Post-revision return drift—the long-term return for stocks after analysts adjust their related recommendations—is estimated to be insignificantly different from zero during 2003–2010. The authors investigate transaction cost changes as a primary explanation and anticipate a reduced information intermediary role for analysts.
What’s Inside?
The authors present new evidence to discount the impact of post-revision return drift (PRD), the historical finding that changes by sell-side analysts of their stock recommendations predict related future long-term returns in the same direction as the change. For the period 2003–2010, average PRD is no longer persistently different from zero, and the authors find no statistical support for a causal relationship between revisions and PRD. Investors do not underreact to the analyst changes, nor are analysts demonstrably better informed about stock prospects. The disappearance of PRD may be because of generally declining transaction costs and the subsequent opportunism of profit-taking arbitrageurs.
How Is This Research Useful to Practitioners?
The authors’ conclusions appear to be worth noting by various investment practitioners.
For asset/wealth managers, the authors offer revisionary thinking with subtlety. They note that the rational pricing of securities in the form of competition for returns combined with changes in real inefficiencies as illustrated by falling transaction prices can extend market efficiency, which may contribute to a better understanding of indexing. Yet for those still willing to pursue active management, some modest rewards may be available. In partitioning the investigation by size of firm and associated trading volume, the authors find modest PRD in some small subsamples (particularly for the lowest decile and for NASDAQ-listed firms), but the results are not particularly robust.
For stock analysts, their ability to add investment value through research seems modest at best and disappointing in terms of their significance as information intermediaries in securities markets. There is little average investor reaction to analyst reports, which seems sensible because those reports do not appear to provide useful new data about long-run stock price behavior. Focusing attention on primary sources as opposed to a filtering mediator is the more prevalent investor technique for gauging opportunities and managing risks; the authors note measurably greater scarcity of neglected public information with heightened competition to react quickly to corporate announcements.
How Did the Authors Conduct This Research?
The First Call Historical Database of analyst recommendation revisions is the authors’ key data source, with “revision” defined as a status shift on the five-point continuum from strong buy to strong sell. Contemporaneous daily stock price and volume movements from the CRSP database are tracked to gauge return effects. The authors cover 1993–2010, and to control for potential earnings announcement effects and high-frequency algorithmic trading on PRD, they split the data into two periods: 1993–2003 and 2003–2010.
Several PRD measures are investigated across three matching samples (event time, portfolio, and earnings) to reflect the various research methods used in previous studies of the topic. With an event-study model, the authors align revisions with their announcement dates. Using portfolios in calendar time, they compare the performance of buy portfolios of upgraded stocks with that of sell portfolios of downgraded stocks. Finally, they innovate methodology to align firms on their earnings report announcement dates and compare (1) drifts for firms with upgrades or downgrades with (2) drifts for firms with unchanged ratings.
Regressions are run with a 12-factor specification across the three samples to document PRD trends over time in one-, three-, and six-month intervals. The effects of transaction costs on PRD’s persistence (or lack thereof) is tested in light of volume, firm size, and coverage effects, with larger ratings on these dimensions indicating lower transaction costs. Underreaction and informed analyst hypotheses are also tested to assess any divergence in knowledge between investors at large and the specialist analysts. Finally, patterns in major foreign markets are checked to confirm that PRD trends parallel US findings.
Abstractor’s Viewpoint
More than four decades ago, the Financial Analysts Journal warned CFA® charterholders via such insightful observers as Charles Ellis, CFA, that active managers’ pursuit of significant portfolio alpha was a Holy Grail–type quest. Plus ça change is now a better summary of contemporary conditions. Perhaps the real value from the authors’ research is not the analyst recommendations but the comparative framework their research provides for gauging one’s own reasoning and independently derived conclusions.