Gaming Data Gives No Edge. Instead, Game the Data Devoted
Ken Fisher
Founder, Executive Chairman and Co-Chief Investment Officer of Fisher Investments
How can you gain an edge? That timeless question has confounded investors almost forever. Today, many fixate on data, thinking quick data access and manipulation mines magical statistical relationships. “Quantitative” fund managers run vast sums on this premise in New York, Singapore, London, Tokyo, Sydney and all the world’s financial hubs. It isn’t just these managers, though—belief in data’s dominance abounds.
This approach suffers one huge flaw: Data has been commoditized. Smartphones harness almost endless data. Professionals consider parallel stats. A plethora of American universities ranging from Princeton and Carnegie Mellon to the Universities of Illinois and Minnesota offer master’s programs in Quantitative Finance. That, too, ripples globally. Data devotion is worldwide, with many convinced statistical deep dives reign supreme in capital markets. The fatal flaw? Anyone can crunch the numbers. Anything anyone can do holds no edge in capital markets. Today’s investing requires looking past data and numbers, seeking truths the data devotees shun. Let me explain.
Decades ago, in pre-Internet days, information was scarce—hence, precious. Morning newspapers brought yesterday’s news. Magazines came weekly or monthly—even older news. Deeper analysis often required library trips or diving into musty archives, microfiche and library records.
Using data was tough, too. There were no algorithms or spreadsheets to array data and quickly test hypotheses. In the mid-20th century, my father, Philip Fisher, possessed such high-powered data crunching tools as a hand-crank comptometer and pencil. This was enough then to become an investing legend. When I started in markets 49 years ago, I had little more. Testing new theories required painstaking effort.
When I developed the price-to-sales ratio (PSR) in the early 1980s, it worked wonders. PSRs compared a firm’s market capitalization to its revenues, identifying temporarily troubled, profitless firms with oversold shares. If I could identify the potential for future profit rebounds, I really had something. And it worked. I gathered data from The Wall Street Journal’s then-daily “Earnings Digest,” quarterly company reports and monthly brokerage “stock guides” to calculate PSRs. It was tediously time-consuming. Eventually, while I was still virtually broke, I paid Goldman Sachs $25,000 (around $66,800 today) for a one-time PSR screen of all NYSE listings. It was a lot of money for me then but it was … valuable!
PSRs worked then precisely because data was scarce and no one used them. Few conceived of comparing prices to sales. Fewer still had the time, tenacity or resources to deploy PSRs broadly. After my 1984 book, Super Stocks, their popularity slowly grew. Professors taught them. More demand made PSRs more available. Stocks priced them in, sapping their power. They don’t really work anymore.
The Internet’s abundant, accessible data accelerates this process. Any widely known metric—from price-to-earnings and price-to-book ratios to high-frequency economic data, Delta variant case counts, vaccination progress, highway traffic, on and on—is yours instantly. If anybody can do it, it holds little stock market power—always.
Rather than accept this, data devotees double down, seeking ever-more data and deeper statistical complexities. Their exertion makes truly unique findings near impossible. Any data edge is fleeting. All their number crunching didn’t help data-driven managers navigate 2020. Early in the pandemic, they saw an inverse relationship between COVID case counts and the S&P 500. Many presumed this meant the former would dictate returns. But that relationship broke down after mere weeks, leaving devotees confused by the supposedly irrational stock market. Whenever you think stocks are irrational, it is more likely you are. This irrationality view, plus a persistent value bias—because the data say value leads early in bull markets—torpedoed devotees’ returns last year. Data-driven managers have overall been more in line with their benchmarks this year, but many are still lagging.
领英推荐
Instead of following them, look to what the data-devoted dismiss because of biases. Humans are wired to seek and believe data supporting pre-existing beliefs and dismiss contradictory information. Psychologists call this “confirmation bias.” And it creates overconfidence. It benefited our ancient ancestors, building the confidence needed to face daunting tasks with high failure rates—like running up to big wild beasts with stone-pointed sticks to try to get something to eat. But in markets, it hurts. Instead, note that others’ susceptibility to bias can be your edge.
Presently, confirmation bias thwarts the data-devoted. Despite 2020’s quant lag, many still bet on the value leadership nearly everyone expected last year. But this time? Global growth stocks beat value 81.4% to 58.1% from the bear market’s March 23 low through 2020’s close. Too many expected value leadership. When many believe in something everyone can do, it is pre-priced—Markets 101.
Still, every time value jumps, devotees start crowing. Twin value countertrends last November and early this year had many shouting they weren’t wrong—just early. Vaccinations, they argued, would deliver lasting fast growth and higher inflation, supercharging economically sensitive value stocks. But their edge didn’t last. From mid-May through September 24, global growth stocks rose 15.2%. Value? Just 0.9%. On the year, value is again lagging.
Data devotion blinds adherents to contradictions explaining why value stocks didn’t lead last year and likely won’t ahead. Value leadership immediately following bear market lows is what usually happens, but data says nothing about why—and why is always harder to grasp and much more powerful. Bear markets typically start slow and grind lower as recession approaches. The majority of the decline, accompanied by the really big plunges, usually comes late. Economically sensitive value stocks are highly credit-reliant. They tumble as banks recoil from lending. Eventually, sentiment overshoots. Credit conditions ease. Relief sends value surging.
But the ultra-swift contraction last year didn’t give value time to grind lower, reach capitulation and prime a bounce. Value plunged -37.7% in a month! Without a relief rally, value will need far higher long-term interest rates to help banks’ profits—and encourage them to lend. Despite widespread inflation fears, that looks unlikely in any time frame stocks care a hoot about. “Stimulus” seems to be stuck in gridlocked Washington, and it has no history of driving swift growth and inflation anyway, as I have detailed before. Meanwhile many inflation metrics’ jumps this year are fleeting—skewed by last year’s lockdowns temporarily depressing year-earlier results—and the subsequent re-openings.
Historical analysis has big benefits. But you must look beyond raw numbers—investing isn’t mathematics. Remember: While data are cold and unfeeling, everyone interpreting them is biased and emotional—human. Investing successfully now is about gaming others’ blinding biases.
Ken Fisher is founder and executive chairman of Fisher Investments. Follow him on Twitter @KennethLFisher.
?????? ???? ???????? ?????????????? | Creating generational wealth for clients with hands off passive investments ?? | Alternative Assets | Energy | Real Estate | ForEx
3 年Great article Ken!! People would learn a lot about data from this. They would understand it more and learn how to gain an edge. Thanks for sharing!! ??