A recent study from the University of Kansas indicates that inconsistencies in analyst revisions may not be the result of analyst bias or conflicts of interest, as has been written about regularly by both academics and the press, but instead may have rational explanations.
Recap of KU Study
Recently, three professors at the University of Kansas, Messrs. Min Park, Michael Iselin and Andrew Van Buskirk, published a study in the Journal of Accounting and Economics called, Seemingly Inconsistent Analyst Revisions. In this study, the professors analyzed whether inconsistent analyst revisions were of lower-quality than revisions that were revised consistently.
According to the KU study, about 20% to 30% of the time when sell-side analysts revise their stock recommendations, target prices, or earnings recommendations, at least two of these estimates are revised in the opposite direction. Traditionally, this is deemed to be an inconsistent analyst revision. Existing academic literature indicates that these “inconsistent revisions” are driven primarily by analyst bias, are less accurate, and are viewed to be less credible by investors.
The authors found that the determinants of analyst inconsistency depend on the different estimate pair being evaluated. For example, inconsistency in target price/earnings estimate pairs is strongly associated with earnings volatility, whether the firm reported a loss, and with changes in aggregate P/E ratios. However, none of those factors exhibit a statistical relation with inconsistency in target price/recommendation pairs.
In addition, the professors found only limited support for the idea that analyst inconsistency is necessarily driven by conflicts of interest related to investment banking activities. They did discover that debt issuance is significantly associated with recommendation/earnings estimate inconsistency. However, they also found that neither debt issuance nor equity issuance are significantly associated with target price/recommendation inconsistency.
In the KU study, the professors examined the effect of analyst consistency on the ex post accuracy of earnings estimates and target prices. They found that earnings estimates are no more or less accurate when they are inconsistent with contemporaneous target price revisions. In addition, they found that analysts who tend to issue more inconsistent revisions are more accurate in both their inconsistent and their consistent target price estimates.
Lastly, the authors evaluated the relationship between analyst consistency and how this impacts investors’ uncertainty, as measured by the change in implied volatility around the report. What they discovered was there was no relation between analyst inconsistency and changes in implied volatility. In other words, they found no evidence that investors perceive inconsistent outputs as less valid or credible.
The results of the KU study indicate that the general assumption/conclusion in prior research that inconsistent analyst revisions reflect lower quality, lower validity, or biased information is not broadly true of analyst inconsistencies. Specifically, the authors found that accounting and economic factors drive inconsistency as much if not more than investment banking activity and that inconsistent analyst revisions are no less accurate, nor perceived as less valid, than consistent revisions.
The publication of the recent University of Kansas study on analyst revisions called Seemingly Inconsistent Analyst Revisions is interesting as it calls into question the traditional view that inconsistent analyst revisions are driven primarily by the conflicts of interest related to investment banking activities. In addition, this study suggests that inconsistent revisions are not necessarily less accurate than consistent revisions, nor are they less credible to investors.
Min Park, assistant professor of business at the University of Kansas, and one of the co-authors of the study explained the general findings of their analysis, “Everybody has pressures that would lead to inconsistencies. That’s well-known in the prior literature. I wouldn’t deny that happens. But the bigger picture is that there is a rational reason behind these inconsistencies. Basically, there are accounting or economic factors why analysts revise their outputs in what looks to be an inconsistent manner.”
In other words, most sell-side analyst revisions – whether they are consistent or inconsistent – are typically driven by logical rationale rather than by analyst biases or conflicts of interest