A Glimpse Behind a Closed Door: The Long‐Term Investment Value of Buy‐Side Research and Its Effect on Fund Trades and Performance

• Published date: 01 June 2014
• Author: MICHAEL REBELLO,KELSEY D. WEI
• DOI: http://doi.org/10.1111/1475-679X.12042
• Date: 01 June 2014

DOI: 10.1111/1475-679X.12042

Journal of Accounting Research

Vol. 52 No. 3 June 2014

Printed in U.S.A.

A Glimpse Behind a Closed Door:

The Long-Term Investment Value of

Buy-Side Research and Its Effect on

Fund Trades and Performance

MICHAEL REBELLO

∗AND KELSEY D. WEI

†

Received 29 November 2012; accepted 8 January 2014

ABSTRACT

We examine proprietary research produced by buy-side analysts working for

a large fund management company. We find that the buy-side research has

investment value for a one-year horizon, and the analysts producing this re-

search exhibit differential ability that tends to persist over time. The buy-side

research strongly influences trades made by the company’s funds, especially

when it coveys information that is independent of the fund managers’ own in-

formation, when it is produced by buy-side analysts with good track records,

and when the underlying stocks have little sell-side coverage. The influence of

sell-side research is concentrated primarily in stocks not followed by buy-side

analysts and in funds with low overall buy-side coverage. The company’s funds

that rely more heavily on buy-side research generate superior performance.

∗Naveen Jindal School of Management, The University of Texasat Dallas †Federal Reser ve

Board, Division of Research and Statistics.

Accepted by Philip Berger.For useful comments, we would like to thank an anonymous ref-

eree, Ashiq Ali, Jonathan Clarke, Steven Crawford, Tom George, Zuhair Khan, Stan Markov,

Amit Seru, and seminar participants at the 2011 Lone Star Conference, 2013 American Fi-

nance Association meeting, 2013 Financial Management Association Asian meeting, and 2013

Financial Management Association Annual meeting. All errors are our own. The views pre-

sented in this paper are solely those of the authors and do not necessarily represent those of

the Federal Reserve Board or its staff.

775

Copyright C, University of Chicago on behalf of the Accounting Research Center,2014

776 M.REBELLO AND K.D.WEI

1. Introduction

By some estimates, fund managers place greater reliance on research from

buy-side analysts, who work exclusively for them, than they do on research

from sell-side analysts, who typically work for brokerage firms. For example,

Cheng, Liu, and Qian [2006] find that fund managers rely more heavily

on buy-side research than on sell-side or independent research. Moreover,

Johnson [2006] estimates that fund management companies in the United

States and the United Kingdom spend more on buy-side research annually

than on sell-side research (USD 7.7 billion vs. USD 7.0 billion). Despite

the importance fund managers place on buy-side research, in contrast to

the large number of studies that have investigated sell-side research (Beyer

et al. [2010]), only a handful of studies have investigated the investment

value of buy-side research, its use by fund managers, or its effect on fund

performance.1Moreover, there is also no systematic analysis of the value of

buy-side research over horizons that match those of typical fund managers.

This void is especially important since fund managers can only benefit from

superior information that matches their investment horizons.2

In this paper, we contribute to the literature on buy-side research by ex-

amining the investment value of research produced by a sample of buy-side

analysts over a one-year investment horizon, its use by fund managers, and

its effect on fund performance.3This research was generated between 1994

and 2008 by analysts working for a large global fund management company

(the company) that manages well over $300 billion of assets, and is consis-

tently rated among America’s largest money managers by Institutional In-

vestor. The buy-side analysts provide all the company’s fund managers with

exclusive access to research on large-capitalization U.S. stocks that are fol-

lowed by more than 10 sell-side analysts on average, and are the only source

of in-house research on these stocks. They are compensated based solely on

the investment value of their research, which is assessed through a combina-

tion of evaluations by the company’s fund managers and the performance

1See, for example, Cheng, Liu, and Qian [2006], Kacperczyk and Seru [2007], Groysberg,

Healy, and Chapman [2008], Frey and Herbst [2011], and Groysberg et al. [2013].

2Consistent with an investment horizon longer than one year, the typical mutual fund

turnover ratio is around 70% (see, e.g., Nanda, Wang, and Zheng [2004], Wermers, Yao,

and Zhao [2012]). Perhaps due to sell-side analysts’ incentives to promote short-term trad-

ing (Irvine [2004], Jackson [2005], Cowen, Groysberg, and Healy [2006], Agrawal and Chen

[2008], Groysberg, Healy, and Maber [2011]), investments based on sell-side research gener-

ate abnormal returns only for horizons of at most a few months (Womack [1996], Brav and

Lehavy [2003], Jegadeesh et al. [2004]).

3While we find that buy-side research helps predict future abnormal returns over horizons

ranging from one month to one quarter, this return predictive power seems to be largely a

function of past price movements. Buy-side research also predicts returns beyond a one-year

horizon. However, these relations tend to be less stable.

LONG-TERM INVESTMENT VALUE OF BUY-SIDE RESEARCH 777

of test portfolios based on their research.4Since the company’s funds have

a large-cap value-oriented investment objective and moderate turnover ra-

tios (typically around 60% a year), the analysts have little incentive to focus

on information that is valuable only for short-term trading. Therefore, our

data provide us with a unique opportunity to examine analysts’ ability to

produce research that matches the investment horizons of their affiliated

fund managers, and assess the impact of such research on fund trades and

performance.

We focus our analysis on the primary research output of our sample of

buy-side analysts: an estimate of the “excess return” for each stock they

cover. This excess return, which we refer to as buy-side rating, is the dif-

ference between an analyst’s expectation of a stock’s return based on her

estimate of its fundamental value and the discount rate she uses to value

the stock.5Despite their extensive exposure to sell-side research, the buy-

side analysts produce research that is very different from sell-side research

on these stocks: The buy-side ratings are decreasing in consensus sell-

side recommendations, recommendations from sell-side analysts working

for top brokers, and recommendations from contrarian sell-side analysts.

Buy-side rating changes are similarly negatively correlated with alternative

measures of sell-side recommendation changes. Moreover, in stark con-

trast to the positive relation between sell-side analyst recommendations and

stock returns documented in prior research (see, e.g., Bradshaw [2004],

Jegadeesh et al. [2004]), buy-side analysts tend to assign higher ratings to

stocks with lower past returns and operating performance.

The buy-side ratings and their changes have significant investment value

over a one-year horizon, which is economically meaningful both before and

after we adjust for risk and stock characteristics. On average, stocks in the

top buy-side rating quintile generate a Daniel et al. [1997] (henceforth

“DGTW [1997]”) characteristic-adjusted abnormal return that is 4.51%

higher than that on stocks in the bottom quintile over the following year.

Similarly, zero-investment portfolios with long positions in stocks with large

rating increases relative to a year ago and short positions in stocks with

large rating decreases over the same period, generate an average abnor-

mal return of 2.69% in the year after portfolio formation. The investment

values of both buy-side ratings and buy-side rating changes remain econom-

ically and statistically robust when we control for stock returns and sell-side

recommendation changes. We also find that analysts with superior past per-

formance continue to significantly outperform their peers in the following

4The investment company studied by Groysberg et al. [2013] uses a similar compensation

structure for its buy-side analysts.

5All the analysts working for the company have to use the same proprietary fundamental

valuation model, which uses estimates of future earnings and profitability made by individual

analysts as inputs. To maintain consistency in valuations, some of the inputs are not made by

individual analysts covering the stock but by a team of senior analysts.