Cream‐Skimming in Financial Markets

• Date: 01 April 2016
• DOI: http://doi.org/10.1111/jofi.12385
• Author: TANO SANTOS,JOSE A. SCHEINKMAN,PATRICK BOLTON
• Published date: 01 April 2016

THE JOURNAL OF FINANCE •VOL. LXXI, NO. 2 •APRIL 2016

Cream-Skimming in Financial Markets

PATRICK BOLTON, TANO SANTOS, and JOSE A. SCHEINKMAN∗

ABSTRACT

We propose a model in which investors may choose to acquire costly information

that identifies good assets and purchase these assets in opaque (OTC) markets. Un-

informed investors access an asset pool that has been cream-skimmed by informed

investors. When the quality composition of assets for sale is fixed, there is too much

information acquisition and the financial industry extracts excessive rents. In the

presence of moral hazard in origination, the social value of information varies in-

versely with information acquisition. Low quality origination is associated with large

rents in the financial sector. Equilibrium acquisition of information is generically

inefficient.

WHAT DOES THE FINANCIAL industry add to the real economy? What is the optimal

organization of financial markets, and how much talent is required in the

financial industry? We revisit these fundamental questions in light of recent

events and criticisms of the financial industry. The core issue underlying these

questions is whether the financial industry extracts excessively high rents

from the provision of financial services and whether these rents attract too

much talent.1Figure V, Panel B, from Philippon and Reshef (2012, p. 1569)

plots the evolution of U.S. wages (relative to average nonfarm wages) for three

subsegments of the financial services industry: credit, insurance, and “other

finance.” “Credit” refers to banks, S&Ls, and similar institutions; “insurance” to

life and P&C insurers and “other finance” to the financial investment industry

and investment banks. The most remarkable finding is that the bulk of the

∗Patrick Bolton and Tano Santos are with Columbia University and NBER. Jose A. Scheinkman

is with Columbia University, Princeton University, and NBER. This article was previously circu-

lated under the title “Is the Financial Sector TooBig?” We thank seminar participants at the Bank of

Spain, Cambridge University, CEMFI, EHESS, HEC-Paris, HKUST, INSEAD, London Business

School, London School of Economics, MIT-Sloan, Rochester, Rutgers Business School, Studien-

zentrum Gerzensee, Toulouse School of Economics, University of Virginia-McIntire, Washington

University-Olin Business School, Wharton, and the 6th Banco de Portugal Conference on Mone-

tary Economics for their feedback and suggestions. We also thank Guido Lorenzoni and Marzena

Rostek for detailed comments and Rishab Guha and Linan Qiu for research assistance. This paper

was awarded the 2011 Jaime Fern´

andez Araoz Prize. Bolton acknowledges financial support from

Columbia University and the Fondation Louis Bachelier, Santos acknowledges financial support

from Columbia University, and Scheinkman from Columbia University and Princeton University.

All authors declare no conflict of interest related to this article.

1Goldin and Katz (2008) document that the percentage of male Harvard graduates with posi-

tions in finance 15 years after graduation tripled from the 1970 to the 1990 cohort, largely at the

expense of occupations in law and medicine.

DOI: 10.1111/jofi.12385

709

710 The Journal of Finance R

growth in remuneration in the financial services industry took place in the

“other finance” sector.

In this paper, we attempt to explain the outsized remuneration in this latter

sector by modeling a financial industry that consists of two sectors: an orga-

nized, regulated, standardized, and transparent market where most retail (i.e.,

plain vanilla) transactions take place, and an informal, opaque sector, where in-

formed professional investors trade and bespoke services are offered to clients.

We refer to the transparent, standardized markets as organized exchanges,

and we refer to the opaque sector as the over-the-counter (OTC) sector even

though in practice not all OTC markets are opaque (e.g., markets for foreign

exchange are quite transparent). A central idea we develop is that, although

OTC markets provide special valuation services to sellers of securities, their

opacity also allows informed dealers to extract informational rents. What is

more, OTC markets cannibalize organized exchanges by “cream-skimming”

the juiciest deals away from retail investors.2And, informed investors don’t

just get access to the best assets—they also benefit from the cream-skimming

activities of other informed dealers that help them extract better terms from

the asset sellers. The reason is that cream-skimming worsens the quality of

the pool of assets flowing into the exchange, thereby lowering the price a seller

of a good asset can obtain in the exchange as an outside option. This is how

informed dealers are able to extract outsized terms from the sellers of good

assets. Cream-skimming and outsized remuneration are two sides of the same

coin. Moreover, the outsized informational rents in OTC markets attract tal-

ent to the financial industry, enhancing the cream-skimming externality and

further increasing the information rents extracted by informed dealers. In this

way, growth in OTC markets does not lead to the dissipation of these rents. On

the contrary, dealer remuneration rises as the OTC sector grows, even when

growth is not due to increased demand.

The role of the financial sector in our model is to provide liquidity by allowing

asset originators to sell their assets to investors. A related role is the valuation

of assets for sale when assets vary in quality. This valuation service is offered

by informed dealers in OTC markets. However,although originators are willing

to pay dealers to help them identify valuable assets, in our model there is no

social benefit from these valuation services unless more accurate valuations

also serve the role of providing better origination incentives; otherwise, costly

information acquisition by professional investors only serves the purpose of

cream-skimming the best assets, which is valuable to the trading parties but

has no social value. This result is related to the general observation first made

by Hirshleifer (1971) about the private value of “foreknowledge information,”

2Rothschild and Stiglitz (1976) considered a different form of cream-skimming in insurance

markets with adverse selection. In that setting, insurers are uninformed about risk types, but offer

contracts that induce informed agents to self-select into insurance contracts. For an application

of the Rothschild-Stiglitz framework to competition among organized exchanges, see Santos and

Scheinkman (2001).