Brokers versus Retail Investors: Conflicting Interests and Dominated Products

• DOI: http://doi.org/10.1111/jofi.12763
• Published date: 01 June 2019
• Date: 01 June 2019
• Author: MARK EGAN

THE JOURNAL OF FINANCE •VOL. LXXIV, NO. 3 •JUNE 2019

Brokers versus Retail Investors: Conflicting

Interests and Dominated Products

MARK EGAN∗

ABSTRACT

I study how brokers distort household investment decisions. Using a novel convert-

ible bond data set, I find that consumers often purchase dominated bonds—cheap and

expensive otherwise-identical bonds coexist in the market. Brokers are incentivized

to sell the dominated bonds, typically earning two times greater fees for selling them.

I develop and estimate a broker-intermediated search model that rationalizes this

behavior. The estimates indicate that costly search is a key friction in financial mar-

kets, but the effects of search costs are compounded when brokers are incentivized to

direct the search of consumers toward high-fee inferior products.

THE PRICES AND FEES OF SEEMINGLY identical financial products often differ dras-

tically. Previous research documents price heterogeneity across mutual funds,

mortgages, bonds, and other financial products.1What drives this behavior?

Hortac¸su and Syverson (2004) highlight the importance of search in a con-

sumer’s investment decision process. But consumer search does not happen

in a vacuum. Broker intermediation plays a critical role in a consumer’s in-

vestment decision and search process. In 2010, 56% of American households

sought investment advice from a financial professional (Survey of Consumer

Finances [available at https://www.federalreserve.gov/econres/scfindex.htm]).

∗Mark Egan is at Harvard Business School, I am grateful for the comments received from

Eric Budish; Clare C´

el´

erier; Adam Copeland; Ali Hortac¸ su; Juhani Linnainmaa; NealeMahoney;

Gregor Matvos; Robert McDonald; Salvatore Miglietta; Tobias Moskowitz; Casey Mulligan; Derek

Neal; Tomas Philipson; Monika Piazzesi; Amit Seru; Norman Sch¨

urhoff; Jesse Shapiro; Amir

Sufi; Chad Syverson; Frank Wolak; the members of the University of Chicago Booth School of

Business Fama Miller Corporate Finance Reading Group; as well as seminar participants at

AQR Captial Management, the 2016 AFA Meetings, the 2016 EFA Meetings, the 2016 FIRS

Conference, Harvard Business School, the 2016 International Industrial Organization Confer-

ence, the 2014 North American Econometric Society Meetings, NYU Stern School of Business,

the University of Chicago, the University of Chicago Booth School of Business, the University

of Minnesota Carlson School of Management, the University of Pennsylvania, the University of

Texas McCombs School of Business, and the 2015 Washington University in St. Louis Graduate

Student Conference. The author has read the Journal of Finance’s disclosure policy and has no

conflicts of interest to disclose.

1For example, see Hortac¸su and Syverson (2004) and Elton, Gruber, and Busse (2004)forthe

mutual fund industry; Gurun, Matvos, and Seru (2016) and Allen, Clark, and Houde (2014,2018)

for mortgages; Green, Hollifield, and Schc¸rhoff(2007) for bonds; Duarte and Hastings (2012)for

privatized social security plans; Christoffersen and Musto (2002) for money funds; and Brown and

Goolsbee (2002) for life insurance.

DOI: 10.1111/jofi.12763

1217

1218 The Journal of FinanceR

Despite their prevalence, however, brokers—commonly referred to as financial

advisors—may not be acting in the best interest of their clients. Brokers in the

United States are not held to a fiduciary standard, which means that they are

not legally obligated to act in a client’s best financial interests. A broker may

therefore legally choose to subordinate his client’s interests to his own financial

interests by directing his client to inferior products with high brokerage fees.

Former President Barack Obama (2015) called for regulators to “update the

rules and requirements that retirement advisors put the best interests of their

clients above their own financial interests. It’s a very simple principle: You want

to give financial advice, you’ve got to put your client’s interests first.” Using a

novel data set that precisely measures the conflicts of interest between brokers

and consumers in a clean setting, in this paper I illustrate how conflicts of

interest in the brokerage industry distort investment decisions and compound

search/information frictions in retail financial markets.

The paper has two goals. The first is to use a unique clean setting to show

that brokers distort the investment decisions of consumers. I show that brokers

play a critical role in determining consumers’ investment outcomes and are in-

centivized to sell inferior products. Accounting for the incentives of financial

intermediaries is of first-order importance in understanding investment flows

for the financial products studied in this paper. The second goal is to rational-

ize the behavior of brokers and consumers in equilibrium by developing and

estimating a broker-intermediated search model. In the model, brokers facili-

tate consumers’ search process by selecting financial products to offer to each

consumer. In equilibrium brokers maximize expected brokerage commissions

by selecting products based on product quality (risk and return), commissions,

and the consumer’s level of sophistication.2I then use the model to quantify

the distortions in the market due to heterogeneity in consumer sophistication,

search costs, and broker incentives.3

Several data challenges complicate investigating the above questions. First,

data on how brokers are compensated are scarce. The product studied most

extensively in the literature are mutual funds, but there are limited data on

how brokers are compensated for selling mutual funds. Mutual funds report

12b-1 fees, which are used to cover aggregate marketing, distributional, and

service expenses, but 12b-1 fees do not tell us how brokers/intermediaries are

specifically compensated. Additionally, since mutual fund fees are paid by the

investor, a change in 12b-1 fees directly impacts the incentives of both brokers

and consumers, which complicates identification.4I address these challenges

2A handful of papers highlight the importance of consumer sophistication and trust in fi-

nancial markets, including: Campbell (2006), Guiso, Sapienza, and Zingales (2008), Carlin and

Manso (2011), Gennaioli, Shleifer,and Vishny (2015), Garleanu and Pedersen (2018), and Agarwal,

Ben-David, and Yao (2017).

3This paper relates to the growing literature on regulating consumer financial products, includ-

ing Agarwal et al. (2009), Campbell et al. (2010,2011), and Agarwal, Chomsisengphet, Mahoney,

and Stroebel (2014).

4Improved fee data from mutual fund N-SAR filings ameliorate some of these issues, but the

N-SAR data still only report the average fees paid and do not provide full details on revenue sharing.

Brokers versus Retail Investors 1219

by constructing a new retail bond data set. In the data set broker compensation

comes in the form of a “kick-back” that is paid by the bond issuer to the broker.

Because consumers do not pay the kick-back, conditional on the other product

characteristics, consumers are indifferent to the commission/kick-back. This

unique feature of the data allows me to separately identify the preferences of

brokers and consumers.

Second, a related challenge is that it can be difficult to directly compare and

establish a definitive rank ordering of financial products. Financial products,

such as mutual funds, vary across several observable and unobservable dimen-

sions, making direct comparisons of products challenging. For example, we may

think that Vanguard’s S&P 500 fund, which charges 0.17% a year, is better than

Guggenheim’s Rydex S&P 500 fund, which charges 2.28% a year, but product

differentiation and preference heterogeneity make such claims tenuous (Del

Guercio and Reuter (2014)).5Moreover, attempting to draw any inference from

variation in fees across funds is confounded by other product characteristics.

Fees/commissions are endogenous and are almost certainly correlated with the

unobserved product characteristics of the funds. I construct a new bond data set

in which the products are relatively simple and are characterized completely

by a few observable features so that a rank ordering of financial products is

indeed possible.

I address these data challenges by constructing a new retail bond data set

covering all one-year reverse convertible bonds issued in the United States over

the period 2008 to 2012. A reverse convertible is similar to a standard fixed-

rate bond in that it pays fixed monthly/quarterly coupons based on the initial

investment, but differs in that the final principal payment is convertible into

shares of some prespecified reference equity. At maturity, bondholders receive

either their full principal or a fixed number of shares of the reference equity

if the reference equity share price falls below some predetermined convertible

threshold. Reverse convertibles are typically purchased by wealthy individuals

and households with incomes in excess of $100,000, liquid assets in excess of

$100,000, and total net worth of at least $250,000.6

The advantage of studying reverse convertible bonds over other financial

products is twofold. First, reverse convertibles are characterized completely

by a small number of dimensions, namely, a fixed coupon and an equity-linked

principal payment. As a result, it is easy to locate simultaneously issued reverse

convertibles for which the payout of one reverse convertible is unambiguously

dominated by another—the bond with the higher coupon. Consider two nearly

identical one-year reverse convertibles issued by JPMorgan Chase on June 30,

5Randall, David, January 13, 2014, Analysis: High-priced index funds? The worst deal for in-

vestors, Reuters. Available at: https://www.reuters.com/article/us-indexfunds-costs-analysis/ana-

lysis-high-priced-index-funds-the-worst-deal-for-investors-idUSBREA0C0N920140113 (accessed

December 28, 2016).

6Royal Bank of Canada deems reverse convertibles as “suitable” for investors meet-

ing this criterion. Source: Schafer, Lee, April 24, 2015, Buy a reverse convertible?

Sure! What is it?” The Star Tribune. Available at: http://www.startribune.com/buy-a-reverse-

convertible-sure-what-is-it/301260161/ (accessed April 20, 2016).