Can artificial intelligence usefully estimate intrinsic firm value without cues from market prices? In his September 2026 paper entitled "Valu(AI)tion: Machine Value and Market Prices", Marc Schmitt builds a machine analyst that each year forecasts profits strictly from published firm accounting data (no stock prices/returns or analyst estimates) and calculates firm value by inserting forecasted profits into a residual income model. He assumes accounting data are available four months after fiscal year ends. He then each month reforms an equal-weighted hedge portfolio that is long (short) the tenth, or decile, of stocks with the highest (lowest) estimated valuation-to-price ratios. Using publicly available accounting data for U.S. firms during 1975 through 2024, with valuation testing starting in 1985, he finds that:
Subscribe to Keep Reading
Get the research edge serious investors rely on.
- 1,200+ research articles
- Monthly strategy signals
- 20+ years of backtested analysis
Cancel anytime