Analyst Behavior: The Geography of Social Interaction

ABSTRACT

In this paper, I provide empirical evidence that an analyst working in Germany is more likely to publish a high (low) price target regarding a DAX30 stock when other Germany based analysts are also optimistic (pessimistic) about the same stock. This effect of geographical proximity is not biased by the fact that DAX30 companies are headquartered in Germany. Shedding light on how influence takes place, I show that influence through communication and the exchange of opinion within small groups of analysts plays a vital role. This mainly applies during a bullish market environment. When markets are bearish, analysts' incentives induce them not to deviate too much from the overall average, such that then observational learning has a greater impact.

KEYWORDS:

• Analyst behavior
• Social interaction
• Herding
• Word-of-mouth

Notes

1. This number has been published by the World Federation of Exchanges members on www.world-exchanges.org and refers to the stocks that are traded on the 54 major stock exchanges in the world. Double counting has tried to be eliminated by only considering domestic stocks from the perspective of each stock exchange.

2. Hereafter, the term analyst always refers to an equity analyst. Due to empirical data availability, the term analyst furthermore always refers to a sell-side analyst. See, for instance, Groysberg et al. [2007] for a detailed comparative analysis between buy-side and sell-side analysts.

3. I use the term “exchange of opinion” to emphasize that information is not only transmitted but also discussed.

4. One could also think of the situation where information is only given away, however, with the intention to influence other analysts such that they skew their valuation results into a desired direction, which makes the own already published result more credible for investors.

5. This website is operated by aktiencheck.de AG, one of the leading providers of stock research in Germany. The company acts as an independent research firm that collects analyst reports and publishes them together with own reports on its web page, free of charge.

6. Ljungqvist et al. [2009] reported systematic errors in the historical I/B/E/S recommendation database. The comparison of price targets that appear in both databases but did not reveal remarkable deviations.

7. The remaining analysts could not be contacted, as they either left their firm or because no or not a valid email address could be found.

8. The results of the questions number 7–9 and 11 are adjusted by the number of intrafirm exchanges as being asked by question 10. This is done for two reasons. Analysts in the same firm act as one unity and only publish one result. Furthermore, the exchange of opinion in a research team takes place by definition and does not provide any insight.

9. Some respondents annotated that from the formulation of the questions it is not perceptible whether the word analyst refers to sell-side or buy-side analysts. Buy-side analysts are sell-side analysts' clients. Hence, the discussion of forecast results between these two types of analysts is natural and not worth of analysis. The fact that 95% of the answers to question number 7 are below or equal to 5 proves that the questions have been understood as being indented, if one assumes that number of sell-side analysts' clients is usually greater than five. As a precaution, the two answers above 15 have been taken off.

10. The second and third most respondents come from the United Kingdom (22%) and France (10%).

11. Using the answers from the survey, it is assumed that if two analysts exchange their opinions then they do so regarding all companies they have in common.

12. See www.cfainstitute.org for more information.