Does one model fit all in global equity markets? Some insight into market factor based strategies in enhancing alpha

• Published date: 01 July 2019
• Author: Subhransu S. Mohanty
• Date: 01 July 2019
• DOI: http://doi.org/10.1002/ijfe.1710

RESEARCH ARTICLE

Does one model fit all in global equity markets?

Some insight into market factor based strategies in

enhancing alpha

Subhransu S. Mohanty

1,2

1

Faculty of Finance, St. Francis Institute

of Management and Research, Mumbai,

India

2

President Emeritus of SMART

International Holdings,Inc. in Delaware,

USA

Correspondence

Subhransu S. Mohanty, St. Francis

Institute of Management and Research,

Mumbai, India.

Email: drssmohanty@gmail.com;

director@sfimar.org; drssmohanty

@smartinternationalholdings.org

JEL Classification: G11; G12; G14; G15;

G17

Abstract

The sources of risk in a marketplace are systematic, cross‐sectional, and time

varying in nature. Though the capital asset pricing model (CAPM) provides an

excellent risk–return framework and the market beta may reflect the risk asso-

ciated with risky assets, there are opportunities for investors to take advantage

of dimensional and time‐varying return anomalies in order to improve their

investment returns. In this paper, we restrict our analysis to return variations

linked to market factor anomalies or factor or dimensional beta using the

Fama–French three‐factor; Carhart four‐factor; Fama‐French five‐factor; and

Asness, Frazzini, and Pederson (AFP)'s five‐and six‐factor models. We find sig-

nificant variations in explaining sources of risk across 22 developed and 21

emerging markets with data over a long period from 1991 to 2016. Each market

is unique in terms of factor risk characteristics, and market risk as explained by

the CAPM is not the true risk measure. Hence, contrary to the risk–return effi-

ciency framework, we find that lower market risk results in higher excess return

in 19 out of the 22 developed markets, which is a major anomaly. However,

although in majority of the markets, the AFP models result in reducing market

risk (15 countries) and enhancing alpha (11 countries), it is also very interesting

to note that the CAPM is second only in generating excess returns in the devel-

oped markets. We are conscious of the fact, however, that each market is unique

in its composition and trend even over a long time horizon, and hence, a gener-

alized approach in asset allocation cannot be adopted across all the markets.

KEYWORDS

betting against beta, conservative‐minus‐aggressive, high‐minus‐low,quality‐minus‐junk, robust‐

minus‐weak, small‐minus‐big

1|INTRODUCTION

Investments in global equity markets are driven by

expected return, the variance of such return, and the level

of informational efficiency. Because these factors are

future‐oriented and uncertain, according to Hicks,

expected returns from investments include an allowance

for risk (Hicks, 1939).

1

This risk varies from security to

security or from market to market. The efficient market

hypothesis (EMH) or informational efficiency in markets

is the basic premise upon which expected return and risk

framework was built. It propounds that at any point in

time, market prices of assets truly reflect all the informa-

tional content, and performance of an asset cannot be

Received: 5 November 2017 Revised: 19 August 2018 Accepted: 9 September 2018

DOI: 10.1002/ijfe.1710

1170 © 2018 John Wiley & Sons, Ltd. Int J Fin Econ. 2019;24:1170–1192.wileyonlinelibrary.com/journal/ijfe

predicted based on its historical prices, as they behave

randomly. Hence, such prices are fair to the extent that

an investor cannot earn abnormal returns either through

stock selection or market timing. The Sharpe–Lintner

capital asset pricing model (CAPM), by far the most sig-

nificant development in modern capital market theory,

is based on the above assumptions within Markowitz's

(1959, 1991) mean–variance optimization framework.

However, both the random walk theory and CAPM

are marred with criticism. The former with its indepen-

dent incremental assumption (Cootner, 1964) and sta-

tionaries (Osborne, 1959), and the latter with empirical

findings of Black, Jensen, and Scholes that demonstrated

that “low beta”assets earn a higher return on average

and “high beta”assets earn a lower return on average

(Black, 1972), which are against the CAPM return–risk

framework that the expected excess return from holding

an asset is proportional to the covariance of its return

with the market portfolio (its “beta”). Further, it also

assumes that investors have homogeneity of return

expectations, which according to Tobin (1958)

2

is not

true, as he observes, “for a given amount of risk, an

investor always prefers a greater to a smaller expectation

of return.”The utility theorists, behavioural finance

researchers, and psychologists have a number of expla-

nations to investors' varied risk and return preferences.

Second, the model also assumes a single period, that is,

the investment opportunity sets do not change over

time, though in reality, they do. Intertemporal studies

show that risk and return associated with assets and

markets, as well as their correlations, change over time.

Despite the above, the CAPM provides a strong basis of

the relationship among asset returns and explains a

significant fraction of the variation in asset returns

(Merton, 1973).

Continued academic and nonacademic empirical

research shows that there are many anomalies to counter

the risk–return efficiency of assets and their markets.

Merton (1973) says that up to four unspecified state

variables lead to risk premiums that are not captured by

the market factor. Because these unspecified state vari-

ables have not been identified and measured, the later

empirical studies mostly deal with excess return (alpha)

generation through factor portfolios providing different

combinations of exposures to the unknown state vari-

ables within the relevant multifactor efficient set along

with the market portfolio and the risk‐free asset (Fama

& French, 2015). Notable among them are the Fama–

French three‐factor model, the Carhart four‐factor model,

the Fama–French five‐factor model, and the Asness and

Frazini's six‐factor model. All these models are highly

intuitive and provide additional cross‐sectional risk–

return dimensions to the market risk. These factors are

size, value, momentum, profitability, investment, quality,

and low beta.

It was observed that small companies are considered

more risky (the size effect) than the big ones, as they

are less liquid, and companies with a high book‐to‐mar-

ket price ratio (the value effect) generally outperform

companies with a low book‐to‐market price ratio. More-

over, stock returns have a certain momentum, as some

findings show that stock returns tend to exhibit positive

autocorrelation in the short to medium term, stocks that

have performed well in the past, generally perform well

in the future, whereas stocks that have performed poorly,

generally perform poorly. Similarly, more profitable

companies (profitability premium) are expected to have

a higher valuation compared with the less profitable ones

and high book equity growth (investment premium)

means a lower valuation growth. In the most recent

models, the quality factor (premium) is being used, which

adds a few more parameters to profitability, such as,

growth (higher price for stocks with growing profits),

safety (both return‐based measure of safety, i.e., volatility

risk relative to market risk and fundamental‐based

measures [such as stocks with low leverage, low volatility

of profitability, and low credit risk]), and payout ratio

(higher payout means less of management agency prob-

lems; Jensen, 1986). Additionally, the low‐risk anomaly

has been further substantiated with more findings on

liquidity preference (Harvey, Lundblad, & Bekaert,

2007), liquidity funding risk (Acharya & Pedersen,

2005), and portfolio constraints in a generic market

setting, which shows that low‐beta stocks have high

expected returns.

All the above factors are cross‐sectional but interde-

pendent in nature, change over time, and maybe useful

in defining country or market‐specific systematic (mar-

ket) risk and factor (dimensional idiosyncratic) risk. In

this paper, we attempt to analyse country‐specific idio-

syncratic risk with the help of factor alphas and the risks

associated with them, so that in a global equity invest-

ment setting, investors will be in a position to differenti-

ate between equity markets of the various countries and

position them with a strategy to maximize their portfolio

performance as well as reduce or monitor risk.

2|DATA SET

We have used MSCI Global Equity Markets Standard

Price Monthly Index Data (ACWI and its country compo-

nents) from the date of availability till December 2016 for

all developed markets, from January 1991, except Israel

in which case it is from January 1993; in the case of

emerging markets, from January 1991 for Brazil, Chile,

MOHANTY 1171