Which Types Of Analyst Firms Make More Optimistic

Which Types Of Analyst Firms Make More Optimistic Forecasts? Amanda Cowen, Boris Groysberg, and Paul M. Healy Discussion by Bruce N. Lehmann Graduate School of International Relations and Pacific Studies University of California at San Diego Financial Markets Conference 2004 Federal Reserve Bank of Atlanta Sea Island, Georgia August 15, 2004 Copyright © Bruce N. Lehmann 2004

The Paper in a Nutshell • Price and earnings forecasts by analysts at: + underwriters are least optimistic + syndicates are in the middle + brokerage firms are most optimistic + pure research firms have no clear bias • Analysts are more optimistic about firms they have followed for a long time • Conclusions + Sales and trading incentives more important than underwriting Copyright © Bruce N. Lehmann 2004

Road Map for Discussion • Point of view + Nature of the experiment + My brother, the (sort of) analyst • Issues + Reinterpreting the results + Benchmarks + Nature of product + The uses of language in the analyst business + Distribution method for analyst reports Copyright © Bruce N. Lehmann 2004

A Pox on All Their Houses • The paper: + measures the relative differences in average forecasts across analyst domiciles + concludes regulators need not worry about underwriting firm analysts who are relatively pessimistic (on average) • Alternative interpretation + All analysts are sinners + Those at brokerage firms are the worst sinners + Regulators should punish all sinners and require penance and regular attendance at. Copyright church © Bruce N. Lehmann 2004

Benchmarks • Maybe optimism should be measured relative to domicile-specific benchmarks + If underwriters must be optimistic to generate business at the margin, they need not be optimistic relative to non-underwriting analysts • Maybe optimism only counts when actively trying to take a specific firm public + Maybe analysts at underwriters are only biased conditional on the target being in their sights + Maybe they even signal they can market ipo’s Copyright © Bruce N. Lehmann 2004 about which they are relatively pessimistic

Nature of Product • Analysts are judged on: + Many reports, not one forecast at a time + the ideas and insights underlying their forecasts, not on the specific point forecast + how bad their forecasts are when far from peer forecasts, not so much on the good ones • What if analysts are compensated in this way and they were the unit of observation? + What data would you need? Copyright © Bruce N. Lehmann 2004

Code and Language • Hold means buy, weak buy means hold, strong buy means buy… + Code might be optimism relative to domicile specific benchmark or language in report • Implications for analysts with long tenures + Likely their message is in the language of the report, not in specific point forecast • Negative reports represent a career decision + Analysts are responding to different incentives in this case Copyright © Bruce N. Lehmann 2004

What is the Production Function for Distributing Research? • Typically a layer of “sales” analysts between actual analysts and buy side clients • “Sales” analysts pick and choose the research they tout from their firm’s menu • Do “sales” analysts tout biased reports when they hope to take a firm public? • How do relations between actual and “sales” analysts affect the reports that are taken to the buy side firms? Copyright © Bruce N. Lehmann 2004

Differential Incentives Copyright © Bruce N. Lehmann 2004

Endogeneity Copyright © Bruce N. Lehmann 2004