Do Asset Prices Reflect Fundamentals? Freshly Squeezed Evidence from the OJ Market

Motivation(s)

In the seminal paper on frozen concentrate orange juice (FCOJ) future prices and weather, the author identified that fundamentals (e.g. cash flows, expected returns) have little explanatory power for asset returns. Even though the results lack any mention of market irrationality, the excess volatility literature have subsequently adopted and frequently cited it in that context.

Recent behavioral finance research proposed an alternative explanation that the relation between asset prices and fundamental information is incorrectly measured or not measured at all.

Proposed Solution(s)

The authors assert that existing literature has misinterpreted the data by ignoring the state dependence inherent in the structural relation between FCOJ returns and their underlying fundamentals. Asset prices drift away from their underlying fundamentals due to irrational agents who ignore these fundamentals. The previously identified weather surprises should impact future returns only around freezing temperatures. Examining non-weather related news about supply-side shocks provided further evidence for the proposed theory’s prediction power.

Evaluation(s)

The data used in the experiments spanned from 1967 to 1998 and consisted of temperature forecasts, USDA production forecasts conditioned on a non-freeze season, resolution of production uncertainty, FCOJ futures returns, and production from areas outside of Orlando specifically Brazil.

The empirical data revealed that the original FCOJ study incorrectly assigned nonzero importance to months that can not trigger a crop freeze. The linear regression model it used cannot capture the nonlinearity of the weather: only a surprise that causes a transition between the freezing and nonfreezing state matters. Moreover, the regression of asset returns on forecast errors implicitly assumes that a zero forecast error should have no effect. In contrast to the original study’s conclusion, the FCOJ futures return variability and direction seems to coincide more with the author’s proposed piecewise linear model.

Since the goal is to demonstrate that relevant information about fundamentals affect asset prices, the authors chose to ignore return-relevant information. However, to build the correct structural model with all the relevant interactions and accurately measure the fundamentals is a difficult task.

Future Direction(s)

• What pieces of information returned by unsupervised learning can count as fundamental?

• How can counterfactual simulations be used to identify fundamental information?

Question(s)

• What percentage of finance papers blindly accept results like FCOJ?

Analysis

Fundamental information do have a strong predictable influence on asset returns assuming the model represents the relationships accurately enough. When encountering any interpretation of data, always question the methodology that led to the conclusion. The paper could have been summarized more succinctly while still maintaining a strong argument. It is surprising that the excess volatility literature would simply cite a result without questioning its applicability. The data in the appendix is very useful as labeled data for future exploration of identifying fundamental information.

References

BRSW07

Jacob Boudoukh, Matthew Richardson, YuQing Jeff Shen, and Robert F Whitelaw. Do asset prices reflect fundamentals? freshly squeezed evidence from the oj market. Journal of Financial Economics, 83(2):397–412, 2007.