Brokerage Relationships and Analyst Forecasts: Evidence from the Protocol for Broker Recruiting

Abstract

In this study, we offer novel evidence on how the nature of brokerage-client relationships can
influence the quality of equity research. We exploit a unique setting provided by the Protocol for
Broker Recruiting to examine whether relaxed broker non-compete agreement (NCA) enforcement
generates spillover effects on sell-side analysts. Entry into this agreement reassigns ownership of
the client relationship from the brokerage house to individual brokers, potentially generating a
greater standard of care. Using a generalized difference-in-difference research design, we provide
evidence consistent with brokers reducing pressure on analysts to produce optimistic research
following protocol entry. This effect is concentrated among less experienced and non-All Star
analysts, who previously may have faced the greatest pressures to sacrifice objectivity.
Additionally, we find that analysts issue more accurate forecasts and generate reports with
heightened market reactions following protocol entry. Our collective evidence sheds new light on
how the nature of brokerage relationships can influence analysts’ research production.

Keywords: analysts, forecasts, bias, protocol

JEL Classification: G14, G24