September 13th, 2019 | by The Gerstein Fisher Team
Accounting Facts and Security Analyst Errors
- Security analysts, like individual investors, fall prey to behavioral biases
- These errors are quite systematic.
- “Hard” accounting characteristics such as profitability are useful tools for security analysis but are misinterpreted by analysts.
Behavioral finance is a topic on which I’ve conducted a considerable amount of research and written before.* But this is mainly about how biases lead individual investors astray. What about investment-industry pros, are they also vulnerable to behavioral biases and errors when they analyze securities? I recently completed a comprehensive research project with two partners, Ronnie Shah and Sheridan Titman, in which we studied the link between sell-side security analysts’ forecasts and subsequent realized earnings and stock returns. I’d like to briefly summarize a few of the findings.
For our study, we used consensus analyst long-term growth forecasts drawn from Institutional Brokers Estimate System (I/B/E/S), and we examined several balance sheet and income-statement accounting ratios (i.e., “hard” data) at the security level that have been shown by academic studies to predict stock returns. Specifically, we studied the following four factors, or firm characteristics: profitability, and three yardsticks of growth—sales, assets, and external financing or equity dilution.
A Pattern of Errors
What we found was quite perverse. In reality (and as borne out in research), firm profitability is quite persistent and positively correlated with stock returns. But analysts mistakenly assume that profits are mean-reverting and therefore set security-price and growth expectations too high for stocks with low profitability (i.e., too optimistic) and too low (too pessimistic) for highly profitable companies.
The pattern with the three growth measures is similar. These are, in fact, negatively correlated with stock returns, but analyst perceptions are the inverse. Specifically, whereas analysts perceive high past sales growth to be a good predictor of future earnings growth and returns, high sales growth is actually negatively associated with future earnings.
Asset growth and external financing are two characteristics over which company managements exercise considerable discretion. Analysts are optimistic when firms take actions to promote growth, whereas the relationship between these choices and realized earnings growth is actually negative (perhaps analysts fail to account for managers who are overly optimistic or guilty of “empire-building” tendencies in the businesses). In other words, there appears to be an erroneous belief that growing firms are making positive net-present value investments that will generate future earnings.
Exhibit 1 provides one illustration of the discrepancy between analysts’ long-term growth forecasts, from 1982 to 2014, and the 5-year realized EPS annualized growth rate from 1982 to 2009. Note that the mean projected growth rate is remarkably stable, ranging from 15% in 1982 to 20% in 2001, before falling to 14% in 2014. The actual 5-year growth rate fluctuates far more dramatically, from just above zero to 18%. Note that realized growth tends to be high right after recessions (e.g., 2001 and 2008), when expectations are low.
In sum, “hard” accounting measures such as profitability and asset growth are very useful predictors of future excess stock returns. The problem is that analysts misinterpret these characteristics in a systematic way, which results in faulty growth expectations and security mispricing. These biases are consistent with return patterns that we observe: For instance, the high returns associated with profitability may be due to the belief that earnings are mean reverting when in reality earnings are quite persistent.
I have scarcely scratched the surface of our study and invite you to review the full paper: “Do Growth Expectations Help Explain Characteristic-Sorted Portfolio Returns?”
*See, for example, https://gersteinfisher.com/gf_article/past-performance-is-indicative-of-future-beliefs/