Expertise about technology, markets, economics and politics

Think and Grow Rich

Why do small-cap stocks, on average and over time, produce larger returns than large-cap stocks?

Wall Street has a standard answer. Small-cap stocks, like rowboats on the Atlantic, are more risky than the big ocean liners of the S&P 500. They get tossed around in storms. One leak can send them to the bottom.

With all this added risk, goes the theory, that investors would never buy small caps unless they received or expected to receive a larger return. Small caps pay more, supposedly, because the efficient market noticing their riskiness, suppresses their price. Arithmetic says if they survive, their percentage return will be higher.

This idea—that investors are, on average compensated for risk—has dominated Wall Street for more than 50 years. The idea has been frequently challenged, we would say discredited, but it hangs on. Part of the reason is that academic finance is wedded to statistical methods in research.

Suppose for your graduate thesis you set out to test whether the application of certain investing principles can raise returns over time. You might, for instance, wonder whether traditional value investing can increase returns.

Your thesis advisor lays out the rules. You must pick an arbitrary period of say, 20 years, though 40 would be better. You may not consider any special information or insight about markets during the chosen period; that would violate the standard of “arbitrary.” You must put your investing rules in place before you run the data and they may not be altered at any time.

Suppose you find that for the first five years, value worked. For the next seven, it did not. For the next three, it did. And for the final five years, it didn’t.

How exciting! You conclude you are on to something. In your thesis, you offer several explanations as to why value worked in period A, not in period B, and again in period C and not in D. Naturally, you are very proud of your discovery.

Alas, your academic career is now over. You broke the most sacred rules of academic research. You drew conclusions based on known results. You were not “arbitrary.” You “data mined,” committing the very grave sin of forming your thesis based on known results. Take this quarter, call your mother, and tell her you are not going to be a professor of finance.

Take heart. Good investors make more money than good professors. This is because what professors call “cheating,” investors call “learning from experience,” or, Heaven forbid, “thinking.”

The statistical dogmas of academic finance, including “returns are a reward for risk,” reduce to a single idea “thinking does not work.” Like flipping a coin, it’s all stats in the long term and luck in the near term.

So please, when you are persuaded by our brilliant colleagues at George Gilder’s Moonshots to invest in a wonderfully innovative start-up, and you make 200% on your money, don’t go bragging about your brilliance and certainly don’t give any credit to that silly Moonshots team. You made the money only because you took a bigger risk; next time you may not be so lucky.

We take a somewhat different view. We believe in thinking (which happily has not yet been reduced to statistics, despite AI efforts to the contrary). Thinking is just about the best thing an investor can do.

The reason small stocks pay more is that thinking works especially well for them. That’s thanks to the fundamental principles of information economics. As George explains in his new book, “Life After Capitalism,” the four pillars of information economics are:

  • Wealth is Knowledge
  • Growth is Learning
  • Information is Surprise
  • Money is Time

The world’s biggest and most successful companies are worth hundreds of billions because, over time, they have accumulated vast stores of relevant knowledge. If you doubt this and want to say  “no it’s the physical assets they have accumulated that make their wealth” observe how quickly an old company collapses when it fails to add to its stock of knowledge. Indeed, knowledge is accumulated learning, and learning is growth. These companies have grown greatly by progressing along the learning curve, which dictates that for every accumulated doubling of volume, costs drop, or value-added rises, between 20-30%.

Eventually, each accumulated doubling gets a bit harder, with two results. Learning slows and the ratio of new information to well-established knowledge declines.

Because only surprise counts as information, the new learning may have a smaller impact on a company that already knows a great deal. It may even know too much, if that hoary knowledge makes it less appreciative of surprise.

We started with the metaphor of a small boat and a great ship on the Atlantic. Think of the waves as surprises. They hardly affect the great ship at all, such is its accumulated mass, but they may be decisive for the little boat.

The metaphor is imperfect because the little boat does not make the waves. But when a tiny start-up is making some of the waves—generating the surprise—it may be transformed in a way the ocean liner is not.

Crucial is the ratio of new information—surprise—to the mass of accumulated knowledge.

What’s your nominee for the most surprising firm of the decade?

How about Nvidia? From 2010 through 2016 when most people, including perhaps Nvidia, believed NVDA was making graphics cards, revenues rose 50%. Then it gradually dawned on the world that NVDA was making the world’s first practical AI processors. From 2016 to date, NVDA’s annualized revenues are up 552%.


When you invest in an innovative small-cap company, the most important thing is not the size of the company but how much beneficent surprise it is generating.

That’s what you should be thinking about.

P.S. You’ve got to come to COSM 2023, Nov. 1-3 in Bellevue, Washington. COSM 2022 and 2021 were probably the best tech gatherings we’ve ever been to, and the 2023 version is not to be missed.

COSM is the ultimate expression of George’s worldview, the Gilder Team’s insights into what is happening in tech, how it matters to the world and especially to our readers and tech investors. Save the dates of Nov. 1-3:

The focus this year is on AI and all its works. Key speakers include:

  • The Wall Street Journal’s Andy Kessler on the economics of AI.
  • Juan Lavista Ferres, Microsoft’s chief scientist on AI’s potential for global problem solving.
  • Ray Kurzweill will shock you with the prospects for AI immortality.
  • Archana Vemulapalli, head of solutions architecture at Amazon AWS, will plunge into the AI open or closed debate.
  • Michael Milken will propose a new AI-enabled high-yield healthcare system.
  • …and more

Plus, you will meet lots of key folks from the companies we cover.

  • Ariel Malik, the venture capitalist backing a dozen graphene companies spun out of Jim Tours Rice University Lab, will give important updates on the graphene revolution.
  • Steven Balaban, of Lambda Labs, a George Gilder favorite, will cover the prospects for companies enabling AI on the edge.
  • DO NOT MISS Vered Kaplan CEO of Orgenesis on the amazing prospects of affordable cell therapy.
  • Another half a dozen start-up heroes.

Speaking of heroes, the brilliant and brave Michael Shellenberger (recently harassed by Congress People of Limited IQ) will speak on Free Speech in the Digital Age.

As always, Carver Mead will give a riveting reflection on our three days together.

Social time is great—meet old friends and fellow subscribers and investors.

George and Nini, of course, and the rest of the Gilder Team, John, Steve, Paul and Richard will be there, too.



P.P.S. Come join our Eagle colleagues on an incredible cruise! Set sail on Dec. 4 for 16 days, embarking on a memorable journey that combines fascinating history, vibrant culture and picturesque scenery. Enjoy seminars on the days the ship is cruising from one destination to another, as well as dinners with members of the Eagle team. Some of the places on the itinerary are Mexico, Belize, Panama, Ecuador and more! Click here now for all the details.



Log In

Forgot Password