EXPLOITING CLOSED‐END FUND DISCOUNTS: A SYSTEMATIC EXAMINATION OF ALPHAS

• DOI: http://doi.org/10.1111/jfir.12122
• Published date: 01 June 2017
• Author: Yangru Wu,Louis R. Piccotti,Dilip Patro
• Date: 01 June 2017

EXPLOITING CLOSED-END FUND DISCOUNTS:

A SYSTEMATIC EXAMINATION OF ALPHAS

Dilip Patro

Federal Deposit Insurance Corporation

Louis R. Piccotti

University at Albany, State University of New York

Yangru Wu

Rutgers Business School

Abstract

We systematically study the value of the information contained in closed-end fund

(CEF) premiums. We parametrically estimate CEF expected returns as a function of the

history of CEF premiums, in addition to the current premium, and buy the quintile of

funds with the highest expected returns and sell the quintile of funds with the lowest

expected returns. The return on this portfolio suggests that previous studies, which

examine the information in current premiums only, have understated the value of the

information in premiums. Our strategy values the information in the history of CEF

premiums at an annualized return of 18.2%.

JEL Classification: G01, G11, G12

I. Introduction

Closed-end funds (CEFs) are investment companies that issue a f‌ixed number of shares

and invest the proceeds based on the objective of the fund. The shares of a CEF are traded

on a stock exchange similar to common stock and unlike an open-end fund cannot be

redeemed by the shareholders at its net asset value (NAV). In eff‌icient and frictionless

markets, the share price at which a fund trades must equal its NAV. In reality, however,

share prices are often signif‌icantly below their respective NAVs (termed the CEF

discount puzzle). Furthermore, the difference between share prices and NAVs, referred

to as the premium,

1

exhibits large time-series and cross-sectional variation. A large body

We thank the editor, two anonymous referees, Ren-raw Chen, Bjorn Flesaker, George Jiang, Andreas

Michlmayr, Robert Patrick, David Smith, Daniel Weaver, Fei Wu, Feng Zhao, the editors, two anonymous referees,

and seminar and conference participants at Rutgers University, Central University of Finance and Economics, City

University of New York, and the International Symposium on Financial Engineering and Risk Management for

helpful conversations and comments. Part of this work was completed while Wu visited the Chinese Academy of

Finance and Development of the Central University of Finance and Economics, whose hospitality was greatly

appreciated. The views expressed in this article are those of the authors and do not necessarily ref‌lect those of their

institutions. We are responsible for any remaining errors.

1

A fund’s discount is the negative of its premium.

The Journal of Financial Research Vol. XL, No. 2 Pages 223–248 Summer 2017

223

© 2017 The Southern Finance Association and the Southwestern Finance Association

RAWLS COLLEGE OF BUSINESS, TEXAS TECH UNIVERSITY

PUBLISHED FOR THE SOUTHERN AND SOUTHWESTERN

FINANCE ASSOCIATIONS BY WILEY-BLACKWELL PUBLISHING

of research has tried to explain this puzzling behavior. Leading explanations include

investor sentiment effects (see, e.g., De Long et al. 1990; Lee, Shleifer, and Thaler 1991),

open-ending frictions (see, e.g., Brickley and Schallheim 1985; Bradley et al. 2010;

Brauer 1988), agency costs (see, e.g., Barclay, Holderness, and Pontiff 1993; Khorana,

Wahal, and Zenner 2002; Del Guercio, Dann, and Partch 2003), managerial skills (Chay

and Trzcinka 1999; Coles, Suay, and Woodbury 2000; Johnson, Lin, and Song 2006;

Berk and Stanton 2007), and market segmentation (see, e.g., Bonser-Neal et al. 1990;

Bodurtha, Kim, and Lee 1995; Gemmill and Thomas 2002; Nishiotis 2004; Cherkes,

Sagi, and Stanton 2009; Froot and Ramadorai 2008; Elton et al. 2013).

Thompson (1978) f‌inds that portfolios of CEFs trading at discounts outperform

the market and Pontiff (1995) shows that funds with premiums accrue negative

abnormal returns and funds with discounts accrue positive abnormal returns. The

traditional long–short strategy, which sorts on the observed premium alone, ignores

the informationcontent that past CEF premiums contain, however. Therefore,if the history

of premiums contain valuableinformation, previous studies have understated the value of

information contained in premiums. We contribute to the literature by using a parametric

method to optimallyexploit the information containedin the history of CEF premiums and

premium innovations. Our f‌irst model relieves the traditional assumptions of Thompson

(1978) and Pontiff(1995) that expected returns are exactly the negative of the currentCEF

premium and allows the coeff‌icient on CEF premium to be freely estimated. Our second

model further incorporatesthe optimally chosen history of CEF premiumchanges (similar

to the Dickey–Fuller, 1981, model) to better match the CEF return-data-generating

process. By estimating thesetwo new models of CEF premiums, we are able to gauge the

value of the information content contained in current premiums, as well as in lagged

premiums, when we form portfolios sorted on our model-predicted expected returns.

First, we f‌ind that the majority of funds exhibit signif‌icant mean reversion in the

premium. Our analysis shows that the bias-adjusted speed for mean reversion to

equilibrium premium levels is 8.6% per month, implying an average half-life of

7.7 months for all the funds in our sample. Furthermore, there exists large cross-sectional

variation in the reversion speeds, indicating substantial heterogeneity across funds. In

general, funds investing in f‌ixed-income securities have faster speeds of reversion than

funds investing in equities. International funds exhibit more signif‌icant evidence of mean

reversion than funds invested domestically. Even for funds within the same fund type,

there exists large cross-sectional heterogeneity in premium mean reversion speeds. Early

explanations for why fund premiums should mean revert are provided by the noise trader

model of De Long et al. (1990) and the investor sentiment hypothesis of Lee, Shleifer,

and Thaler (1991). Alternatively, premiums should also display rational mean reversion

as a result of time-varying contingent liabilities as evidenced in the f‌indings of Malkiel

(1977), Chay, Choi, and Pontiff (2006), and Day, Li, and Xu (2011).

We design a trading strategy that optimally exploits the information content of

CEF premiums by using our two parametric models and then by buying the quintile of

funds with the highest expected returns and selling the quintile of funds with the lowest

expected returns. The returns to this strategy implicitly measure the value contained in

current CEF premiums only and in the history of CEF premium innovations at an

annualized mean return of 17.3% and 18.2%, respectively. Furthermore, commonly used

224 The Journal of Financial Research