The trend premium around the world: Evidence from the stock market

• Published date: 01 June 2023
• Author: Hai Lin,Pengfei Liu,Cheng Zhang
• Date: 01 June 2023
• DOI: http://doi.org/10.1111/irfi.12400

ORIGINAL ARTICLE

The trend premium around the world: Evidence

from the stock market

Hai Lin

1

| Pengfei Liu

2

| Cheng Zhang

1

1

School of Economics and Finance, Victoria

University of Wellington, Wellington,

New Zealand

2

College of Economics and Management,

China Jiliang University, Hangzhou, China

Correspondence

Pengfei Liu, College of Economics and

Management, China Jiliang University,

Hangzhou 310018, China.

Email: pengfei.liu@cjlu.edu.cn

Funding information

China Scholarship Council

Abstract

This paper studies the predictive power of the trend strat-

egy in the international stock market. Using data from

49 markets, we find that a trend signal exploiting the short-,

intermediate-, and long-term price information can predict

stock returns cross-sectionally in the international market.

The significance of the trend strategy is associated with

market-level characteristics such as macroeconomic condi-

tions, culture, and the information environment. The trend

premium is more pronounced in markets with a more

advanced macroeconomic status, a higher level of informa-

tion uncertainty and individualism, and better accessibility

to foreign investors. Nevertheless, the trend strategy only

outperforms the momentum strategy in a relatively short

horizon.

KEYWORDS

international, momentum, return predictability, trend premium

JEL CLASSIFICATION

G12, G14

1|INTRODUCTION

A central issue in finance is whether asset returns are predictable. The literature has documented the importance of

historical price information in predicting stock returns. For example, Greenwood and Shleifer (2014) and Hirshleifer

et al. (2015) follow behavior theory to show that historical average returns or past trends contain information for

expected returns when investors extrapolate expectations from the past. Other studies also show the importance of

technical analysis. Treynor and Ferguson (1985), Brown and Jennings (1989), Brock et al. (1992), Lo et al. (2000),

Received: 29 November 2020 Revised: 28 June 2022 Accepted: 30 September 2022

DOI: 10.1111/irfi.12400

© 2022 International Review of Finance Ltd.

International Review of Finance. 2023;23:317–358. wileyonlinelibrary.com/journal/irfi 317

Cespa and Vives (2011), Neely et al. (2014), and Han et al. (2016), among others, demonstrate that past returns have

predictive power on future returns. Besides, Covel (2004) and Lo and Hasanhodzic (2009) document that moving

averages (MAs) of prices are widely used by practitioners. The knowledge on the applicability of technical indicators

is beneficial for us to understand the real-world operation of the financial market in addition to the fundamental

information. Nevertheless, little attention has been paid to examining the importance of historical price signals in an

international setting, which is crucial to expand our understanding of cross-market variations in the international

stock market.

As shown in a recent study by Han et al. (2016), a trend strategy using the MA information in stock prices rang-

ing from 3 to 1000 days can generate statistically significant and economically meaningful returns in the US market.

Inspired by Han et al. (2016), we address three questions in this study. First, we examine whether the trend strategy

proposed by Han et al. (2016) applies to other stock markets as the case for the momentum strategy (Chan

et al., 2000; Chui et al., 2010; Rouwenhorst, 1998). Second, we investigate the information structure of the trend

strategy to analyze the contribution of information across different horizons. We compare our results with the

momentum strategy that merely uses the information of one particular time horizon. Third, we try to identify factors

that can explain the difference in the trend premia across different markets. Specifically, we explore the relationship

between the performance of the trend strategy and market-level characteristics.

We conduct the analysis using a large sample that consists of 49 markets and more than 72,000 stocks. Follow-

ing Han et al. (2016), we forecast the monthly stock returns cross-sectionally and sort stocks into quintile portfolios

by their expected returns. A portfolio that longs the stocks with the highest expected returns (quintile 5) and shorts

the stocks with the lowest expected returns (quintile 1) is constructed. The performance of this long-short (H-L)

trend portfolio measures the profitability of the trend strategy, which is defined as trend premium throughout the

paper. We analyze the trend premia at both the individual market level and the global level and document several

findings.

First, we find convincing evidence that the trend strategy generates significant returns in the international stock

market. At the individual market level, in 39 out of 49 markets, the trend portfolio can generate statistically signifi-

cant returns when employing the equal-weighted (EW) portfolios. At the global level, the trend strategy generates

1.22%, 1.67%, and 0.87% monthly return under the market average method, the market composite method, and the

value-weighted method, respectively. The results suggest the existence of the trend premium in the international

stock market.

Second, the trend premium could not be explained by the traditional asset pricing models. Specifically, we use

the CAPM, Fama and French (1993) three-factor and Carhart (1997) four-factor models to calculate the adjusted

returns of H-L portfolios and find that the adjusted returns are still statistically significant. Besides, the trend strategy

survives transaction costs, indicating that the trend premium in the international stock market is economically

meaningful.

Third, price information across different horizons jointly contributes to the performance of the trend strategy,

while the most significant proportion comes from the short-term information. To examine the contribution of differ-

ent horizon information to the trend premium, we divide the MA signals into the short-, intermediate-, and long-

term, respectively. We find that even if the short-term information contributes the most to the trend premium in

terms of magnitude, adding intermediate- and long-term signals can improve the performance of the trend strategy

further. A global portfolio that employs only the short-term information generates a monthly return of 1.04% under

the market average method. When adding the intermediate- and long-term information, the trend premium increases

by 0.18%, which is statistically significant at the 1% level.

Fourth, the trend strategy outperforms the momentum strategy in a relatively short horizon. To compare the

performance of the trend strategy and that of the momentum strategy, we use the same holding period returns rang-

ing from 1 to 6 months. We find that, at the 1-month horizon, the profits of the trend strategy are significantly

higher than that of the momentum strategy under both the market average method and the market composite

method. On top of that, the trend premium is not as stable as the momentum premium. Specifically, the trend

318 LIN ET AL.

premium in the international stock market decays quickly as holding period increases. Our finding reveals that the

advantage of the trend strategy over the momentum strategy in the global stock markets mainly lies in the short

investment horizon.

Finally, there exist significant cross-market variations in the trend premium in the international stock mar-

ket. The literature has documented several explanations on the cross-sectional difference in the international

stock market. The first explanation is the macroeconomic fundamentals. For example, Chan et al. (2005)show

that economic development could explain the foreign bias(investorsunderweightoroverweightforeignmar-

kets) of mutual funds from 26 developed and developing countries. The second explanation is related to cul-

ture. For example, Chui et al. (2010) find that individualism could explain the cross-market variations in the

momentum of 41 countries. The momentum strategy is more profitable in markets with high individualism. The

third explanation is about the information environment. Hirshleifer et al. (2009) document that limited inv estor

attention causes market under reactions, while Zhang (2006)andHanetal.(2016) show the importance of

information uncertainty on the cross-sectional of stock returns. Following these studies, we consider three pos-

sible explanations for the cross-market variations in the trend premium, including macroeconomic fundamen-

tals, culture, and information environment. Utilizing both a bivariate portfolio analysis and the Fama-MacBeth

regression, we find that the trend premium is more pronounced in markets with a more advanced macroeco-

nomic status, a higher level of information uncertainty and individualism, and better accessibility to foreign

investors.

Our study contributes to the literature in several ways. First, we use the historical price information to forecast

stock returns out-of-sample, which is closely related to the studies that extrapolate return expectations using techni-

cal analysis. Therefore, our paper complements the existing studies in terms of extrapolative return expectation

(Barberis et al., 2018; Greenwood & Shleifer, 2014; Hirshleifer et al., 2015) and technical analysis (Brock et al., 1992;

Brown & Jennings, 1989; Cespa & Vives, 2011; Han et al., 2016; Lo et al., 2000; Neely et al., 2014; Treynor &

Ferguson, 1985). The findings on the predictive power of trend signals in 49 international stock markets expand the

understanding of investor behaviors in an international setting.

Second, we document that the performance of the trend premium is related to market-level characteristics,

which confirms the importance of market-level variables in explaining asset returns. Consistent with the literature,

we document the importance of macroeconomic fundamentals, culture, and information environment on explaining

the cross-market difference in the trend premium. These findings extend our understanding as to what extent

the cross-market variations can explain stock returns (see, for example, Chan et al., 2005; Chui et al., 2010;

Lau et al., 2010; Jacobs, 2016).

Third, our study also relates to the literature on whether and how synthesizing more information improves

return predictability. Many studies find that using a large set of predictors can get better performance. For

example, Kelly and Pruitt (2013,2015)andHuangetal.(2015) show that the partial least squares (PLS) method

provides a powerful procedure for extracting information from a large set of predictors. Lin et al. (2018)extend

the PLS method and document its superior forecasti ng performance on the corporate bond market. In our

study, we show that using more information does not necessarily yield a higher return than using less informa-

tion. Although the trend strategy utilizes more informationthan the momentum strategy, it merely outperforms

the momentum strategy in short horizons. This finding extends the study of Han et al. (2016)abouttherela-

tionship between the trend and the momentum strategy. In this regard, our analysis helps to understand

the conditions under which a better out-of-sample performance can be achieved when employing

more information.

The remainder of the paper is organized as follows. Section 2explains how we implement the trend strategy.

Section 3describes the data. Section 4reports the performance of the trend strategy at the individual market level,

while Section 5shows the results of the trend strategy at the global level. Section 6explores market-level character-

istic variables that explain variations in trend premia across different markets. Section 7reports the results of robust-

ness tests. Section 8concludes the paper.

LIN ET AL.319