Foreign market portfolio concentration and performance

• Date: 01 March 2020
• DOI: http://doi.org/10.1111/fima.12263
• Author: Sturla Lyngnes Fjesme
• Published date: 01 March 2020

DOI: 10.1111/fima.12263

ORIGINAL ARTICLE

Foreign market portfolio concentration and

performance

Sturla Lyngnes Fjesme

Oslo Business School at Oslo Metropolitan

University in Oslo, Oslo, Norway

Correspondence

SturlaLyngnes Fjesme, Oslo Business School at

OsloMetropolitan University in Oslo, Norway.

Email:sturla.fjesme@oslomet.no

Abstract

Using security holdings of 49,857 foreign investorson the Oslo Stock

Exchange (OSE), I test whether concentrated investment strate-

gies in international markets result in excess risk-adjusted returns.

I find that investors with higher learning capacity increase returns,

while investors with lower learning capacity decrease returns from

the portfolio concentration. I measure learning capacity as institu-

tional classification, geographical proximity to Norway,and cultural

closeness to Norwegian investors (as based on the Hofstede cul-

tural closeness measures). I conclude, consistent with the informa-

tion advantage theory, that concentrated investment strategies in

foreign markets can be optimal (disastrous)for investors with higher

(lower) learning capacity.

1INTRODUCTION

Traditionalasset pricing theory suggests that holding the market portfolio through international diversification opti-

mizes risk-return characteristics (Markowitz, 1952). In practice, we find that investors concentrate portfolios in a

small number of securities thereby missing out on obvious diversification benefits. More recent theoretical research

explainsthis apparent discrepancy using investor learning. When investors obtain information before they invest, port-

folio holdings are concentrated in assets with more information as this will optimize returns (Van Nieuwerburgh &

Veldkamp, 2009, 2010). Empirically,Choi, Fedenia, Skiba, and Sokolyk (2017) find a positive relation between foreign

market portfolio concentrationand risk-adjusted returns for higher learning capacity (smart) investors.

In this paper, I investigatewhether investors with lower learning capacity also improve returns from foreign mar-

ket portfolio concentrations. Obtaining data to investigatethis research question has proven difficult in the past as it

requires portfolio holdings for many investorsin a foreign market over a long period of time. In this paper, I investigate

all of the portfolios held by foreign (non-Norwegian) investors on the Oslo Stock Exchange (OSE)from January 1993

to July 2006 from the Central Depository (the OSE VPS). There are 38,776 unique foreign institutional investors and

11,081 unique foreign retail investors with a combined 1.5 million investor-month portfolio holdings from 152 differ-

ent countries. I identify foreign retail investorsfrom residential addresses and foreign institutional investors from main

office addresses.

c

2019 Financial Management Association International

Financial Management. 2020;49:161–177. wileyonlinelibrary.com/journal/fima 161

162 FJE SM E

My main empirical finding is that low learning capacity investors reduce risk-adjusted returns from foreign market

portfolio concentrations. I use two measures of learning capacity.First, I follow Hanley and Wilhelm (1995), Michaely

and Shaw (1994), and Aggarwal, Prabhala, and Puri (2002) and measure low learning capacity for the retail investor

classification(as opposed to institutional investor classification). I find that retail investors who increase foreign market

portfolio concentration by one standard deviation reduce annual risk-adjusted returns by –3.29%.This is significantly

lower than institutional investorswho increase annual risk-adjusted returns by +1.57% from increasing foreign market

portfolio concentration by one standard deviation.

In addition, I follow Bernile, Kumar, and Sulaeman (2009), Coval and Moskowitz (2001), and Baik, Kang, and Kim

(2010) and measure investor learning capacity based on investor geographicalproximity. Since there is quite an over-

lap in geographical and cultural proximity, I also measure learning capacity based on investor cultural closeness as

documented by Hofstede (2003, 2018) and Hofstede and Hofstede (2004). I find that investors who are distant from

Norway (lower learning capacity investors)reduce annual risk-adjusted returns by –1.13% by increasing portfolio con-

centration by one standard deviation. This is significantly lower than closer (higher learning capacity) investors who

increase annual risk-adjusted returns by +1.32% by increasing portfolio concentrationby one standard deviation.

I follow Choi et al. (2017) and measure portfolio concentrationas the sum of the absolute deviation in investor port-

folio weights from the market value weights. I investigate the relation between monthly portfolio excessreturns and

concentration while controlling for standard risk factors (RM-RF,SMB, HML, and Momentum), investor size (portfolio

values), the number of unique companies in the investor portfolios, investortypes, as well as various fixed effects (Choi

et al., 2017). Any relation between concentration and return is in excessof what is expected based on the differences

in portfolio values, investor types, the number of actual investments,portfolio risk characteristics, and year effects.

I makethree important contributions to the literature. First, I find that investors with arguably lower learning capac-

ity lose money from portfolio concentration. Choi et al. (2017) determine that investors with higher learning capacity

earn positive risk-adjusted-returns from foreign market portfolio concentration. These authors are limited to investi-

gating institutional investors with company holdings equal to or larger than 0.1% of the issued shares. I first replicate

the results in Choi et al. (2017) for a single market (the OSE). I then demonstrate that the investors not included in

Choi et al. (2017) (foreign retail investors) experience the opposite effect from portfolio concentration. This finding

extends Odean (1998a, 1998b) and Barber and Odean (2000) by suggesting that retail investors’ exhibit overconfi-

dence in portfolio formations. This finding also extends to Hanleyand Wilhelm (1995), Michaely and Shaw (1994), and

Aggarwal et al. (2002) by further documenting that institutional investors have a higher leaning capacity than retail

investors. This is an important contribution as it confirms that concentration is a poor strategy for investorswith low

learning capacity.

In addition, I contribute by confirming that cultural and geographical proximity are important indicators of infor-

mation learning in portfolio concentration. Choi et al. (2017) measure learning capacity based on investor types, such

as hedge fund versus mutual fund. I argue that investors that are geographically and/or culturallycloser to Norway

have a higher learningc apacityabout the Norwegian market than other foreign investors. I find that investors who are

geographically and/or culturally closer to Norwayearn higher returns than other foreign investors from concentrat-

ing portfolios on the OSE. I extend VanNieuwerburgh and Veldkamp (2009, 2010) and Choi et al. (2017) by providing

additional evidence linking portfolio concentration to returns for skilled investors.

Finally, I contribute to the literature investigating smart money. Gruber (1996) and Zheng (1999) find that

some smart investors are able to predict which funds will earn higher returns in the future (smart money). I con-

tribute to Gruber (1996) and Zheng (1999) by further documenting the existence of smart money.I determine that

smart investors increase while less informed investors decrease risk-adjusted returns from foreign market portfolio

concentration.

I only study those shares held on the OSE, so it is possible that investors hold other assets in addition to the portfo-

lios investigated. Toevaluate how the OSE portfolios contribute to broader held international portfolios, I also investi-

gate information ratios in addition to the portfolio returns. The information ratio evaluates whether investorsgener-

ate OSE portfolios that contribute positively to well-diversified international portfolios (Treynor& Black, 1973). The

results remain largely unchanged. Institutional investors increase, while retail investors decrease information ratios