How does the volatility‐timing strategy perform in mutual funds portfolios

• Published date: 01 March 2023
• Author: Zhida Yin,Jilin Jiang,Zongxin Qian
• Date: 01 March 2023
• DOI: http://doi.org/10.1111/irfi.12387

ORIGINAL ARTICLE

How does the volatility-timing strategy perform

in mutual funds portfolios

Zhida Yin

1

| Jilin Jiang

2

| Zongxin Qian

3,4

1

International Monetary Institute, Renmin

University of China, Beijing, China

2

Global Decision Science Department,

American Express, New York, USA

3

School of Finance, Renmin University of

China, Beijing, China

4

China Financial Policy Research Center,

Renmin University of China, Beijing, China

Correspondence

Zongxin Qian, International Monetary

Institute, Renmin University of China,

No. 59 Zhongguancun Street, Haidian District,

Beijing, China, 100872.

Email: qianzx@ruc.edu.cn

Funding information

National Natural Science Foundation of China,

Grant/Award Numbers: 71773126, 71850009

Abstract

Literature suggests that a volatility-timing strategy improves

the performance of factor portfolios in the stock market

and currency carry trade. This paper shows that the perfor-

mance of this strategy is mixed when applied to mutual

fund portfolios. More specifically, its performance not only

depends on the investment style of the mutual funds but

also the time periods when it is applied.

KEYWORDS

mutual fund, skew, volatility-timing

JEL CLASSIFICATION

G11, G23

1|INTRODUCTION

Recent studies on stock and currency market find that scaling the investment positions by past volatility from time

to time could help reduce the uncertainty and improve performance. For example, Barroso and Santa-Clara (2015)

show that by scaling the position of momentum strategy on the stock market, the Sharpe ratios are doubled and

huge losses are avoided during the stock market crashes. Moreira and Muir (2017) further demonstrate that scaling

positions by volatility also increase Sharpe ratios of other equities portfolios and the currency carry trade. Therefore,

it is interesting to know whether the same risk management by volatility timing (VT) could improve the performance

of investments in mutual funds. Wang et al. (2021) apply the VT technique to mutual funds and find that VT also

improve the performance of mutual fund portfolios.

In our study, we extend the analysis of Wang et al. (2021) in two aspects. First, we compare the performance of

VT in mutual fund portfolios with different investment styles. More specifically, we follow Hunter et al. (2014)to

divide the mutual funds into different active peer groups by investment styles, small capitalization (cap) core, small

cap value, small cap growth, mid cap growth, large cap core, large cap value, large cap growth, and compare the

Received: 10 February 2021 Revised: 18 April 2022 Accepted: 17 June 2022

DOI: 10.1111/irfi.12387

© 2022 International Review of Finance Ltd.

International Review of Finance. 2023;23:87–102. wileyonlinelibrary.com/journal/irfi 87

performance of VT for those different groups.

1

Second, we compare the performance of VT in different sample

periods. Figure 1plots the cumulative excess return of equal weighted equity mutual funds from January 1998 to

December 2020. Obviously, the trend of the cumulative mutual fund excess return changes after June 2009. More

specifically, the return grows faster. A formal structural break test (Chow test) confirms that the excess return of the

equal weighted equity mutual funds portfolio experienced a mean shift in June 2009 (the test pvalue is 0.042).

Hence, it is interesting to compare the performance of VT before and after June 2009.

We follow the methodology by Barroso and Santa-Clara (2015) to dynamically scale the position in mutual fund

investments by the return volatility of the past six-month and compare the excess return series of the scaled portfo-

lios of mutual funds with the unscaled portfolios. The results are mixed. In the period from September 1998 to June

2009, the VT technique significantly raises the Sharpe ratios of mutual fund portfolios which invest in growth stocks.

It also raises Fama–French five-factor alphas of those portfolios. Return distributions of volatility-managed portfolios

are also less left-skewed than the original portfolios. However, the performance of VT is mixed when applied to

mutual fund portfolios which is balanced in growth and value stocks or focused on value stocks. Particularly, VT

increases Sharpe ratio and skewness in these fund portfolios and reduces their kurtosis, but VT also reduces the

Fama–French five-factor alphas of mutual fund portfolios which is balanced or focus on value stocks. More impor-

tantly, VT fails in the period from July 2009 to December 2020. It reduces the Sharpe ratios and Fama–French five-

factor alphas of all mutual fund portfolios while raises the crash risk.

Moreira and Muir (2017) suggest that the fact that VT improves portfolio return challenges the risk-based theory

of asset pricing. Economically speaking, taking on more risk when volatility is high reduces investor utility, therefore,

the VT strategy which reduces risk-taking should require lower risk premium. Our tests using the mutual fund portfo-

lio somehow reduces the puzzle posed by Moreira and Muir (2017). In some mutual fund portfolio groups, VT does

decrease the Fama–French five-factor alphas. Moreover, VT fails to improve performance in more recent years.

What drives the change in VT performance is an interesting question for future theoretical research.

To further explore the reasons behind this difference, we investigate the out-of-sample (OOS) exposures to fac-

tor risks of the mutual fund portfolios. The exposures to factor risks significantly differ between the scaled portfolios

and the unscaled ones.

We find that the changes in the exposure to market risk brought about by scaling tend to increase the Sharpe

ratios of the growth stocks-focused mutual fund portfolios before June 2009. After June 2009, this effect weakens.

Actually, the ability of VT to assume market risk in proper time is worse for all fund portfolios after June 2009.

Before June 2009, the changes in the exposure to Fama–French SMB factor and HML factor brought about by

scaling increase the Sharpe ratios of all mutual fund portfolios, except for the SCG group in which scaling improves

its Sharpe ratio through changing loadings to SMB factors, but does not through the HML factor. After June 2009,

FIGURE 1 Cumulative equal-weighted excess returns of equity mutual funds. Source: WRDS

88 YIN ET AL.