Fundamental Extrapolation and Stock Returns

Abstract

Recent studies show that most accounting anomalies are either decayed or explained away by
stock market risk factors, challenging the importance of fundamental analysis. We propose
three pooling strategies to exploit information in multiple fundamental signals and their trends.
We find that our fundamental extrapolation yields strong evidence that fundamental analysis
matters, providing again economically significant alphas. To explain our evidence that supports
the fundamental analysis, we propose further a simple equilibrium model allowing for both
a cash flow effect and a discount rate effect. The novel discount rate effect, which pushes
stock price up relative to the fundamental value and depresses stock price with an increased
exposure on the consumption volatility, helps to understand the return predictability. The
theory provides a unified framework and can be applied more broadly to explain seemingly
contradicting findings in the accounting literature.

Keywords: Fundamental extrapolation; Pooling extrapolation; Price extrapolation; Machine learning; Expectation

JEL Classification: G10, G14, G15, G30, G32, G41