Professional Documents
Culture Documents
A
JIAN CHEN sset allocation strategies designed can outperform the corresponding passive
is an assistant professor to exploit stock return predict- buy-and-hold benchmark strategies ignoring
of finance at the School
ability are of great interest to return predictability.
of Economics and Fujian
Key Laboratory of Sta- portfolio managers and finan- The portfolio strategies are constructed
tistical Sciences, Xiamen cial economists. Voluminous studies report as follows: First, at the end of each month, the
University, in Xiamen, positive evidence that active portfolio strat- mean-variance portfolio manager recursively
China. egies incorporating return predictability forecasts both the mean and the covariance of
jchenl@xmu.edu.cn can strongly beat the passive buy-and-hold excess Chinese stock returns in the coming
FUWEI JIANG
benchmark and deliver economically and sta- month. Second, the portfolio manager allo-
is an assistant professor of tistically significant gains for portfolio man- cates the weights of his investment in propor-
finance at Central Univer- agers in the U.S. stock market, for example, tion to the forecasted mean to covariance ratio
sity of Finance and Eco- Kandel and Stambaugh [1996]; Campbell and of stock returns, and then he holds the port-
nomics’s School of Finance Thompson [2008]; Cochrane [2008, 2011]; folios for the whole month until he rebalances
in Beijing, China.
and Kong et al. [2011], among many others. the portfolios at the end of next month. We
jfuwei@gmail.com
However, there is much less research ana- assess the performance of the active portfolio
JUN T U lyzing asset allocation choices in the Chi- strategies with four performance criteria: the
is an associate professor nese stock market context, which recently has Sharpe ratio, the certainty equivalent return
of finance at Singapore attracted considerable attention from practi- (CER) gain, the CER gain net of transac-
Management University’s tioners due to its huge size and rapid growth tion costs, and the turnover for each port-
Lee Kong Chian Business
School in Singapore.
in the past two decades. The Chinese stock folio strategy, which are commonly used in
tujun@smu.edu.sg market now ranks largest among all emerging the existing studies, for example, Campbell
stock markets and the second-largest among and Thompson [2008] and DeMiguel et al.
all national stock markets. [2009].
This article aims to design suitable asset To implement the investment strategies,
allocation strategies to incorporate Chinese we employ a host of eight return predictors
stock market return predictability. Specifi- to forecast the next month Chinese stock
cally, we examine whether a hypothetical returns. Specifically, Merton [1980] suggests
mean-variance portfolio manager can take that stock market risk measures are related
advantage of the return predictability of to expected stock returns. Barro [2006] and
the Chinese stock market, and whether the Bali et al. [2009] provide empirical evidence
accordingly constructed portfolio strategies that downside risk capturing the worst loss
such as aggregate market timing strategy of the stock market over a target horizon can
and component portfolio rotation strategies forecast future stock returns. Bali and Peng
available in the Chinese stock market. large certainty equivalent return (CER) gains for the
The Journal of Portfolio Management 2014.41.5:71-83. Downloaded from www.iijournals.com by University of Tennessee on 08/16/15.
We then construct combination forecasts to pool portfolio manager, which range from a low of 253 basis
forecasting information in all eight individual predictors points (mean method) to a high of 597 basis points (PLS
together and to improve upon the conventional univar- method). It indicates that the investor would be willing
iate predictive regression forecasts. Rapach et al. [2010] to pay up to 597 basis points per annum to have access to
and others argue that model uncertainty and parameter the portfolios constructed on combination forecasts rela-
instability surrounding the data-generating process for tive to the passive buy-and-hold benchmark portfolio.
stock returns seriously impair the forecasting ability of The aggregate market timing portfolios generate high
individual predictive regression models. In addition, monthly Sharpe ratios of 0.20 to 0.24, which are more
although some individual predictors may generate good than two times larger than 0.09 of the buy-and-hold
forecasting performance over certain sample periods, strategy, which ignores return predictability.
research aiming to identify the best individual pre- We then explicitly take into account trading costs
dictor may be subject to survival bias in that ex ante the of implementing active portfolio strategies. Assuming
investor cannot know which one of the predictors to use, a proportional transaction cost of 50 basis points per
and the best individual predictor may change over time transaction, because of the relatively low share turnover
due to parameter instability. Rapach et al. [2010] hence of our market timing strategies (about 4.67% per month
recommend combining information in all predictors for PLS combination forecasts), the transaction costs are
together to stabilize individual forecasts and improve economically small. After deducting these trading costs,
forecasting performance. In this article, we consider four the market timing portfolio strategies still deliver supe-
combination forecasting methods, including the mean, rior performance, with large net of transaction costs
trimmed mean, principle component (PC), and par- CER gains ranging from 217 basis points (trimmed
tial least square (PLS) methods, which are commonly mean method) to 582 basis points (PLS method) per
used in the related studies such as Rapach et al. [2010]; annum.
Neely et al. [2013]; and Huang et al. [2014]. Rapach and Industry, size, and value portfolios rotation strate-
Zhou [2013] recently conducted a comprehensive survey gies provide further empirical evidence that the portfolio
on stock return forecasting literature, highlighting the manager can gain significant profits by incorporating
challenges faced by the forecasters, and providing an Chinese stock market predictability. For example,
up-to-date review on strategies to improve forecasting Chinese industry portfolios rotation strategies based on
performance by addressing model uncertainty and combination forecasts deliver Sharpe ratios of 0.15 to
parameter instability. 0.16, which are remarkably larger than that of the equal-
Our empirical analysis shows that the Chinese stock weighted buy-and-hold strategy (0.08) for 13 industries.
market presents significant return predictability. Among Industry rotation portfolios strategies based on combina-
the eight return predictors considered, three of them tion forecasts generate economically large CER gains
generate significant positive out-of-sample ROS 2
statistics ranging from 412 basis points to 654 basis points relative
for the Chinese market portfolio. Most important, when to the passive benchmark. Therefore, the investor would
combining information in all eight individual predictors be willing to pay up to 654 basis points per annum to
together, all of the four combination forecasts generate have access to the industry rotation portfolios formed
significant positive ROS 2
, with statistical significance on combination forecasts relative to the buy-and-hold
at the 5% or better levels. The ROS 2
s of combination benchmark. We detect similarly strong economic value
72 A SSET A LLOCATION IN THE CHINESE STOCK M ARKET: THE ROLE OF R ETURN P REDICTABILITY SPECIAL CHINA ISSUE 2015
of return predictability in size and value rotation strate- CER for a portfolio manager who uses a particular
gies as well. predictive regression-forecasting model of excess stock
returns and that for a portfolio manager who uses the
ASSET ALLOCATION STRATEGY historical average forecasts. We multiply this difference
by 12 so that it can be interpreted as the annual portfolio
We assume that a hypothetical portfolio manager management fee that an investor would be willing to
adopts a mean-variance portfolio investment strategy pay to have access to the predictive regression forecasts
as follows: At the end of each month, the manager first instead of the historical average forecasts.
recursively forecasts both the mean and covariance of The monthly Sharpe ratio is the mean portfolio
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
excess Chinese stock returns in the coming month. He return in excess of the risk-free rate divided by the
The Journal of Portfolio Management 2014.41.5:71-83. Downloaded from www.iijournals.com by University of Tennessee on 08/16/15.
then uses the forecasts to make asset allocation decisions standard deviation of the excess portfolio returns. The
across risky Chinese stocks and risk-free bills. The man- average monthly turnover is the percentage of wealth
ager holds the portfolios until he rebalances his portfo- traded each month. For the CER gain net of transaction
lios in the next month. costs, the costs are calculated using the monthly turnover
Specifically, the mean-variance portfolio manager measures and assuming a proportional transaction cost
chooses the weights of equities in the portfolios at the equal to 50 basis points per transaction.
end of month t as follows:
FORECAST CONSTRUCTION
1 −1 ˆ
wt Σˆ R t +1 (1)
γ t +1 Following Campbell and Thompson [2008] and
Welch and Goyal [2008], we construct the out-of-
where γ is the risk aversion coefficient of five, Rˆ t+1 is sample forecasts for the Chinese excess stock returns
the vector of out-of-sample forecasts of excess Chinese based on recursive predictive regressions, in which the
stock returns, and Σˆ t +1 is the forecast of the variance- predictive regression slopes are estimated recursively by
covariance matrix of excess stock returns. The weights using information available up to the period of forecast
on stocks in the manager’s portfolio are restricted to formation, t, to avoid the use of future data not available
lie between 0 and 1.5 to prevent extreme investments at the time of forecast to investors.
and limit the impact of estimation error. The econo- Specifically, the out-of-sample excess return fore-
metric methodology of out-of-sample return prediction cast at period t + 1 and information available through
is described in the following section. When there are period t is generated by
only two assets involved, the manager then allocates
1−wt of a portfolio to risk-free bills. Rˆ tj+1 αˆ t βˆ t X tk (3)
Following Campbell and Thompson [2008] and
DeMiguel et al. [2009], we consider four criteria to where α̂ t and β̂t are the OLS estimates from regressing
evaluate portfolio performance: (i) the Sharpe ratio, (ii) {R sj 1 }ts−=11 on a constant and {X sk }ts−=11 , in which Rtj+1 rep-
the certainty equivalent return (CER) gain, (iii) the resents the monthly excess stock returns of the aggregate
CER gain net of transaction costs, and (iv) the share Chinese stock market portfolio, 13 industry portfolios,
turnover for each portfolio strategy. 10 size portfolios, and 10 value portfolios, respectively;
The CER of a portfolio is calculated by and X tk denotes the return predictors at period t.
We divide the total sample of length T into n initial
E p = μˆ p − 0.5 γ σˆ 2p
CER (2) estimation sub-sample and q out-of-sample evaluation
2 sub-sample, where T = n + q, and get q out-of-sample
where μˆ p and σˆ p are the sample mean and variance,
{ }
T −1
forecasts: Rˆ tj 1 t n . In this article, we use 2001:12 to
respectively, for the portfolio manager’s portfolio over 2004:12 as the initial estimation period so that the fore-
the forecast evaluation period. The CER can be inter- cast evaluation period spans 2005:01 to 2012:12. The
preted as the risk-free return that the practitioner is length of the initial in-sample estimation period balances
willing to accept instead of adopting the given risky having enough observations for precisely estimating the
portfolio. The CER gain is the difference between the
74 A SSET A LLOCATION IN THE CHINESE STOCK M ARKET: THE ROLE OF R ETURN P REDICTABILITY SPECIAL CHINA ISSUE 2015
and σ in Equation (6) are estimated using the maximum stock exchanges, where dividends and earnings are
likelihood method. With the estimated parameters and measured using a one-year moving sum.
specified loss threshold u, VaR can be calculated at any • Book-to-market ratio (log), B/M: the difference
confidence level p. We employ a rolling window of five- between the logarithm of book value and that of
year daily returns to estimate the 95-percentile VaR. market value for A-share stocks listed in Shanghai
In addition, we calculate the realized volatility using and Shenzhen stock exchanges.
squared daily returns over a five-year rolling window, • Earnings-price ratio (log), E/P: the difference
since Ghysels et al. [2005] suggest that the predictability between the logarithm of earnings and that of
of realized volatility is stronger over longer estimation prices on all A-share stocks listed in Shanghai and
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
intervals due to estimation error. Shenzhen stock exchanges, where earnings are
The Journal of Portfolio Management 2014.41.5:71-83. Downloaded from www.iijournals.com by University of Tennessee on 08/16/15.
The second class is economic variables. Following measured using a one-year moving sum.
Welch and Goyal [2008], we consider a group of six • Inf lation, INF: calculated according to the CPI
Chinese economic variables that are popular predictors from the National Bureau of Statistics. Following
in the return predictability literature and are available Welch and Goyal [2008], since the inf lation rate
for the Chinese market. data are released in the following month, we use
the lagged two-month inf lation in regression.
• Dividend-price ratio (log), D/P: difference
between the logarithm of dividends and that of Exhibit 1 shows that Chinese stock market returns
prices for all A-share stocks listed in Shanghai and f luctuate wildly over time. As shown in Panel A of
Shenzhen stock exchanges, where dividends are Exhibit 2, the average of monthly excess aggregate
measured using a one-year moving sum. Chinese stock market return is 0.86%, and its standard
• Dividend yield (log), D/Y: the difference between deviation is 9.11%. The standard deviation is about two
the logarithm of dividends and that of lagged times larger than that of the U.S. stock market reported
prices, where dividends are measured using a one- in the literature (for example, Bali et al. [2009] and
year moving sum. Huang et al. [2014]). Thus the Chinese stock market
• Dividend-payout ratio (log), D/E: difference on average delivers both high return and high vola-
between the log of dividends and log of earnings tility over our sample periods. Nonetheless, the Chinese
for A-share stocks listed in Shanghai and Shenzhen market has a high Sharpe ratio of 0.09 for a buy-and-
EXHIBIT 1
Aggregate Chinese Stock Market Returns
This figure plots the aggregate Chinese stock market returns over the period from January 2002 through December 2012.
hold investor, which is about 40% higher than that of the OUT-OF-SAMPLE FORECASTING
U.S. stock market reported in the literature. In addition, PERFORMANCE
the aggregate Chinese market has negative skewness and
high kurtosis greater than 3, indicating a left-skewed Panel A of Exhibit 3 presents the out-of-sample
and leptokurtic distribution. Therefore, it is important forecasting performance of eight individual predictors
for the portfolio managers to manage both the market for the monthly excess aggregate Chinese market returns
variance risk and downside risk actively when investing over the 2005:01 to 2012:12 forecast evaluation period.2
in the Chinese stock market. Panel A shows that four out of eight individual predic-
Panel A also presents the descriptive statistics for tors generate negative out-of-sample ROS 2
, indicating
the downside risk (DR), the realized volatility (RV ), higher MSFE than the historical average. The limited
and the six Chinese economic variables. The mean of out-of-sample predictability of individual predictors is
downside risk for the aggregate Chinese stock market is largely consistent with Welch and Goyal [2008]; Rapach
12.04%, which implies a possible extreme loss of 12.04 et al. [2010]; and Rapach and Zhou [2013], which show
RMB in one month at the 95% confidence level when that numerous economic variables with in-sample pre-
investing 100 RMB in the aggregate Chinese stock dictive ability fail to deliver consistent out-of-sample
market. The average realized volatility is 7.81%, which forecasting gains in the U.S. stock market. Our findings
captures the variation risk of the stock market. More- are also consistent with Goh et al. [2013], which detect
over, both the downside risk and realized volatility are weak predictive power of economic variables in the
highly persistent with the first-order autocorrelation Chinese stock market. Economically, Welch and Goyal
coefficients of 0.92. The summary statistics of the six [2008] attribute the poor out-of-sample performance
Chinese economic variables are largely consistent with of individual predictors to structural instability. Rapach
literature: the mean values range from −2.90 for D/P and Zhou [2013] further argue that the data-generating
and D/Y to 0.02 for INF, and all economic predictors process for expected stock return is highly uncertain,
are highly persistent. Panel B shows that both downside complex, and constantly evolving, which is unlikely to
risk and realized volatility have low correlation with the be captured by a single predictor reliably and consistently
economic variables. over time. Moreover, the relatively low data quality for
reported earnings and macroeconomic variables, and the
76 A SSET A LLOCATION IN THE CHINESE STOCK M ARKET: THE ROLE OF R ETURN P REDICTABILITY SPECIAL CHINA ISSUE 2015
EXHIBIT 3 downside risk based on all stocks traded on
Out-of-Sample Prediction NYSE, NASDAQ, and AMEX during the
period from 1973 to 2009, and the long-
This table reports the results of out-of-sample forecasts of the aggregate Chinese stock
market. Panel A reports results for the individual forecasts generated by the recursive short spread portfolio formed on downside
univariate predictive regressions using one of the eight individual predictors given in risk generates sizable monthly alpha of 35
the first column. Panel B reports results for the four combination forecasts based on basis points.
the mean, trimmed mean, principle component (PC), and partial least square (PLS) We then construct four combina-
combining methods, respectively. ROS 2
is the Campbell and Thompson (2008)’s out- tion forecasts to pool information in all
of-sample R 2 statistic, which measures the reduction in mean squared forecast error
(MSFE) for the competing predictive regression forecast relative to the historical
eight individual predictors together and
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
average benchmark forecast. MSFE-adjusted is the Clark and West ( 2007)’s statistic
for testing the null hypothesis that the historical average forecast MSFE is less than regression forecasts. Rapach et al. [2010]
or equal to the competing predictive regression forecast MSFE against the one-sided and Rapach and Zhou [2013] argue that
(upper-tail) alternative hypothesis that the historical average forecast MSFE is greater model uncertainty and parameter insta-
than the competing predictive regression forecast MSFE. All of the predictive regres-
bility surrounding the data-generating
sion slopes in out-of-sample forecasts are estimated recursively using the data available
through period of forecast formation t. The out-of-sample evaluation period extends process for expected stock returns seriously
from January 2005 through December 2012. impair the forecasting ability of individual
predictive regression models. In addition,
while some individual predictors may gen-
erate good forecasting performance over
certain sample periods, research aiming to
identify the best individual predictor may
be subject to survival bias in that ex ante
the investor cannot know which one of the
predictors to use and the best model may
change over time due to parameter insta-
bility. Following Rapach et al. [2010], we
hence combine information in all predic-
low propensity to pay dividends in the Chinese stock tors together to stabilize the individual forecasts and
market may also contribute the weak forecasting per- improve forecasting performance. In this article, we
formance of individual economic variables. consider four combination methods:
Among the four individual predictors with positive
ROS2
, three of them (downside risk, B/M, and E/P) are • Mean combination (Mean): uses the simple “1/N”
statistically significant according to the MSFE-adjusted rule that sets equal weight for each individual pre-
statistics, suggesting that these three individual forecasts dictive regression model forecast, which is used in
produce a significantly smaller MSFE than the historical Rapach et al. [2010].
average benchmark. It is interesting to note that the • Trimmed mean combination (Trimmed mean):
downside risk presents the strongest forecasting power sets weight of zero for the individual forecasts with
in term of MSFE over our sample period, with a high the smallest and largest values and “1/(N-2)” for
ROS2
of 5.99%. The finding is largely consistent with the remaining individual forecasts, which is used
the previous research on downside risk documented in in Rapach et al. [2010].
the U.S. stock market. Bali et al. [2009] detect strong • Principal component combination (PC): extracts
return predictability of downside risk for excess returns the common factor from the eight individual pre-
on a number of U.S. aggregate market proxies including dictors explaining maximally the total variations,
the NYSE/AMEX/Nasdaq, NYSE/AMEX, NYSE, then use the estimated PC factor to make forecasts,
Nasdaq, and S&P 500 portfolios over the 1996 to 2005 which is used in Ludvigson and Ng [2007] and
sample period, with R 2 of about 1.6%. Huang et al. Neely et al. [2013].
[2012] find strong cross-sectional forecasting power of
78 A SSET A LLOCATION IN THE CHINESE STOCK M ARKET: THE ROLE OF R ETURN P REDICTABILITY SPECIAL CHINA ISSUE 2015
generate large positive CER gains, ranging from 242 COMPONENT ROTATION PORTFOLIO
basis points (trimmed mean) to 597 basis points (PLS), PERFORMANCE
consistent with the ROS 2
in Exhibit 3. Hence, the port-
folio manager can exploit the predictive power of com- We then examine the performance of component
bination forecasts to make significant investment profits portfolios rotation strategies, where the portfolio manager
up to 597 basis points. The sophisticated combination rebalances his portfolio between the 13 Chinese industry
strategies like PC and PLS perform particularly well portfolios (10 size portfolios, 10 value portfolios) and
with high CER gains. risk-free bill based on the corresponding forecasts of
Specifically, the CER gain of PC forecasts is 555 Chinese industry (size, value) portfolios returns.
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
basis points larger than seven of the eight individual We start with the industry rotation strategy.
The Journal of Portfolio Management 2014.41.5:71-83. Downloaded from www.iijournals.com by University of Tennessee on 08/16/15.
forecasts, and the CER gain of PLS forecasts is 597 basis Exhibit 5 presents the summary statistics and correla-
points larger than all the eight individual forecasts in tion matrix for excess returns of the 13 Chinese industry
Panel A of Exhibit 4. In addition, seven of the eight portfolios. As shown in Panel A of Exhibit 5, the average
individual forecasts in Panel A and all the four combina- returns of the industry portfolios range from 0.34%
tion forecasts in Panel B produce high monthly Sharpe (CNSTR) to 1.20% (MINES), whereas the standard
ratios, which are larger than that of the passive buy-and- deviations range from 8.82% (TRANS) to 11.49%
hold strategy (0.09). Three individual predictors (down- (MEDIA). The maximum Sharpe ratio is 0.12 for
side risk, B/M, and E/P) and all the four combination MINES, whereas the minimum Sharpe ratio is 0.04 for
methods more than double the Shape ratios compared to CNSTR. Panel B reports the correlation matrix for the
the buy-and-hold strategy, in the range of 0.20 to 0.24, 13 Chinese industry portfolios. The correlation coef-
indicating great economic value of return predictability ficients range from 0.42 to 0.96. We then construct an
in the Chinese stock market. equal-weighted buy-and-hold portfolio based on the 13
We then take into account the trading costs of industry portfolios, which is a stringent asset allocation
implementing the market timing strategies, and calcu- benchmark as argued by DeMiguel et al. (2009). Panel A
late the net-of-transaction-costs CER gains for all the shows that the Sharpe ratio of the equal-weighted port-
individual and combination forecasts. Exhibit 4 shows folio is 0.08, greater than many individual industry port-
that the economic value of both the individual predictors folios due to diversification.
and the combination forecasts is robust to transaction Exhibit 6 reports the results for industry rotation
costs due to the relatively low portfolio turnovers. For strategies. As shown in Panel A of Exhibit 6, three of
example, four of the eight individual predictors deliver eight individual predictors (downside risk, realized
positive net-of-transaction-costs CER gains, and three volatility, and book-to-market ratio) generate positive
individual predictors produce annualized net-of-trans- CER gains for industry rotation strategies, ranging from
action-costs CER gains larger than 100 basis points. 324 basis points to 659 basis points per annum. After
In addition, all the four combination forecasts generate accounting for transaction costs, the CER gains of the
large net-of-transaction-costs CER gains, ranging from three predictors remain economically large. In addition,
217 basis points to 582 basis points, with an average two predictors (downside risk and book-to-market ratio)
monthly turnover about 5%. generate large Sharpe ratios of 0.15 for industry rota-
In summary, the aggregate market timing strate- tion strategies, almost double that of the equal-weighted
gies based on a selected subset of individual predictors industry portfolio strategy (0.08). Again, the downside
and all the four combination forecasts can substantially risk performs best among the eight individual predictors.
outperform the passive buy-and-hold benchmark port- In summary, consistent with the results of market timing
folio strategy in the Chinese stock market and deliver strategies displayed in Exhibit 4, the industry portfolios
substantial economic profits for the portfolio manager. are strongly predictable by some individual predictors
Moreover, the combination forecasting methods, which such as downside risk and book-to-market ratio, which
pool the information in all individual predictors together, can be exploited by the portfolio manager to make siz-
often outperform or perform as well as the best indi- able economic profits.
vidual predictor in market timing. Panel B of Exhibit 6 demonstrates the economic
value of combination forecasts for industry rotation
80
EXHIBIT 5
Summary Statistics for Industry Portfolios
Panel A reports the mean, standard deviation (Std. Dev.), first-order autocorrelation coefficient (AR(1)), and Sharpe ratio (SR) for the monthly excess returns (in per-
centage) of the 13 Chinese industry portfolios and the corresponding equal-weighted buy-and-hold portfolio (EW ). Panel B reports the correlation matrix for the 13
industry portfolios. Our data sample period extends from January 2002 through December 2012.
A SSET A LLOCATION IN THE CHINESE STOCK M ARKET: THE ROLE OF R ETURN P REDICTABILITY
SPECIAL CHINA ISSUE 2015
EXHIBIT 6 buy-and-hold industry strategies. The four combination
Industry Rotation Portfolio Performance forecasts also have large Sharpe ratios of 0.15 to 0.16,
almost double the Sharpe ratio of the equal-weighted
This table reports the asset allocation measures for the mean-vari-
ance portfolio manager with a risk aversion coefficient of five, who industry portfolio strategy.
allocates monthly between the 13 Chinese industry portfolios and In summary, our empirical results for industry
risk-free bill using the out-of-sample forecasts for the excess industry rotation portfolio strategies are in line with those for
portfolio returns. Panel A shows results for individual predictive market timing portfolio strategies. The mean-variance
regression forecasts based on one of the eight individual predictors industry rotation strategies incorporating Chinese stock
given in the first column. Panel B exhibits the results for combi-
nation forecasts according to the mean, trimmed mean, principle
market return predictability often can substantially beat
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
gain for the portfolio manager who uses the predictive regression
forecast instead of the historical average benchmark forecast. The ENDNOTES
monthly Sharpe ratio is the average return of the portfolio formed
on the predictive regression forecast in excess of the risk-free rate We thank Dashan Huang, Qi Luo, Sovan Mitra, Hong-
divided by its standard deviation. Turnover is the average monthly feng Peng, Xiaofen Tan, Heping Xiong, Liqing Zhang, Guofu
turnover for the portfolio based on the predictive regression forecast. Zhou, an anonymous referee, and the editor (Frank J. Fabozzi)
Δ, tc = 50 bps is the CER gain assuming a proportional transactions for their valuable comments. Jian Chen acknowledges finan-
cost of 50 basis points per transaction. The out-of-sample evaluation
cial support from the National Natural Science Foundation
period is over 2005:01–2012:12.
of China (No.71201136). Tu would like to thank Heng-fu
Zou and Institute for Advanced Study of Wuhan University
for their continuous intellectual support for the past 20 years.
Part of the work was done when Tu was visiting the School
of Economics and Management of Wuhan University.
1
The Shanghai Stock Exchange was established in 1990
and the Shenzhen Stock Exchange was established in 1991.
Since December 16, 1996, both exchanges have adopted daily
price change limits of 10%. Therefore, this article focuses
only on the post-1996 sample. We use data from 1997 through
2001 to calculate the initial realized volatility and the down-
side risk, and use the sample from January 2002 to construct
CONCLUSION return forecasts.
2
To save space, we show only the out-of-sample fore-
This article investigates the asset allocation issue casting results for the aggregate market. In an unreported
of the Chinese stock market. We examine whether the table, we find similar results for the in-sample forecasts of
mean-variance portfolio manager can incorporate Chi- the aggregate Chinese market and the in- and out-of-sample
forecasts for the industry, size, and value portfolios.
nese stock market return predictability to improve invest-
ment performance and whether the newly proposed active
strategies can generate significant economic gains relative REFERENCES
to the passive benchmark strategies of buying and holding
Bali, T., O. Demirtas, and H. Levy. “Is There an Intertemporal
the stock market. To implement the active portfolio strat- Relation Between Downside Risk and Expected Returns?”
egies, we consider a host of eight individual return predic- Journal of Financial and Quantitative Analysis, Vol. 44, No. 4
tors in forecasting stock returns. Empirically, we show that (2009), pp. 883-909.
two of them (downside risk and book-to-market ratio)
can deliver large economic gains for the investor. Bali, T., and L. Peng. “Is There a Risk-Return Tradeoff ?
We then combine the forecasting information in Evidence from High-Frequency Data.” Journal of Applied
all predictors together using the mean, trimmed mean, Econometrics, Vol. 21, No. 8 (2006), pp. 1169-1198.
principle component, and the partial least square com-
bination forecasting methods to improve on individual Barro, R. “Rare Disasters and Asset Markets in the Twen-
forecasts. We find that all the market timing strategies tieth Century.” Quarterly Journal of Economics, Vol. 121, No. 3
and component rotation strategies based on combina- (2006), pp. 823-866.
82 A SSET A LLOCATION IN THE CHINESE STOCK M ARKET: THE ROLE OF R ETURN P REDICTABILITY SPECIAL CHINA ISSUE 2015
Campbell, J., and S. Thompson. “Predicting the Equity Pre- Kelly, B., and S. Pruitt. “The Three-Pass Regression Filter:
mium Out of Sample: Can Anything Beat the Historical A New Approach to Forecasting Using Many Predictors.”
Average?” Review of Financial Studies, Vol. 21, No. 4 (2008), Working paper, University of Chicago, 2012.
pp. 1509-1531.
——. “Market Expectations in the Cross-Section of Present
Clark, T., and K. West. “Approximately Normal Tests for Values.” Journal of Finance, Vol. 68, No. 5 (2013), pp. 1721-
Equal Predictive Accuracy in Nested Models.” Journal of 1756.
Econometrics, Vol. 138, No. 1 (2007), pp. 291-311.
Kong, A., D. Rapach, J. Strauss, and G. Zhou. “Predicting
Cochrane, J. “The Dog that Did Not Bark: A Defense of Market Components Out of Sample: Asset Allocation Impli-
It is illegal to make unauthorized copies of this article, forward to an unauthorized user or to post electronically without Publisher permission.
The Journal of Portfolio Management 2014.41.5:71-83. Downloaded from www.iijournals.com by University of Tennessee on 08/16/15.
Return Predictability.” Review of Financial Studies, Vol. 21, cations.” The Journal of Portfolio Management, Vol. 37, No. 4
No. 4 (2008), pp. 1533-1575. (2011), pp. 29-41.
——. “Presidential Address: Discount Rates.” Journal of Ludvigson, S., and S. Ng. “The Empirical Risk-Return
Finance, Vol. 66, No. 4 (2011), pp. 1047-1108. Relation: A Factor Analysis Approach.” Journal of Financial
Economcis, Vol. 83, No. 1 (2007), pp. 171-222.
DeMiguel, V., L. Garlappi, and R. Uppal. “Optimal Versus
Naive Diversification: How Inefficient Is the 1/N Portfolio Merton, R. “On Estimating the Expected Return on the
Strategy?” Review of Financial Studies, Vol. 22, No. 5 (2009), Market.” Journal of Financial Economics, Vol. 8, No. 4 (1980),
pp. 1915-1953. pp. 323-361.
Gençay, R., and F. Selçuk. “Extreme Value Theory and Neely, C., D. Rapach, J. Tu, and G. Zhou. “Forecasting the
Value-at-Risk: Relative Performance in Emerging Markets.” Equity Risk Premium: The Role of Technical Indicators.”
International Journal of Forecasting, Vol. 20, No. 2 (2004), pp. Forthcoming in Management Science, 2013.
287-303.
Rapach, D., J. Strauss, and G. Zhou. “Out-of-Sample Equity
Ghysels, E., P. Santa-Clara, and R. Valkanov. “There Is a Premium Prediction: Combination Forecasts and Links to
Risk Return Trade-Off After All.” Journal of Financial Eco- the Real Economy.” Review of Financial Studies, Vol. 23, No.
nomics, Vol. 76, No. 3 (2005), pp. 509-548. 2 (2010), pp. 821-862.
Goh, J., F. Jiang, J. Tu, and Y. Wang. “Can U.S. Economic Rapach, D., and G. Zhou. “Forecasting Stock Returns.” In
Variables Predict the Chinese Stock Market?” Pacific-Basin Handbook of Economic Forecasting II, edited by G. Elliott and
Finance Journal, 22 (April 2013), pp. 69-87. A. Timmermann, pp. 328-383. Elsevier, 2013.
Gupta, A., and B. Liang. “Do Hedge Funds Have Enough Rossi, A., and A. Timmermann. “What Is the Shape of the
Capital? A Value-at-Risk Approach.” Journal of Financial Risk-Return Relation?” Working paper, University of Cali-
Economics, Vol. 77, No. 1 (2005), pp. 219-253. fornia, San Diego, 2011.
Huang, D., F. Jiang, J. Tu, and G. Zhou. “Investor Sentiment Welch, I., and A. Goyal. “A Comprehensive Look at the Empir-
Aligned: A Powerful Predictor of Stock Returns.” Working ical Performance of Equity Premium Prediction.” Review of
paper, Washington University, 2014. Financial Studies, Vol. 21, No. 4 (2008), pp. 1455-1508.