Moneyball Versus The Investment Crowd

My last post introduced the concept of Moneyball investing, a catchy (I hope) term I borrow from Michael Lewis to describe an approach to investing that stands between Indexers who restrict decisions and subjective Active investors. Analogous to what’s done in baseball, Moneyball investing means making on the basis of factors, rules, evidence, etc. Now, let’s see what, it can bring to the table in terms of return and risk.

My Favorite Guru Quote

Peter Lynch opened Chapter 9 of his classic “One Up on Wall Street” as follows:

“If I could avoid a single stock, it would be the hottest stock in the hottest industry, the one that gets the most favorable publicity, the one that every investor hears about in the car pool or on the commuter train— and succumbing to the social pressure, often buys.

“Hot stocks can go up fast, usually out of sight of any of the known landmarks of value, but since there’s nothing but hope and thin air to support them, they fall just as quickly. If you aren’t clever at selling hot stocks (and the fact that you’ve bought them is a clue that you won’t be), you’ll soon see your profits turn into losses, because when the price falls, it’s not going to fall slowly, nor is it likely to stop at the level where you jumped on.”

I’ve lost count of how many times I quoted this, and I’ll probably keep doing so because this may be the single most-important piece of wisdom any guru has or can pass on to anyone.

The first edition of that book was published in 1989 and Lynch was sharing the benefits of experience he gained during prior years. Since them we’ve had our information revolution. Even the greenest rookie can get vital information as quickly as the most hardened pro, and so, too can everybody in the media who talks and writes about stocks.

So have we entered into an investment version of the Enlightenment? Can we confine Lynch’s warning to the Dark- and Middle-Ages of Wall Street?

So it’s really not about information. Everybody has that. Moneyball is about the benefits of disciplined systematized use of the information.

Testing the Waters

It’s hard to precisely test this since we still lack simple reliable measures of the sort of crowd appeal to which Lynch referred. But perhaps we can come up with some tolerable proxies. Here are some ideas:

• Dollar Volume Traded: This is volume multiplied by the price of the stock. Like basic volume, it measures the extent of trading activity but does so in a way that does not “penalize” stocks with high prices, stocks in which a desired dollar investment can be obtained with fewer shares. But there is still a potential size-based distortion. So rather than using this, I’ll make one more adjustment:
• Scaled Dollar Shares Traded (SDST): Even a boring day in Apple (AAPL) stock is going to be far more active than an exciting trading day for a small cap firm, simply because Apple is so darned big. So to control for this, I’ll divide Dollar Shares Traded by Market Capitalization. It’s not perfect. It measures activity. It doesn’t tell us whether the sentiment is bullish or bearish. But given ad hoc observations regarding how so much more media and analyst commentary is bullish, it’s not unreasonable for me to assume that a significant part of this activity (albeit definitely not the only part) is positive in tone. So I’ll use this (I’ll divide the 60-day average dollar volume by the latest market cap). But it would be nice if I had a bit more.
• Technical Trend: I’ll also conduct some tests in which I’ll apply my SDST sort only to stocks that are trending up. Long-trend samples will consist of stocks for which the 50-day simple moving average is above the 200-day average. Short-trend tests will be based on the 5- and 20- day moving averages.
• S&P 500: Finally, all testing will be done only on stocks that are part of the S&P 500. Since I’m trying to tease out crowd attention, it makes sense to confine the study to the most intensively watched group of stocks on our planet.

In all tests, I’m going to pick the 10 blue-chip stocks that ranked highest in terms of SDST. I assume all such crowd-generated portfolios will be refreshed every four weeks and I include transaction costs through an assumed 0.25% per trade price penalty (slippage).

One type of test, the basic test, tracks a hypothetical portfolio managed this way from 1/2/99 through 11/5/15.

The other test, a rolling test, is not sensitive to the starting date and the particularities of the resulting every-4-week refresh dates. It measures a lot of self-contained 4-week holding periods (875 to be exact). The first of these begins on 1/2/99. The next 4-week test begins 1/9/99, and so on and so forth with a new test beginning each week until the present. We’ll see the average of all of these four-week experiments, as well as separate averages for 4-week periods in which the market rose and periods during which the market fell.

The Results

Table 1 shows the results.

Table 1

The rolling tests in Table 1 suggest it’s OK to run with the bulls in up markets. That makes sense. It’s consistent with Lynch’s observations that there are times when money can be made this way. But also consistent with Lynch is evidence to the effect that down markets can be brutal.

Personally and based on Table 1, I’m not at all impressed by crowd chasing. But I’m a conservative kind of guy. I can easily see somebody (my wife for example) pointing out that the market has, for as long as many can remember, had more up periods than down. The overall rolling-test average might justify some in deciding its OK on the whole, to run with the bulls.

Fair enough – to a point. In response, I suggest first that who knows if the market will change direction (we’ve had decades of falling interest rates, suppose that reverses), and second, that taking on as much volatility (standard deviation) and downside risk as we see seems senseless considering how easy it would be to do so much better using just a dash of Moneyball-style fundamental analysis.

The fundamentals I’ll interject won’t even be all that complex. They’ll involve use of two ranking systems I created for Portfolio123 based on widely known and used metrics (a Value ranking system based on PE, PEG, Price/sales, Price/Free Cash Flow and Price/Book) and a Quality ranking system based on Margins, Turnover, Returns on Capital, and Financial Strength). For this next test, I’ll continue to take the crowd into account (i.e., pick the top 10 S&P 500 stocks in terms of SDST) but apply that sort only to a pre-qualified (i.e., screened) group of blue-chip stocks that rank 75 or higher (with 100 being the best) under both the Value and Quality ranking systems.

Here are the Results:

Table 2

Table 2 does not abandon the crowd. I continue to pick the 10 stocks with the highest SDST (Scaled Dollar shares Traded) data. Nobody has to be a hermit, and nobody has to beg Matt Damon to escort them to Mars in order to hunt for obscure undiscovered companies. We’re dealing with S&P 500 firms, the ones whose fundamental merits are as widely disseminated as any information anywhere. It’s not about the information: It’s about the systemization.

Note too we’re not interviewing management, we’re not listening to conference calls, we’re not communicating with suppliers or competitors, we’re not curling up in front of a fake fireplace with 10-Ks opened to accounting footnotes. In fact, we may or may not even bother looking at the names of the firms in which we invest. (The closest I come to this sort of thing is noticing tickers when I download them from Portfolio123 into Excel and then up into my FolioInvesting.com brokerage account).

To me, the traditional securities-analyst things are about as useful as the five tools of baseball (hitting for average, hitting for power, running, throwing, and fielding) are for Billy Beane. Like the latter, I’m into data. He’s interested in OBP (On Base Percentage), WVAR (Wins Above Replacement Value), etc. I’m interested in the kinds of metrics I put into my models, as listed above.

I play factors, not companies or sectors our countries. Some indexers do a bit of this (tracking a mid-cap value index for example). Some Active investors dabble in legitimate but less systematic ways (the measures I use are well discussed in books by or about such legends as Lynch, Warren Buffett).

I’m not Warren Buffett or Peter Lynch. But with Factor-Based Moneyball investing, I don’t have to be. It’s not simple. It takes work. But it’s do-able, by me, and by and for many others. This is something the new robotic era of stock selection brings to the average investor.