My new book Soldier of Fortune: Warren Buffett, Sun Tzu and the Ancient Art of Risk-Taking is now available via amzn.to/3JfCPBq

This week in value: Japan’s Value-Up All-Stars (Verdad) Warren Buffett steps down as chairman of Berkshire Hathaway: ‘Father Time always wins’ (CNBC) How Share Buybacks Can Create Compelling Value Trades (EFT) Is Markel the New Berkshire Hathaway Blueprint for Value Investors Post-Buffett? (Simply)

This is a chart showing implied and actual total returns using Shiller PE. It assumes mean reversion to average valuations over ten years. Blue: actual total return (10-year CAGR). Each point is the annualized return, dividends included, that you'd have earned by investing on that date and holding for ten years. The line stops around 2015 because the ten-year results after that date aren't known yet. Red: implied total return. This is a model's forecast, made at each date, of the next ten years' annualized return. Yellow: average total return. This is the long-run average up to each point, settling around 9–9.5% a year. It's the "normal" baseline. Cheap markets imply high future returns and expensive markets imply low ones. The model has worked reasonably well. Red and blue move together through most of the 20th century. The model flagged low returns before the 1929 crash, high returns out of the Depression lows, high returns from the cheap markets of the late 1970s and early 1980s, and near-zero returns at the 2000 dot-com peak. Actual returns then came in slightly negative for 2000–2010. The current reading is extreme. The red line ends near 0–1%, the lowest level on the entire chart and at or below the 2000 bubble peak. It sits far below the ~9% historical average.

Join me on @BizBriefing today with @michaelsantoli @KristinSmith @Greenbackd and more!

Let's go.

August 15, 1971. President Nixon closed the gold window, effectively turning the US dollar into a fiat currency. Since that date, the 10-year yield has averaged 5.91%. It closed today at 5.18%, --- Maybe the bond market won't end humanity.

US 10-Year Treasury Yield just blasted through its 2007 peak.

Looks like a Sornette wave. Parabolic moves end in tears.

Off. Mid- and small caps cannot get up off the mat. Despite underperforming for the last decade+ somehow both still outperforming over 3+ decades.

Ouch.

Earnings forecasts are a terrible predictor

Shiller PE at 41.6x. TTM PE at 26.5. HIstorically expensive.

Evergreen.

Small caps have underperformed large caps over the last two decades. Quite the relative underperformance since 2018. Relentless.

This is a scatter plot of every month since 1973, placing each month by two numbers: the 10-year Treasury yield (horizontal axis) and the trailing price-to-earnings ratio of the stock market (vertical axis). The dark diamond marks today, September 2026, at roughly a 4.8% yield and a trailing P/E near 26. When bond yields are high, stocks tend to trade at low multiples, and when yields are low, multiples expand. Today sits in an unusual spot: yields are moderately high and stocks are expensive. Only 29 of 645 months (4.5%) had both a higher yield and a higher P/E than now, and every one of them falls in a single stretch from April 1998 to June 2002, the peak and unwinding of the dot-com bubble. Trailing P/E can look stretched when earnings are about to accelerate, which is the bull case investors are implicitly making (much of it around AI-driven profit growth). The market's composition has also shifted toward higher-margin, faster-growing companies than in the 1970s–90s, which can justify structurally higher multiples. And the relationship between yields and multiples was much weaker during the low-rate 2010s, so the regression line isn't a law.