Ivey Business School’s Value Investing Program | Brian Chingono
Why It Matters
Because growth persistence is essentially random, investors who base valuations on projected high earnings risk systematic overvaluation, while a simple GDP‑growth benchmark can improve return forecasts.
Key Takeaways
- •Growth rates rarely persist beyond random chance across firms.
- •Revenue shows slight persistence, earnings growth does not.
- •Analyst forecasts systematically overestimate earnings growth in future.
- •Analysts correctly identify extreme growth performers, miss median cases.
- •Historical findings on growth persistence hold true post‑1997.
Summary
Brian Chingono presented an out‑of‑sample replication of the seminal Chan et al. study on growth persistence, extending the original 1951‑1997 dataset to 1997‑2022 and covering U.S., European and Japanese firms.
The analysis shows that companies with above‑median growth in one year retain that status at rates matching random coin‑flip probabilities—about 25 % in year 2, 13 % in year 3, and 6‑7 % in year 4. Revenue growth exhibits a modest 2‑3 % excess over chance, while earnings and profit line items are statistically indistinguishable from randomness.
Chingono highlighted the practical fallout: investors often extrapolate linear earnings trends, as illustrated by the Nvidia episode, leading to overvalued high‑growth stocks. Analyst forecasts systematically overshoot actual earnings, especially on lower‑down‑statement items, and while they sort extreme performers reasonably well (≈48 % of top forecasts land in the top outcome quintile), median predictions are no better than assuming nominal‑GDP growth.
These findings caution valuation models against assuming sustained high growth and suggest a baseline of GDP‑linked earnings growth for most firms. Overreliance on analyst optimism can inflate prices, whereas recognizing the rarity of true growth outliers helps investors allocate capital more prudently.
Comments
Want to join the conversation?
Loading comments...