out.

This is a property that statisticians call “consistency.” Specifically, if a process is consistent, then as you increase the number of observations some random outcome, the average value of your observations will tend toward the true “population” average.

[Geek's Note: If R = E[R | X] + u, then over N repetitions, the standard deviation of the average error is the standard deviation of the actual error terms, divided by the square root of N. So if your conditional error terms tend to have a mean of zero, plus or minus 2.5% on a weekly basis, you would expect that over 100 weeks, your average error would be zero, with a standard deviation of about 0.25%. Over a full market cycle, you will have made a lot of individual mistakes in your investment position, but as long as your errors are not systematic, the combination of discipline and the law of large numbers will work strongly in your favor. Your results will be largely as you expected despite the fact that you made lots of individual errors along the way].

This is basically the dynamic at work when you sail a boat. If you hop into a sailboat and start across Lake Michigan, it is not particularly helpful to make predictions about the direction and speed of the wind over your entire journey. Much better to align your sails as those conditions change, making numerous modest errors, but getting across the lake.

Inquiry

“Suppose the mind consciousness is observing an elephant walking. During the time of observation, the object of mind consciousness may not be the elephant in and of itself. It may only be a mental construction of the elephant based on previous images of elephants that have been imprinted in store consciousness.

“Inquiry means not using the mental creation, but allowing yourself to get in touch, and to try to see how things truly are. We practice not to be influenced by the name, because when we are caught in the name we can't see reality.”

Thich Nhat Hanh

It is important that we don't place so much emphasis on “average outcomes” that we ignore the facts about particular instances. We still have to look carefully at reality to make sure that we aren't assuming away particular features that are important.

This is a risk that market participants seem to be taking here in a major way. Specifically, we have seen a great number of research reports with the basic thesis of “The recession is over. Here is how the market (or the economy, or employment, etc) has performed after a recession is over.” The difficulty is that these are basically attempts to say “here is an elephant” and then immediately move to describing elephants in general, when in fact, this particular elephant is very likely to be pink, or white. Specifically, valuations here are far different than they have been at the beginning of the typical economic expansion. Moreover, economic expansions have historically always been paced by rapid expansion in debt-financed classes of expenditure such as housing, capital spending, and sustained (not just one-off cash for clunkers) demand for automobiles. In prior recoveries, debt-financed expenditures have turned up quickly and have typically led other classes of expenditure by nearly a year.

If we want to see things as they truly are, we have to look both at the elephant, and at anything that might set this particular elephant apart. With regard to the investment markets, if we suspect that the particular features of the present situation make things “different” than they have been historically, then it is best to look closely and get more data.

As an example, during the late 1990's, it was often argued that technological innovation had changed the economy so profoundly that the market valuations of the time were actually reasonable, if not incredibly attractive (remember Dow 35,000?). So we had to open ourselves to the possibility that things were different in an important way. But when we actually looked at the data, there was simply no historical example – in any productivity spurt since the Industrial Revolution – that could support the sort of growth rates that were implicitly priced into stocks.

When we look at the current market environment today, it is clear that the enthusiasm about the market here is largely based on the idea that the recent recession is over, and that the economy will form a “V” shaped recovery similar, but much stronger quantitatively, to standard post-war recoveries. This is a very difficult argument to make, because the drivers of economic growth that existed in typical economic recoveries – particularly debt origination and consumption growth – are very compromised at present. Our perspective on the ongoing credit risk in the economy is much like that of economists Kenneth Rogoff and Carmen Reinhart, who foresaw the recent financial crisis, and are far less sanguine about the prospects for sustained recovery.

As I've discussed in several weekly comments, this is a subject that I have struggled with in recent months. Even if we could assume that the recent crisis was a standard post-war downturn, and that we are now in a standard post-war recovery, valuations would still concern us because at these levels, stocks are not priced to deliver satisfactory long-term returns in any event. However, we would have a greater willingness to take a moderate speculative exposure based on market action and prospects for sustained economic improvement. On the other hand, when we include other post-crash periods into our data set, and allow for the possibility that those instances better describe present conditions, the case for accepting speculative exposure is much more limited. Of specific concern is the tendency in those periods for strong advances (as we've seen in recent months) to be followed by spectacular failures.

So we have to be very careful about how we name things. When people label stocks as being in a “bull market,” the implicit suggestion is that stocks will continue to advance for a sustained period of time. When people say that the recession has ended and we're now in a “recovery,” the temptation is to look at how the market has performed in previous recoveries, without noting the profound differences between those instances and the current environment.

As Thay says, “We practice not to be influenced by the name, because when we are caught in the name, we can't see reality.” The picture in our head can be very influenced by the words we attach to it.

As Zig Ziglar says, “You can tell your wife that she looks like the first day of spring, or you can tell her that she looks like the last day of a long, hard winter. There is a difference.”

Koans

In Zen, there is a teaching tool known as a “koan” – a question that serves as the object of meditation, and is intended to reveal something about teachings like mindfulness and interconnectedness. Western observers sometimes mistake these for riddles, non-sequiturs, or nonsensical statements, but if you look at them carefully, they are questions or stories intended to prompt the listener to see things as they really are.

A riddle is something like this:

Q: “How does a Zen monk know his pizza is enlightened?” A: “It's one with everything.”

Here is a koan:

A novice monk approaches his teacher and asks, “Is this a bull market or a bear market?” The teacher replies, “If it is a warm day, and I say that it is winter, will you still wear your heaviest coat?”

Causes and Conditions

“This is, because that is. This is not, because that is not.”

Buddha