Do Underwriting Activities Bias Investment Banks’ Buy Recommendations? Lessons from the Latin American Bond Markets

• Date: 01 December 2014
• DOI: http://doi.org/10.1111/infi.12057
• Published date: 01 December 2014
• Author: Sebastian Nieto‐Parra

Do Underwriting Activities

Bias Investment Banks’Buy

Recommendations? Lessons

from the Latin American

Bond Markets

♣

Sebastian Nieto-Parra

OECD Development Centre, 2, rue André-Pascal, Paris Cedex 16,

France 75775

Abstract

This paper examines whether banks’recommendations to investors in

the Latin Ameri can sovereign bon d market are trul y independent o f the

same banks’underwriting activities in these countries. The main find-

ingsarethefollowing.First,aftercontrollingformacroeconomicand

OECD Development Centre. I am especially grate ful to Olivier Accomin otti, Rolando Avendano,

Sandra Dow, Marc Flandreau, Ugo Panizza, Helmut Reisen , Javier Santiso and two anonym ous

referees for valuable comm ents on previous versions of t his paper. I am also grateful to seminar

participants at the LACEA meeting (Santiago, Chile, 2011) and at the Infiniti Conference on

International Finance (Dubl in, Ireland, 2010) for their comments a nd suggestions on a previous

version of this paper. I thank Thomas Dickins on for excellent research assistance provide d during

interviews in New York in 2006–07 with ma rket participants in the primary sovereign bond market.

The opinions expressed and a rguments employed in this arti cle are the sole responsibilit y of the

author and do not necessari ly reflect those of the OECD or of th e governments of its member

countries. All remainin g errors are the author’s responsibility.

International Finance 17:3, 2014: pp. 323–356

DOI: 10.1111/infi.12057

© 2014 John Wiley & Sons Ltd

financial variables, I find that investment banks tend to provide opti-

mistic recommendations to investors in sovereign bonds on which they

act as lead managers. Second, econometric analysis shows that this

association increases with the underwriting fee. These results are

consistent with interviews with the main institutional investors in

sovereign bonds . The findings of this research suggest the existence of

aconflict of interest between the origination and research departments

ofinvestmentbanksintheLatinAmericanbondmarket.

I. Introduction and Motivation

This paper studies two key activities of investment banks in the emerging sovereign

bond market: research and origination of international debt. These elements

contribute to information and price for mation and shed impor tant light on the

structure and efficiency of emerging bond markets.

Information provided by banks to investors m ay suffer from a bias stemming

from conflicting interests ( Boni and Womack 2002). Banks are confronted with a

trade-off in their recommendations: while sell-side activities that aim to build up

their client base provide a long-term incentive to build reputat ion by giving accurate

information, investment banks may be subject to short-term motivations to

recommend the purchase of securities in which they are participating as

underwriters.

The main finding of this resea rch confirmsthisbiasforthecaseoftheLatin

American bond market before the recent financial crisis . After controlling for

macroeconomic and financial variables, I find that underwriting activities are

associated with more optimistic recommendatio ns. In additi on, I find that favorable

recommendations provide d to investors in sovereign bonds are associated wit h

higher fees paid by issuers to underwriters. These results suggest there is a conflict

of interest between the origi nation and research departments of investment ban ks in

the Latin America n bond market. Intervie ws with main instit utional investors in

emerging sovereign bond s are consistent with these results.

This analysis adds a n ew dimension to the large bo dy of research on the structure

of emergi ng sovereig n bond ma rkets. Alt hough prev ious stud ies highlight the

existence of a conflict of i nterest in investment bank activitie s in emerging b ond

markets, they are limited to stylized or anecdota l facts. Nieto-Parra a nd Santiso

(2007) find a positive correlation bet ween underwrit ing activiti es and recommen-

dations. In referring to emerging ma rket debt crises, Calomiris (2003) notes that

‘investment banks’“research”departments cooperate with the masquerade

because not doing so means that they will be cut off from millions i n underwritin g

revenue from the new d ebt offerings or debt swaps’. Blustein (2003) notes that:

‘Big securities firms reaped nearly $1 billion in fees from underwriting

Argenti ne governme nt bonds duri ng the de cade 1991 –2001, and those firms’

324 Sebastian Nieto-Parra

© 2014 John Wiley & Sons Ltd

analysts were generally the ones producing the most bullish and influential reports

on the country’. Edwards (1997) addresses a number of questions regarding the

predictability of the 1994 Mexican crisis: ‘Should Wall Street ana lysts have known

that things were gettin g out of hand? And if they did, why didn’ttheyalerttheir

clients?’.

By contrast, e vidence for the par allel in th e equit y market i n develope d

countries is abundant (the next section provides an overvie w) and has drawn

the scrutiny of policy-makers. Between 2003 and 2004, the Securities and

Exchange Commis sion reached an agreement in the ‘Global Research Analyst

Settlement’with 12 key investment banks i n international capital markets. Th is

agreement required these banks to pay close to $1.5 billion in penalties, investor

education and funding indepe ndent research. In addition to these payments, the

agreement required ‘a clear separation of the research and investment banking

divisions at firms’.

The remainder of thi s paper is organize d as follows. Secti on II briefly reviews the

literature analyzing information problems in the equity market in developed

economies and high lights key dif ferences from the stru cture of t he emerg ing

sovereign bond market. Section III shows the functionin g of the underwr iting

market and emphasizes th e results of participa nts’opinions towards underw riters’

recommendations. Thi s section also presents the hy potheses and empirical strategy.

Section IV descr ibes the data used and presents st ylized facts about underwriters ’

recommendations. Section V provides the results a nd the robustness che cks. Section

VI concludes and outlines t he major policy implications.

II. Review of the Literature

There is a consensus that investm ent banks that serve as lea d managers for an

equity offering issue m ore optimistic and favorab le recommendations than inde-

pendent and unaffiliated institutions (e.g., Dugar and Nathan 1995; Lin and

McNichols 1998; Michaely and Womack 1999, 2005; Dechow et al. 2000; Madureira

2004; Agrawal and Chen 2008). Michaely and Womack (1999) use 200 IPOs and 360

recommendations documented in First Call during the period 1990–91 and find that

underwriters issued 67% of the buy recommendations in the first t wo months after

an IPO compared with 49% by non-underwriters.

Regarding the determinants of the recomme ndations is sued by investme nt

banks, Agrawal and Chen (2008) find that in the US equity market over the period

1994–2003, investment banks’revenues and commissions are both positive and

significant explanatory variables of recommendations net of the consensus (i.e., the

median of th e recommendations). Greater conflic ts with i nvestment banking

and brokerage lead an analyst to issue more optimistic recommendations on a

stock than the consensus. Conrad et al. (200 4) analyze the US equity market

© 2014 John Wiley & Sons Ltd

Do Underwriting Activities Bias Investment Banks’Recommendations? 325