- Email: [email protected]
Linkages between crude oil and emerging Asian stock markets: New evidence from the Chinese stock market crash
Journal Pre-proof
Linkages between Crude Oil and Emerging Asian Stock Markets: New Evidence from the Chinese Stock Market Crash Imran Yousaf PhD Finance Scholar , Arshad Hasan Professor, Dean of Management and Social Sciences department PII: DOI: Reference:
S1544-6123(19)30683-X https://doi.org/10.1016/j.frl.2019.08.023 FRL 1275
To appear in:
Finance Research Letters
Received date: Revised date: Accepted date:
5 July 2019 19 August 2019 27 August 2019
Please cite this article as: Imran Yousaf PhD Finance Scholar , Arshad Hasan Professor, Dean of Management and Linkages between Crude Oil and Emerging Asian Stock Markets: New Evidence from the Chinese Stock Market Crash, Finance Research Letters (2019), doi: https://doi.org/10.1016/j.frl.2019.08.023
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
Highlights
Return spillover from oil to majority emerging stock markets in the Chinese crash.
Volatility transmitted from oil to Indian and Korean stock markets in the Chinese crash.
The weights of oil assets in oil-stock portfolios decrease in the Chinese crash.
Less oil assets are required to minimize portfolio risk in the Chinese crisis.
Linkages between Crude Oil and Emerging Asian Stock Markets: New Evidence from the Chinese Stock Market Crash First Author (Corresponding Author) Name: Imran Yousaf Department: Management and Social Sciences Designation: PhD Finance Scholar University: Capital University of Science and Technology, Islamabad Email: [email protected] Address: Capital University of Science and Technology, Islamabad
Second Author Name: Arshad Hasan Department: Management and Social Sciences Designation: Professor, Dean of Management and Social Sciences department at Capital University of Science and Technology, Islamabad Email: [email protected] Address: Capital University of Science and Technology, Islamabad
Abstract This study examines returns and volatility spillover between crude oil and emerging Asian stock markets during the Chinese stock market crash of 2015. The empirical findings reveal a positive causal effect from crude oil price changes to the majority stock markets. Volatility is transmitted from oil to the Indian and Korean stock markets. The weights of oil assets in oilstock portfolios decrease during the Chinese market crash compared to the full sample and the US financial crisis. Lastly, less oil assets are required to minimize portfolio risk in the Chinese crisis than in the full sample or US crisis. Keywords: Return spillover, volatility spillover, oil markets, stock markets, Chinese stock market crash, oil-stock portfolio JEL Classification: F3, G1, Q43
1. Introduction
Along with return links between oil and stock markets (Dagher and Hariri, 2013), the area of volatility transmission or link has recently attracted considerable attention from regulators and investors. Previous studies suggest that volatility transmission across asset markets generates a considerable challenge for investors. For example, high oil prices and volatility shocks could cause a downswing in stock markets and the wealth of stock market investors, which could ultimately reduce consumer spending and economic growth, especially in crisis periods, when the return and volatility transmission between markets vary intensely (Jang and Sul, 2002; Yang et al., 2003; Bouri, 2014; Bouri, 2015). These challenges push investors to better risk assessment and management. Investors need to continuously adjust their asset allocation to make their investment portfolio more resistant to crisis or turmoil periods, whereas financial market regulators need to take appropriate action during turmoil periods to maintain the financial stability of markets (Bouri, 2013).
This study aims to explore the return and volatility spillover between oil and emerging Asian stock markets particularly during Chinese stock market crash of 2015. The Chinese stock market crashed in 2015 (Lieo and Ziemba, 2016; Han and Liang, 2017; Ahmed and Huo, 2019). The CSI 300 index reached 5,178 points in mid-June of 2015, then took a rollercoaster ride and declined by 34% in just 20 days, losing 1,000 points within just one week. Almost 50% of stocks lost more than half their pre-crash market value. Some lost more than 77% of their pre-crash market value. This was the most dramatic stock market crisis in history (Han and Liang, 2017). China is undeniably the main trading partner of the emerging Asian stock markets. Therefore, the Chinese stock markets are highly integrated with emerging Asian stock markets. Ahmed and Huo, (2019) find that price and volatility spillover from China to other Asia-Pacific stock
markets become stronger during the Chinese stock market crash, thus strong regional integration may reduce diversification opportunities in both China and the Asia-Pacific neighbours. As an alternate, portfolio managers may also add the oil asset in their Asian stocks portfolios to get the benefit of diversification during the Chinese stock market crash. Diversification benefit is higher in case of insignificant spillover between oil and stock markets and vice versa. However, in the finance literature, the link/spillover between oil and stock markets during the Chinese stock market crash is not yet explored. The majority of previous work focuses only on examining the spillover between crude oil and stock markets during the US global financial crisis and other economic crises (Liu et al., 2017; Aydoğan et al., 2017; Wei et al., 2019; Ferreira et al., 2019; Cai et al., 2019). To address this literature gap, this study examines the returns and volatility spillover between the crude oil market and the emerging Asian stock markets during the Chinese stock market crash. To better understand the impact of the Chinese stock market crash, this study also analyses the return and volatility transmissions between crude oil and stock markets during the full sample period (2000 to 2018) and the US global financial crisis of 2008. Moreover, the study computes the optimal weights and hedge ratios for oil-stock portfolios during the full sample, US crisis, and Chinese crisis. The following section includes a data description and methodology, and Section 3 provides the empirical results. Lastly, section 4 concludes the discussion.
2. Data and Methodology 2.1 Data Description This study uses the daily data of accepted benchmark stock indices of nine Asian countries, China, India, Korea, Indonesia, Pakistan, Malaysia, Philippines, Thailand and Taiwan. The study selects the emerging economies of Asia by taking a list of countries included in the MSCI emerging market index. The index is assumed to be the same on non-trading days
(holidays except weekends) as on the previous trading day, suggested by Malik and Hammoudeh (2007). The data for stock indices are taken from the Data Stream database. The data for daily Brent oil prices are taken from the ―Energy Information Administration (EIA)‖. We use Brent spot prices as an indicator of international oil prices, because Brent serves as a pricing benchmark for around two-thirds of the international oil trade (Maghyereh 2004; Fattouh, 2011). The study uses three time periods, the full sample (January 2000 to June 2018), the US subprime crisis period (August 2007 to July 2010) and the Chinese stock market crash period (June 2015 to May 2018). He (2001) mentions that changes in market correlations take place continuously, not only as a result of crises but also due to the consequences of many financial, economic and political events. Therefore, this study uses the same duration of three years for both crises to compare the coefficients of spillover. Li and Giles (2015) use the same time frame for two crises to examine the spillover effect from the USA and Japan to six Asian markets during the Asian financial crisis and the US subprime financial crisis. The details of stock indices and trading hours are given in Table 1.
Table 1: Trading Hours of Emerging Asian Stock Markets Code
Index
Country
Open
Close
GMT
INDEXCSI
SSE Composite Index
China
09:30
15:00
(UTC +8 Hours)
SENSEX
BSE SENSEX
India
09:15
15:30
(UTC +5:30 Hours)
KOSPI
KOSPI Index
Korea
09:00
15:30
(UTC +9 Hours)
JCI
Jakarta Composite Index
Indonesia
09:00
16:00
(UTC +7 Hours)
FBMKLCI
Kuala Lumpur Composite Index
Malaysia
09:00
17:00
(UTC +8 Hours)
KSE100
KSE-100 index
Pakistan
09:30
15:30
(UTC +5 Hours)
PCOMP
Philip Composite (or PSEi) Index
Phillipine
09:30
15:30
(UTC +8 Hours)
SET Index
SET Index
Thailand
10:00
16:30
(UTC +7 Hours)
TWSE
Taiwan Stock Exchange Weighted Index
Taiwan
09:00
13:30
(UTC +8 Hours)
2.2: Methodology The risks, returns and correlations between portfolio assets (oil and stock) are important fundamentals of empirical finance, particularly for constructing hedging strategies. Ling and
McAleer’s (2003) multivariate ―vector autoregressive
moving average–generalized
autoregressive conditional heteroskedasticity‖ (VAR-GARCH) model for return and volatility spillover is used for analysis. For two series, the conditional mean equation of the VAR-GARCH model is: (1)
,
where
is the vector of returns on the stock market and oil market, at time t,
is a 2 × 2 matrix of parameters giving the impacts of own lagged and cross mean transmissions between two series,
is the vector of error terms of the
conditional mean equations for stock and oil returns, respectively, at time t, and is a sequence of identically and independently distributed random vectors. diag (√
,√
), where,
and
=
are the conditional variances of stock and oil returns,
given as:
(
)
(2)
(
)
(3)
Equations (2) and (3) demonstrate how shock and volatility are transmitted across time and across the relevant return indexes. Furthermore, the conditional covariance
between stock
and oil market returns can be estimated as: √
√
(4)
In the above equation, ρ (rho) indicates the constant conditional correlation. The VARGARCH model assumes that negative and positive shocks have the same impact on the conditional variance. This study estimates spillover by using the VAR-AGARCH model proposed by McAleer et al. (2009). The VAR-AGARCH model incorporates asymmetry.
Specifically, the conditional variance is defined as follows, instead of as in equations (2) and (3): (
)
(
)
[
(
)
]
(
)
(
)
[
(
)
]
(
)
(5)
(6)
Here, A(
) and
[
reveal the association between a
] as well as
(
) and
[
(
)
]
volatility of market and its own lagged positive and
negative returns (Lin et al., 2014). Equations (5) and (6) show the conditional variance of each market and how it depends on its own past shock and past volatility, as well as the past shock and past volatility of other (cross) markets. In equation (5),
and
show
how own past shocks and oil market shocks, respectively, affect the current conditional volatility of stock returns. Here,
and
are measures of how own past volatility and
oil market volatility affect current conditional volatility in the stock market. The estimates of the VAR-AGARCH model are used for calculating the optimal portfolio weights and hedge ratios. This study follows Kroner and Ng (1998) in calculating the optimal portfolio weights of stock and oil:
{
where
is the weight of stock in a $1 oil-stock portfolio at time t,
covariance between the stock and oil markets, stocks and oil, respectively, and 1suggested by Kroner and Sultan (1993):
and
is the conditional
are the conditional variance of
is the weight of oil in a $1 oil-stock portfolio. As
(8) where
represents the hedge ratio. This shows that a short position in the oil market can
hedge a long position in stock. RATS 10.0 software is used to estimate the spillover between markets.
3. Empirical Results and Discussion 3.1: Return and Volatility Spillover Analysis Table 2 shows the spillover during the full sample period. The lagged crude oil returns have a significant and positive impact on the emerging Asian stock markets (except India). These results are consistent with the findings of Bouri (2015), Nath et al. (2014) and Amin and Alamgir (2018). This positive relationship can be explained in different ways for each emerging Asian stock market, such as through the role of oil price subsidies, remittances from Gulf countries, and imports from oil-producing countries to emerging Asian markets. In contrast, returns spillover from emerging Asian stock markets to oil markets is highly insignificant, which implies that return spillover is unidirectional from oil to Asian markets during the full sample period. The results for volatility spillover show that volatility is significantly transmitted from the oil market to the Indian, Indonesian, Korean and Malaysian stock markets. Furthermore, the past volatility of the majority of Asian stock markets (except Philippines) does not transmit to the oil market. This suggests that risk spillover is unidirectional from oil to the Indian, Indonesian, Korean and Malaysian stock markets during the full sample period. Table 3 shows the spillover during the US financial crisis. The findings reveal that the return spillover from oil to the majority of Asian markets (except India and Pakistan) is highly significant. In addition, the return transmission effect from the majority of Asian markets (except India) to oil markets is highly insignificant. So there is evidence of unidirectional
return spillover from oil to the majority of emerging Asian stock markets during the US financial crisis. The results reveal that volatility transmission is evident from oil to the Korean, Taiwanese and Thai stock markets. However, the conditional volatility of oil markets is not affected by the past volatility of the majority of Asian stock markets (except Taiwan and Indonesia). This infers that volatility spillover is unidirectional from oil to the few Asian stock markets during the US financial crisis. Thus, investors in a few emerging Asian markets may not get considerable diversification benefit, because of the unidirectional spillover (or integration). It also implies that the investors in the remaining majority emerging Asian stock markets may still attain substantial diversification benefits by holding oil assets within their stock portfolios during the US financial crisis. Table 4 shows the spillover during the Chinese stock market crash. The results show that the return transmissions are significant from oil to the Korean, Malaysian, Philippine and Taiwanese stock markets during the Chinese stock market crash. However, none of the Asian markets transmit return spillover to the oil market, which implies that return spillover is unidirectional from oil to the Korean, Malaysian, Philippine and Taiwanese stock markets during the Chinese stock market crash. The results for volatility spillover reveal that the volatility is only transmitted from oil to the Indian and Korean stock markets during the Chinese stock market crisis. However, there is significant evidence of volatility spillover from the Chinese, Indian, Korean and Thai stock markets to the oil market during the Chinese stock market crash. This infers that volatility spillover is bidirectional between oil-India and oil-Korea during the Chinese stock market crash. It implies that the portfolio investors in Indian and Korean stock markets may not attain diversification benefits by holding oil assets within their stock portfolios during the Chinese stock market crash. 3.2: Optimal Weights and Hedge Ratio Portfolio Implications
Table 5 shows the optimal weights and hedge ratios for an oil-stock portfolio during the full sample, US crisis and Chinese stock market crash. The average optimal weight range is 0.68 for CHN/OIL to 0.90 for MYS/OIL during the full sample period, indicating that for a Chinaoil portfolio of $1, 68 cents should be invested in Chinese stocks and the remaining 32 cents in the oil market. The average hedge ratio range is 0.01 for PAK/OIL to 0.07 for IND/OIL during the full sample period, indicating that a $1 long position in Pakistani stocks can be hedged for 1 cent with a short position in oil assets. During the US financial crisis, the average optimal portfolio weight ranges from 0.62 for CHN/OIL to 0.92 for MYS/OIL. Furthermore, the average hedge ratio range is -0.01 for PHL/OIL to 0.20 for IND/OIL. During the Chinese stock market crash, the average optimal portfolio weights vary from 0.81 for CHN/OIL to 0.93 for MYS/OIL. However, the average hedge ratio range is -0.01 for MYS/OIL to 0.05 for KOR/OIL. Overall, all optimal weights suggest that investors and portfolio managers should allocate at least two-thirds of investment in emerging Asian stocks and remaining in oil assets for optimal oil-stock portfolios. Moreover, the weights of oil assets in oil-stock portfolios decrease during the Chinese stock market crash, compared to the weights in the full sample and US financial crisis. This suggests that portfolio investors should maintain less crude oil during Chinese crash as compare to US financial crisis in their portfolios. The results reveal that hedge ratios are mostly low for each portfolio, meaning few oil assets are required to minimize stockholder risk during normal and crisis periods. Moreover, hedge ratios are lowest during the Chinese stock market crash, compared to the full sample and US subprime crisis. This implies that less oil assets are needed to minimize the risk for stock investors during the Chinese market crash.
4. Conclusion
The main aim of this study is to examine the return and volatility transmission from oil to emerging Asian stock markets during the Chinese stock market crash of 2015. The VARAGARCH model results are used to compute the optimal weights and hedge ratios for an oilstock portfolio. The comprehensive analysis reveals that the association between oil and the stock markets varies across the emerging Asian markets and the various crises. Based on return linkages, the empirical results reveal a positive causal effect from crude oil price changes to the majority of Asian markets during the US subprime crisis. However, the crude oil price changes transmit to the Korean, Malaysian, Philippine and Taiwanese stock markets during the Chinese stock market crash. Based on volatility linkages, the empirical results show that volatility is transmitted from oil to the Korean, Taiwanese and Thai stock markets during the US subprime crisis period. The volatility spillover is insignificant from emerging Asian stock markets to the oil market during the US subprime crisis. Furthermore, volatility is only transmitted from oil to the Indian and Korean stock markets during the Chinese stock market crash. However, volatility is transmitted from a few emerging stock markets to the crude oil market during the Chinese stock market crash. This infers that volatility spillover is bidirectional between oil-India and oil-Korea during the Chinese stock market crash. It implies that the portfolio investors in Indian and Korean stock markets may not attain diversification benefits by holding oil asset within their stock portfolios during Chinese stock market crash. The study reveals that the weights of oil assets in oil-stock portfolios decrease during the Chinese stock market crash, compared to the full sample and the US financial crisis. This study has following policy implications for portfolio managers regarding the decision of asset allocations: portfolio investors should maintain less crude oil during Chinese crash as compare to US financial crisis in their oil-emerging Asian stock portfolios. Moreover,
portfolio managers should invest at least two-thirds of total investment in majority emerging Asian stocks, whereas, remaining proportion on oil assets during all periods. Hedge ratios are lowest during the Chinese crisis, which implies that less oil assets are required to minimize the portfolio risk of a stock investor during the Chinese crisis, compared to the full sample period and the US financial crisis. These findings are useful for international portfolio managers and policymakers in the countries concerned.
References: Ahmed, A. D., & Huo, R. (2019). Impacts of China's crash on Asia-Pacific financial integration: Volatility interdependence, information transmission and market comovement. Economic Modelling, 79, 28-46. Amin, S. B., & Alamgir, F. (2018). The Nexus between the Oil Price and Stock Market: Evidence from South Asia. USAEE Working Paper No. 18. Aydoğan, B., Tunç, G., & Yelkenci, T. (2017). The impact of oil price volatility on net-oil exporter and importer countries’ stock markets. Eurasian Economic Review, 7(2), 231-253. Bouri, E. I. (2013). Correlation and volatility of the MENA equity markets in turbulent periods, and portfolio implications. Economics Bulletin, 33(2), 1575-1593. Bouri, E. I. (2014). Israeli-Hezbollah war and global financial crisis in the Middle East and North African equity markets. Journal of Economic Integration, 1-19. Bouri, E. I. (2015). Return and volatility linkages between oil prices and the Lebanese stock market in crisis periods. Energy, 89, 365-371. Cai, X. J., Tian, S., Yuan, N., & Hamori, S. (2017). Interdependence between oil and East Asian stock markets: Evidence from wavelet coherence analysis. Journal of International Financial Markets, Institutions and Money, 48, 206-223. Dagher, L., & El Hariri, S. (2013). The impact of global oil price shocks on the Lebanese stock market. Energy, 63, 366-374. Ewing, B. T., & Malik, F. (2016). Volatility spillovers between oil prices and the stock market under structural breaks. Global finance journal, 29, 12-23. Fattouh, B. (2011). An anatomy of the crude oil pricing system. Oxford Institute for Energy Studies. Ferreira, P., Pereira, L., da Silva, M. F., & Pereira, H. B. (2019). Detrended correlation coefficients between oil and stock markets: The effect of the 2008 crisis. Physica A: Statistical Mechanics and its Applications, 517, 86-96.
Han, Q., & Liang, J. (2017). Index Futures Trading Restrictions and Spot Market Quality: Evidence from the Recent Chinese Stock Market Crash. Journal of Futures Markets, 37(4), 411-428. He, L. T. (2001). Time variation paths of international transmission of stock volatility—US vs. Hong Kong and South Korea. Global Finance Journal, 12(1), 79-93. Jang, H., & Sul, W. (2002). The Asian financial crisis and the co-movement of Asian stock markets. Journal of Asian Economics, 13(1), 94-104. Kroner, K. F., & Ng, V. K. (1998). Modeling asymmetric comovements of asset returns. The review of financial studies, 11(4), 817-844. Kroner, K. F., & Sultan, J. (1993). Time-varying distributions and dynamic hedging with foreign currency futures. Journal of financial and quantitative analysis, 28(4), 535-551. Li, Y., & Giles, D. E. (2015). Modelling volatility spillover effects between developed stock markets and Asian emerging stock markets. International Journal of Finance & Economics, 20(2), 155-177. Lieo, S., & Ziemba, W. T. (2016). The bond-stock earnings yield differential model: additional applications and other models for stock market crash prediction. Quantitative Finance Letters, 4(1), 26-34. Lin, B., Wesseh, P. K., & Appiah, M. O. (2014). Oil price fluctuation, volatility spillover and the Ghanaian equity market: Implication for portfolio management and hedging effectiveness. Energy Economics, 42, 172-182. Ling, S., & McAleer, M. (2003). Asymptotic theory for a vector ARMA-GARCH model. Econometric theory, 19(2), 280-310. Liu, X., An, H., Huang, S., & Wen, S. (2017). The evolution of spillover effects between oil and stock markets across multi-scales using a wavelet-based GARCH–BEKK model. Physica A: Statistical Mechanics and its Applications, 465, 374-383. Maghyereh, A. (2004). Oil price shocks and emerging stock markets. A generalized VAR approach. International Journal of Applied Econometrics and Quantitative Studies, 1(2), 27−40. Malik, F., & Hammoudeh, S. (2007). Shock and volatility transmission in the oil, US and Gulf equity markets. International Review of Economics & Finance, 16(3), 357-368. McAleer, M., Hoti, S., & Chan, F. (2009). Structure and asymptotic theory for multivariate asymmetric conditional volatility. Econometric Reviews, 28(5), 422-440. Nath Sahu, T., Bandopadhyay, K., & Mondal, D. (2014). An empirical study on the dynamic relationship between oil prices and Indian stock market. Managerial Finance, 40(2), 200215. Ready, R. C. (2017). Oil prices and the stock market. Review of Finance, 22(1), 155-176.
Wei, Y., Qin, S., Li, X., Zhu, S., & Wei, G. (2019). Oil price fluctuation, stock market and macroeconomic fundamentals: Evidence from China before and after the financial crisis. Finance Research Letters. Yang, J., Kolari, J. W., & Min, I. (2003). Stock market integration and financial crises: the case of Asia. Applied Financial Economics, 13(7), 477-486.
Appendix Table 2: Estimates of bivariate VAR-AGARCH for Oil and Asian stock markets during the full sample period CHN
OIL
IND
OIL
INDO
OIL
KOR
OIL
MYS
OIL
PAK
OIL
PHL
OIL
TAIW
OIL
THA
OIL
3.35e-04a (0.039) 0.051a (0.001)
7.20e-05 (0.777) 0.018 (0.361)
4.97e-04a (0.001) 0.137a (0.000)
9.10e-05 (0.715) 0.042b (0.029)
5.37e-04a (0.000) 0.151a (0.000)
8.57e-05 (0.714) 0.026 (0.242)
2.68e-04b (0.047) 0.073a (0.000)
1.57e-04 (0.554) 0.004a (0.042)
1.63e-04b (0.046) 0.188a (0.000)
1.29e-04 (0.662) 0.028 (0.469)
9.26e-04a (0.000) 0.165a (0.000)
1.11e-04 (0.673) -6.16e-03 (0.790)
3.56e-04b (0.031) 0.143a (0.000)
7.45e-05 (0.766) -0.016 (0.516)
2.33e-04 (0.102) 0.096 (0.119)
1.17e-04 (0.646) 0.027 (0.299)
6.09e-04a (0.000) 0.108a (0.000)
1.30e-04 (0.613) -0.014 (0.519)
0.023a (0.002)
0.063a (0.000)
3.4325e-03 (0.630)
0.060a (0.000)
0.016b (0.024)
0.062c (0.000)
0.034a (0.000)
0.060c (0.000)
0.024a (0.000)
0.062a (0.000)
0.016b (0.025)
0.065a (0.000)
0.027a (0.001)
0.061c (0.000)
0.039a (0.000)
0.063a (0.000)
0.028a (0.000)
0.071 (0.000)
Panel A: Mean Equation Constant
Panel B: Variance Equation Constant
Asymmetry
1.09e-06a (0.004) 0.068a (0.000)
6.18e-07 (0.225) -2.27e-04 (0.833)
2.57e-06a (0.000) 0.052a (0.000)
9.53e-07c (0.071) 3.32e-03 (0.120)
5.24e-06a (0.000) 0.070a (0.000)
8.36e-07 (0.198) 0.007a (0.000)
1.21e-06a (0.000) 0.032a (0.000)
1.34e-06a (0.006) 1.66e-03 (0.189)
8.70e-07a (0.000) 0.081a (0.001)
1.24e-06b (0.014) 1.42e-03a (0.003)
5.92e-06a (0.000) 0.096a (0.000)
9.75e-07c (0.079) -9.63e-04 (0.550)
8.82e-06a (0.000) 0.062a (0.000)
-7.77e-09 (0.992) -1.94e-03 (0.330)
1.07e-06a (0.000) 0.032a (0.000)
9.18e-07c (0.061) -9.62e-04 (0.392)
2.46e-06a (0.000) 0.068a (0.000)
3.93e-07 (0.358) -2.56e-03 (0.005)
-3.602e-04 (0.951) 0.920a (0.000) 7.91e-03 (0.226) 0.019b (0.025)
0.018a (0.001) 1.39e-03 (0.282) 0.953a (0.000) 0.051a (0.000)
0.021a (0.008) 0.868a (0.000) -0.014b (0.046) 0.118a (0.000)
0.016a (0.001) 3.72e-04 (0.839) 0.955a (0.000) 0.051a (0.000)
0.005 (0.654) 0.814a (0.000) -0.094b (0.093) 0.129a (0.000)
0.020a (0.000) 2.87e-04 (0.868) 0.954a (0.000) 0.047a (0.000)
0.027a (0.001) 0.922a (0.000) -0.015c (0.061) 0.079a (0.000)
0.019a (0.000) -1.42e-03 (0.304) 0.950a (0.000) 0.051a (0.000)
0.105a (0.000) 0.854a (0.000) -0.077a (0.002) 0.094a (0.000)
0.016a (0.001) -7.2e-05 (0.893) 0.956a (0.000) 0.047a (0.000)
0.019c (0.071) 0.788a (0.000) -0.010 (0.380) 0.148a (0.000)
0.018a (0.002) 4.79e-03 (0.197) 0.954a (0.000) 0.050a (0.000)
-2.44e-03 (0.744) 0.772a (0.000) 0.015 (0.326) 0.115a (0.000)
0.014a (0.003) 0.019a (0.000) 0.958a (0.000) 0.051a (0.000)
0.015 (0.143) 0.919a (0.000) 4.18e-04 (0.170) 0.078 (0.079)
0.018a (0.000) 2.02e-03 (0.141) 0.953c (0.052) 0.046a (0.000)
0.014 (0.217) 0.831a (0.000) -4.98e-03 (0.694) 0.143a (0.000)
0.012 (0.007) 0.010 (0.160) 0.963 (0.000) 0.044 (0.000)
Panel C: Constant Conditional Correlation 0.070a (0.000)
0.110a (0.000)
0.100a (0.000)
0.105a (0.000)
0.083a (0.000)
0.071a (0.000)
0.024 (0.115)
0.078a (0.000)
0.099a (0.000)
Panel D: Diagnostic Tests LogL
25950.9
26536.5
25332.9
26526.1
29316.5
26797.3
AIC
-8.593
-8.558
-8.971
-8.334
-10.126
-9.260
-9.332
-9.177
SIC
-8.199
-8.164
-8.577
-7.939
10.024
-8.866
-8.938
-8.783
JB
1468.694a (0.000)
503.265a (0.000)
723.688a (0.000)
425.896a (0.000)
1346.223a (0.000)
503.078a (0.000)
726.561a (0.000)
326.056a (0.000)
1601.079a (0.000)
301.083a (0.000)
2131.7132a (0.000)
26577.4
405.212a (0.000)
54819.475a (0.000)
26800.2
504.142a (0.000)
482.572a (0.000)
26780.8 -9.184 -8.790 441.107a (0.000)
13921.778a (0.000)
455.573a (0.000)
52.640a 5.379 17.070 5.733 12.535 4.789 9.591 5.246 15.176 5.768 54.140a 5.437 15.554 6.073 25.302b 5.264 35.047a 4.901 (0.000) (0.944) (0.147) (0.929) (0.404) (0.965) (0.652) (0.949) (0.232) (0.927) (0.000) (0.942) (0.213) (0.912) (0.013) (0.949) (0.000) (0.961) c 11.325 15.645 9.937 16.135 11.387 17.188 4.636 9.296 16.677 13.283 6.200 14.658 8.244 19.828 12.703 15.093 6.073 18.630c (12) (0.501) (0.208) (0.622) (0.185) (0.496) (0.143) (0.969) (0.678) (0.162) (0.349) (0.906) (0.261) (0.766) (0.070) (0.391) (0.236) (0.812) (0.098) Notes: The number of lags for VAR is decided using AIC and SIC criteria. JB, Q(12) and Q2(12) refer to the empirical statistics of Jarque -Bera test for normality, Ljung-Box Q statistics of order 12 for autocorrelation applied to the standardized residuals and squared a b c standardized residuals respectively. CHN, China; IND, India; INDO, Indonesia; KOR, Korea; MYS, Malaysia; PAK, Pakistan; PHL, Philippine; TAIW, Taiwan; THA, Thailand. Values in parentheses are the P-values. , , indicate the statistical significance at 1%, 5% and 10% respectively. Q(12)
Table 3: Estimates of bivariate VAR-AGARCH for Oil and Asian stock markets during the US Subprime Crisis period CHN
OIL
IND
OIL
INDO
OIL
KOR
OIL
MYS
OIL
PAK
OIL
PHL
OIL
TAIW
OIL
THA
OIL
7.00e-04 (0.382) 0.051 (0.186) 0.076b (0.027)
9.08e-04 (0.113) 0.143a (0.000) -0.023 (0.353)
9.71e-04 (0.233) 0.083c (0.056) 0.058 (0.113)
4.04e-04 (0.452) 0.133a (0.001) 0.054a (0.007)
9.9e-04 (0.220) 0.037 (0.480) 0.032b (0.042)
-9.4e-05 (0.836) 0.046 (0.241) 0.046b (0.037)
1.01e-03 (0.206) 0.040 (0.485) 0.075b (0.040)
3.85e-04 (0.140) 0.157a (0.000) 0.053a (0.000)
9.68e-04 (0.239) 0.013 (0.874) 0.071c (0.055)
-4.60e-04c (0.056) 0.228a (0.000) 1.7317e-03 (0.379)
1.58e-03b (0.021) -0.055 (0.254) 0.047a (0.000)
4.27e-04 (0.266) 0.131a (0.000) 0.078a (0.000)
1.24e-03c (0.088) 4.13e-03 (0.906) 0.027b (0.044)
1.50e-04 (0.757) 0.078b (0.039) 0.067a (0.001)
9.8011e-04 (0.498) 0.035 (0.522) 0.077b (0.032)
4.96e-04 (0.347) 0.081b (0.035) 0.045b (0.029)
1.19e-03 (0.146) 0.013 (0.805) 0.080b (0.030)
4.58e-06 (0.258) 0.011 (0.183) 1.27e-03 (0.929) -4.64e-03 (0.658) 0.957a (0.000) 0.060a (0.001)
4.90e-06 (0.112) 0.090a (0.000) 0.011 (0.612) 0.867a (0.000) 4.37e-03 (0.851) 0.097a (0.009)
9.01e-06c (0.068) 0.011 (0.197) 0.012 (0.433) -0.013 (0.318) 0.924a (0.000) 0.080b (0.022)
3.46e-05a (0.000) -0.048c (0.059) -1.59e-03 (0.953) 0.581a (0.000) 0.019 (0.721) 0.456a (0.000)
1.05e-05 (0.131) 7.84e-04 (0.872) 0.067b (0.011) 0.041a (0.002) 0.874a (0.000) 0.082b (0.024)
5.31e-06c (0.071) -0.030c (0.052) 0.046 (0.123) 0.862a (0.000) -0.054c (0.074) 0.205a (0.000)
8.93e-06b (0.016) 0.019a (0.007) 8.09e-03 (0.557) -2.96e-03 (0.672) 0.947a (0.000) 0.068a (0.004)
2.14e-06c (0.084) 0.102a (0.000) 0.032 (0.681) 0.777a (0.000) -0.030 (0.694) 0.175a (0.004)
1.07e-05b (0.026) 1.45e-04 (0.930) 0.015 (0.288) 2.89e-03 (0.257) 0.936a (0.000) 0.067a (0.009)
4.64e-05a (0.000) 0.115a (0.007) 0.055b (0.019) 0.550a (0.000) -0.057 (0.221) 0.506a (0.000)
4.35e-05a (0.000) -1.72e-03 (0.229) 0.293a (0.000) -3.76e-03 (0.230) 0.702a (0.000) 0.015a (0.000)
6.02e-05a (0.000) 0.074a (0.000) -9.63e-03b (0.035) 0.611a (0.000) 8.95e-03 (0.352) 0.235a (0.000)
1.26e-05a (0.000) 2.01e-03 (0.278) 0.079a (0.000) -5.24e-03 (0.141) 0.872a (0.000) 0.073a (0.000)
4.33e-06b (0.041) 0.025 (0.141) -0.046c (0.061) 0.911a (0.000) 0.059c (0.098) 0.077a (0.006)
1.21e-05b (0.057) 0.017b (0.012) 0.011 (0.441) -0.013c (0.065) 0.919c (0.000) 0.092c (0.005)
2.79e-05a (0.007) 0.124a (0.007) -0.080 (0.103) 0.538a (0.000) 0.196b (0.045) 0.266a (0.001)
-6.81e-06 (0.339) 6.29e-03 (0.131) 0.018 (0.241) 0.032 (0.102) 0.937a (0.000) 0.017b (0.045)
Panel A: Mean Equation Constant
-5.96e-06 (0.993) 0.041 (0.290) 0.070a (0.007)
Panel B: Variance Equation Constant
Asymmetry
1.58e-05a (0.002) 0.016 (0.407) 0.012 (0.443 0.863a (0.000) -6.53e-03 (0.734) 0.141a (0.000)
Panel C: Constant Conditional Correlation 0.166a (0.000)
0.242a (0.000)
0.184a (0.000)
0.190a (0.000)
0.206a (0.000)
0.055b (0.023)
0.070c (0.060)
0.217a (0.000)
0.256a (0.000)
Panel D: Diagnostic Tests LogL
3729.843
3792.372
3888.261
3951.892
4392.589
3990.862
3889.633
3929.026
3977.646
AIC
-8.904
-9.075
-9.143
-9.411
-10.469
-9.520
-9.361
-9.610
-9.667
SIC
-8.611
-8.782
-8.850
-9.118
-10.175
-9.227
-9.067
-9.317
-9.374
372.813a 683.462a 505.907a 707.356a 641.227a 636.453a 347.741a 695.302a 591.304a 655.213a 631.118a 680.070a 211.200a 665.809a 329.484a 698.102a 347.502a 673.404a (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Q(12) 12.910 5.531 13.998 5.525 12.347 4.892 12.056 5.469 11.343 6.007 14.387 5.199 11.364 5.345 14.639 4.925 14.713 5.483 (0.376) (0.938) (0.301) (0.938) (0.418) (0.961) (0.441) (0.940) (0.500) (0.916) (0.277) (0.951) (0.498) (0.945) (0.262) (0.960) (0.258) (0.940) a b 9.860 13.770 9.287 13.609 6.729 12.770 5.781 14.214 5.987 14.787 6.314 13.2410 25.369 13.438 6.094 14.412 7.670 14.045 (12) (0.628) (0.316) (0.678) (0.326) (0.875) (0.386) (0.927) (0.287) (0.917) (0.253) (0.899) (0.000) (0.013) (0.338) (0.911) (0.275) (0.810) (0.298) Notes: The number of lags for VAR is decided using AIC and SIC criteria. JB, Q(12) and Q2(12) refer to the empirical statistics of Jarque -Bera test for normality, Ljung-Box Q statistics of order 12 for autocorrelation applied to the standardized residuals and squared a b c standardized residuals respectively. CHN, China; IND, India; INDO, Indonesia; KOR, Korea; MYS, Malaysia; PAK, Pakistan; PHL, Philippine; TAIW, Taiwan; THA, Thailand. Values in parentheses are the P-values. , , indicate the statistical significance at 1%, 5% and 10% respectively. JB
Table 4: Estimates of bivariate VAR-AGARCH for Oil and Asian stock markets during the Chinese Stock market crash period CHN
OIL
IND
OIL
INDO
OIL
KOR
OIL
MYS
OIL
3.68e-04 (0.147) 0.153a (0.000) 0.017 (0.163)
1.95e-04 (0.766) 0.110 (0.220) 0.022 (0.537)
-4.38e-05 (0.865) 0.136b (0.001) 0.021 (0.117)
1.91e-05 (0.979) 2.51e-03 (0.976) 0.060c (0.098)
0.000 (0.531) 0.024 (0.629) 0.053a (0.000)
-0.001 (0.232) 0.020 (0.908) 0.087c (0.077)
0.000 (0.488) 0.081c (0.085) 0.035a (0.000)
-0.002 (0.195) -0.064 (0.755) 0.095c (0.070)
3.31e-06a (0.000) -0.082a (0.000) 0.257b (0.022) 0.856a (0.000) -0.233c (0.091) 0.202a (0.000)
2.62e-06 (0.365) 8.88e-04 (0.638) -0.011 (0.257) 6.92e-03b (0.030) 0.960a (0.000) 0.087a (0.000)
2.99e-06 (0.124) 0.053c (0.096) -0.018 (0.911) 0.764a (0.000) 0.143 (0.646) 0.173a (0.002)
9.57e-08 (0.986) -2.23e-03 (0.313) -6.66e-03 (0.772) 0.011 (0.150) 0.948a (0.000) 0.085 (0.129)
0.002 (0.427) -0.028 (0.504) 0.913b (0.020) -0.427 (0.116) 0.289b (0.045) 0.212a (0.001)
0.000 (0.721) 0.004 (0.270) 0.028 (0.294) 0.087a (0.007) 0.551b (0.017) 0.197b (0.022)
0.000 (0.822) 0.100a (0.009) 0.021 (0.977) 0.822a (0.000) 2.220 (0.386) -0.001 (0.977)
0.000 (0.367) -0.002 (0.122) 0.025 (0.553) 0.006 (0.176) 0.675a (0.005) 0.297c (0.056)
PAK
OIL
PHL
OIL
TAIW
OIL
THA
OIL
3.63e-04 (0.177) 0.289a (0.000) 0.016 (0.161)
2.64e-04 (0.706) -0.015 (0.845) 0.077b (0.046)
-4.52e-05 (0.898) 0.044 (0.291) 0.062a (0.000)
-7.25e-05 (0.918) 0.113 (0.125) 0.069b (0.049)
1.91e-05 (0.941) 0.094b (0.015) 0.057a (0.000)
4.58e-04 (0.453) -0.014 (0.880) 0.037b (0.021)
6.93e-05 (0.710) 0.158a (0.000) 0.015 (0.138)
9.58e-05 (0.879) 0.196 (0.524) 0.070c (0.053)
5.33e-06a (0.001) -0.011 (0.581) 0.012 (0.746) 0.835a (0.000) -0.033 (0.299) 0.264a (0.000)
4.99e-06 (0.110) 1.136e-03 (0.415) 2.41e-03 (0.877) -3.23e-03 (0.184) 0.957a (0.000) 0.073a (0.001)
5.33e-06 (0.255) 0.018 (0.687) 6.52e-03 (0.966) 0.845a (0.000) 0.138 (0.708) 0.103a (0.004)
-6.93e-06 (0.459) -1.91e-03 (0.541) -0.017 (0.142) 7.43e-03 (0.577) 0.968a (0.000) 0.078a (0.000)
5.91e-06a (0.004) -0.031 (0.150) 0.185c (0.097) 0.722a (0.000) -0.205 (0.310) 0.264a (0.000)
3.32e-06 (0.534) 9.49e-04 (0.696) -7.00e-03 (0.354) 7.94e-03 (0.164) 0.971a (0.000) 0.064a (0.000)
1.44e-06b (0.042) -0.036c (0.074) 0.086 (0.477) 0.777a (0.000) 0.193 (0.434) 0.266a (0.000)
6.41e-07 (0.758) -1.54e-03 (0.124) 5.65e-03 (0.642) 0.010b (0.015) 0.946a (0.000) 0.042b (0.025)
Panel A: Mean Equation Constant
0.000 (0.478) 0.049 (0.201) 0.032b (0.017)
3.73e-04 (0.579) 0.088 (0.139) 0.070b (0.043)
Panel B: Variance Equation Constant
Asymmetry
1.33e-06a (0.001) 0.014 (0.383) 0.026 (0.134) 0.962a (0.000) -0.017 (0.386) 0.033c (0.091)
3.00e-06 (0.220) 2.744e-03c (0.053) 0.027 (0.141) -5.31e-03b (0.014) 0.949a (0.000) 0.029 (0.166)
Panel C: Constant Conditional Correlation 0.086b (0.016)
0.100a (0.002)
0.097a (0.002)
0.185a (0.000)
0.161a (0.001)
0.086b (0.013)
0.083b (0.019)
0.098a (0.004)
0.107a (0.002)
Panel D: Diagnostic Tests LogL
4383.560
4641.700
4641.700
2502.320
2618.887
4568.196
4449.762
4645.519
4768.204
AIC
-9.939
-11.088
-11.215
-11.190
-11.603
-11.022
-11.059
-11.236
-11.491
SIC
-9.645
-10.795
-10.921
-10.790
-11.203
-10.728
-10.765
-10.942
-11.197
414.288a 466.084a 249.893a 331..790a 663.567a 855.230a 167.276a 198.407a 40.767a 68.018a 38.763a 54.888a 275.086a 122.360a 127.505a 152.307a 71.062a 82.969a (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Q(12) 17.168 8.073 13.966 9.565 15.588 9.431 10.556 10.627 9.491 7.131 12.165 9.165 9.654 8.827 12.048 9.824 5.066 10.443 (0.143) (0.779) (0.303) (0.654) (0.211) (0.666) (0.567) (0.561) (0.661) (0.849) (0.433) (0.689) (0.646) (0.718) (0.442) (0.631) (0.956) (0.577) 20.263c 13.811 4.153 3.920 10.745 16.411 12.726 9.883 6.358 18.003 8.133 14.740 3.033 15.671 8.341 3.014 24.439b 13.386 12) (0.062) (0.313) (0.981) (0.985) (0.551) (0.173) (0.389) (0.626) (0.897) (0.116) (0.775) (0.256) (0.995) (0.207) (0.758) (0.995) (0.018) (0.342) Notes: The number of lags for VAR is decided using AIC and SIC criteria. JB, Q(12) and Q2(12) refer to the empirical statistics of Jarque -Bera test for normality, Ljung-Box Q statistics of order 12 for autocorrelation applied to the standardized residuals and squared a b c standardized residuals respectively. CHN, China; IND, India; INDO, Indonesia; KOR, Korea; MYS, Malaysia; PAK, Pakistan; PHL, Philippine; TAIW, Taiwan; THA, Thailand. Values in parentheses are the P-values. , , indicate the statistical significance at 1%, 5% and 10% respectively. JB
Table 5: Optimal Weights and Hedge Ratios CHN/OIL
IND/OIL
KOR/OIL
INDO/OIL
MYS/OIL
PAK/OIL
PHL/OIL
THA/OIL
TAIW/OIL
0.68
0.74
0.74
0.75
0.90
0.74
0.75
0.76
0.76
0.05
0.07
0.07
0.06
0.03
0.01
0.04
0.06
0.05
0.62
0.66
0.77
0.76
0.92
0.70
0.70
0.74
0.75
0.14
0.20
0.12
0.11
0.08
0.01
-0.01
0.14
0.14
0.81
0.91
0.91
0.89
0.93
0.85
0.85
0.94
0.89
0.04
0.03
0.05
0.04
-0.01
0.04
0.04
0.03
-0.00
Full Sample Period
US Subprime Crisis
Chinese Stock Market Crash
Note:
and
refer to the optimal weights and hedge ratios respectively.
Linkages between Crude Oil and Emerging Asian Stock Markets: New Evidence from the Chinese Stock Market Crash Imran Yousaf PhD Finance Scholar , Arshad Hasan Professor, Dean of Management and Social Sciences department PII: DOI: Reference:
S1544-6123(19)30683-X https://doi.org/10.1016/j.frl.2019.08.023 FRL 1275
To appear in:
Finance Research Letters
Received date: Revised date: Accepted date:
5 July 2019 19 August 2019 27 August 2019
Please cite this article as: Imran Yousaf PhD Finance Scholar , Arshad Hasan Professor, Dean of Management and Linkages between Crude Oil and Emerging Asian Stock Markets: New Evidence from the Chinese Stock Market Crash, Finance Research Letters (2019), doi: https://doi.org/10.1016/j.frl.2019.08.023
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
Highlights
Return spillover from oil to majority emerging stock markets in the Chinese crash.
Volatility transmitted from oil to Indian and Korean stock markets in the Chinese crash.
The weights of oil assets in oil-stock portfolios decrease in the Chinese crash.
Less oil assets are required to minimize portfolio risk in the Chinese crisis.
Linkages between Crude Oil and Emerging Asian Stock Markets: New Evidence from the Chinese Stock Market Crash First Author (Corresponding Author) Name: Imran Yousaf Department: Management and Social Sciences Designation: PhD Finance Scholar University: Capital University of Science and Technology, Islamabad Email: [email protected] Address: Capital University of Science and Technology, Islamabad
Second Author Name: Arshad Hasan Department: Management and Social Sciences Designation: Professor, Dean of Management and Social Sciences department at Capital University of Science and Technology, Islamabad Email: [email protected] Address: Capital University of Science and Technology, Islamabad
Abstract This study examines returns and volatility spillover between crude oil and emerging Asian stock markets during the Chinese stock market crash of 2015. The empirical findings reveal a positive causal effect from crude oil price changes to the majority stock markets. Volatility is transmitted from oil to the Indian and Korean stock markets. The weights of oil assets in oilstock portfolios decrease during the Chinese market crash compared to the full sample and the US financial crisis. Lastly, less oil assets are required to minimize portfolio risk in the Chinese crisis than in the full sample or US crisis. Keywords: Return spillover, volatility spillover, oil markets, stock markets, Chinese stock market crash, oil-stock portfolio JEL Classification: F3, G1, Q43
1. Introduction
Along with return links between oil and stock markets (Dagher and Hariri, 2013), the area of volatility transmission or link has recently attracted considerable attention from regulators and investors. Previous studies suggest that volatility transmission across asset markets generates a considerable challenge for investors. For example, high oil prices and volatility shocks could cause a downswing in stock markets and the wealth of stock market investors, which could ultimately reduce consumer spending and economic growth, especially in crisis periods, when the return and volatility transmission between markets vary intensely (Jang and Sul, 2002; Yang et al., 2003; Bouri, 2014; Bouri, 2015). These challenges push investors to better risk assessment and management. Investors need to continuously adjust their asset allocation to make their investment portfolio more resistant to crisis or turmoil periods, whereas financial market regulators need to take appropriate action during turmoil periods to maintain the financial stability of markets (Bouri, 2013).
This study aims to explore the return and volatility spillover between oil and emerging Asian stock markets particularly during Chinese stock market crash of 2015. The Chinese stock market crashed in 2015 (Lieo and Ziemba, 2016; Han and Liang, 2017; Ahmed and Huo, 2019). The CSI 300 index reached 5,178 points in mid-June of 2015, then took a rollercoaster ride and declined by 34% in just 20 days, losing 1,000 points within just one week. Almost 50% of stocks lost more than half their pre-crash market value. Some lost more than 77% of their pre-crash market value. This was the most dramatic stock market crisis in history (Han and Liang, 2017). China is undeniably the main trading partner of the emerging Asian stock markets. Therefore, the Chinese stock markets are highly integrated with emerging Asian stock markets. Ahmed and Huo, (2019) find that price and volatility spillover from China to other Asia-Pacific stock
markets become stronger during the Chinese stock market crash, thus strong regional integration may reduce diversification opportunities in both China and the Asia-Pacific neighbours. As an alternate, portfolio managers may also add the oil asset in their Asian stocks portfolios to get the benefit of diversification during the Chinese stock market crash. Diversification benefit is higher in case of insignificant spillover between oil and stock markets and vice versa. However, in the finance literature, the link/spillover between oil and stock markets during the Chinese stock market crash is not yet explored. The majority of previous work focuses only on examining the spillover between crude oil and stock markets during the US global financial crisis and other economic crises (Liu et al., 2017; Aydoğan et al., 2017; Wei et al., 2019; Ferreira et al., 2019; Cai et al., 2019). To address this literature gap, this study examines the returns and volatility spillover between the crude oil market and the emerging Asian stock markets during the Chinese stock market crash. To better understand the impact of the Chinese stock market crash, this study also analyses the return and volatility transmissions between crude oil and stock markets during the full sample period (2000 to 2018) and the US global financial crisis of 2008. Moreover, the study computes the optimal weights and hedge ratios for oil-stock portfolios during the full sample, US crisis, and Chinese crisis. The following section includes a data description and methodology, and Section 3 provides the empirical results. Lastly, section 4 concludes the discussion.
2. Data and Methodology 2.1 Data Description This study uses the daily data of accepted benchmark stock indices of nine Asian countries, China, India, Korea, Indonesia, Pakistan, Malaysia, Philippines, Thailand and Taiwan. The study selects the emerging economies of Asia by taking a list of countries included in the MSCI emerging market index. The index is assumed to be the same on non-trading days
(holidays except weekends) as on the previous trading day, suggested by Malik and Hammoudeh (2007). The data for stock indices are taken from the Data Stream database. The data for daily Brent oil prices are taken from the ―Energy Information Administration (EIA)‖. We use Brent spot prices as an indicator of international oil prices, because Brent serves as a pricing benchmark for around two-thirds of the international oil trade (Maghyereh 2004; Fattouh, 2011). The study uses three time periods, the full sample (January 2000 to June 2018), the US subprime crisis period (August 2007 to July 2010) and the Chinese stock market crash period (June 2015 to May 2018). He (2001) mentions that changes in market correlations take place continuously, not only as a result of crises but also due to the consequences of many financial, economic and political events. Therefore, this study uses the same duration of three years for both crises to compare the coefficients of spillover. Li and Giles (2015) use the same time frame for two crises to examine the spillover effect from the USA and Japan to six Asian markets during the Asian financial crisis and the US subprime financial crisis. The details of stock indices and trading hours are given in Table 1.
Table 1: Trading Hours of Emerging Asian Stock Markets Code
Index
Country
Open
Close
GMT
INDEXCSI
SSE Composite Index
China
09:30
15:00
(UTC +8 Hours)
SENSEX
BSE SENSEX
India
09:15
15:30
(UTC +5:30 Hours)
KOSPI
KOSPI Index
Korea
09:00
15:30
(UTC +9 Hours)
JCI
Jakarta Composite Index
Indonesia
09:00
16:00
(UTC +7 Hours)
FBMKLCI
Kuala Lumpur Composite Index
Malaysia
09:00
17:00
(UTC +8 Hours)
KSE100
KSE-100 index
Pakistan
09:30
15:30
(UTC +5 Hours)
PCOMP
Philip Composite (or PSEi) Index
Phillipine
09:30
15:30
(UTC +8 Hours)
SET Index
SET Index
Thailand
10:00
16:30
(UTC +7 Hours)
TWSE
Taiwan Stock Exchange Weighted Index
Taiwan
09:00
13:30
(UTC +8 Hours)
2.2: Methodology The risks, returns and correlations between portfolio assets (oil and stock) are important fundamentals of empirical finance, particularly for constructing hedging strategies. Ling and
McAleer’s (2003) multivariate ―vector autoregressive
moving average–generalized
autoregressive conditional heteroskedasticity‖ (VAR-GARCH) model for return and volatility spillover is used for analysis. For two series, the conditional mean equation of the VAR-GARCH model is: (1)
,
where
is the vector of returns on the stock market and oil market, at time t,
is a 2 × 2 matrix of parameters giving the impacts of own lagged and cross mean transmissions between two series,
is the vector of error terms of the
conditional mean equations for stock and oil returns, respectively, at time t, and is a sequence of identically and independently distributed random vectors. diag (√
,√
), where,
and
=
are the conditional variances of stock and oil returns,
given as:
(
)
(2)
(
)
(3)
Equations (2) and (3) demonstrate how shock and volatility are transmitted across time and across the relevant return indexes. Furthermore, the conditional covariance
between stock
and oil market returns can be estimated as: √
√
(4)
In the above equation, ρ (rho) indicates the constant conditional correlation. The VARGARCH model assumes that negative and positive shocks have the same impact on the conditional variance. This study estimates spillover by using the VAR-AGARCH model proposed by McAleer et al. (2009). The VAR-AGARCH model incorporates asymmetry.
Specifically, the conditional variance is defined as follows, instead of as in equations (2) and (3): (
)
(
)
[
(
)
]
(
)
(
)
[
(
)
]
(
)
(5)
(6)
Here, A(
) and
[
reveal the association between a
] as well as
(
) and
[
(
)
]
volatility of market and its own lagged positive and
negative returns (Lin et al., 2014). Equations (5) and (6) show the conditional variance of each market and how it depends on its own past shock and past volatility, as well as the past shock and past volatility of other (cross) markets. In equation (5),
and
show
how own past shocks and oil market shocks, respectively, affect the current conditional volatility of stock returns. Here,
and
are measures of how own past volatility and
oil market volatility affect current conditional volatility in the stock market. The estimates of the VAR-AGARCH model are used for calculating the optimal portfolio weights and hedge ratios. This study follows Kroner and Ng (1998) in calculating the optimal portfolio weights of stock and oil:
{
where
is the weight of stock in a $1 oil-stock portfolio at time t,
covariance between the stock and oil markets, stocks and oil, respectively, and 1suggested by Kroner and Sultan (1993):
and
is the conditional
are the conditional variance of
is the weight of oil in a $1 oil-stock portfolio. As
(8) where
represents the hedge ratio. This shows that a short position in the oil market can
hedge a long position in stock. RATS 10.0 software is used to estimate the spillover between markets.
3. Empirical Results and Discussion 3.1: Return and Volatility Spillover Analysis Table 2 shows the spillover during the full sample period. The lagged crude oil returns have a significant and positive impact on the emerging Asian stock markets (except India). These results are consistent with the findings of Bouri (2015), Nath et al. (2014) and Amin and Alamgir (2018). This positive relationship can be explained in different ways for each emerging Asian stock market, such as through the role of oil price subsidies, remittances from Gulf countries, and imports from oil-producing countries to emerging Asian markets. In contrast, returns spillover from emerging Asian stock markets to oil markets is highly insignificant, which implies that return spillover is unidirectional from oil to Asian markets during the full sample period. The results for volatility spillover show that volatility is significantly transmitted from the oil market to the Indian, Indonesian, Korean and Malaysian stock markets. Furthermore, the past volatility of the majority of Asian stock markets (except Philippines) does not transmit to the oil market. This suggests that risk spillover is unidirectional from oil to the Indian, Indonesian, Korean and Malaysian stock markets during the full sample period. Table 3 shows the spillover during the US financial crisis. The findings reveal that the return spillover from oil to the majority of Asian markets (except India and Pakistan) is highly significant. In addition, the return transmission effect from the majority of Asian markets (except India) to oil markets is highly insignificant. So there is evidence of unidirectional
return spillover from oil to the majority of emerging Asian stock markets during the US financial crisis. The results reveal that volatility transmission is evident from oil to the Korean, Taiwanese and Thai stock markets. However, the conditional volatility of oil markets is not affected by the past volatility of the majority of Asian stock markets (except Taiwan and Indonesia). This infers that volatility spillover is unidirectional from oil to the few Asian stock markets during the US financial crisis. Thus, investors in a few emerging Asian markets may not get considerable diversification benefit, because of the unidirectional spillover (or integration). It also implies that the investors in the remaining majority emerging Asian stock markets may still attain substantial diversification benefits by holding oil assets within their stock portfolios during the US financial crisis. Table 4 shows the spillover during the Chinese stock market crash. The results show that the return transmissions are significant from oil to the Korean, Malaysian, Philippine and Taiwanese stock markets during the Chinese stock market crash. However, none of the Asian markets transmit return spillover to the oil market, which implies that return spillover is unidirectional from oil to the Korean, Malaysian, Philippine and Taiwanese stock markets during the Chinese stock market crash. The results for volatility spillover reveal that the volatility is only transmitted from oil to the Indian and Korean stock markets during the Chinese stock market crisis. However, there is significant evidence of volatility spillover from the Chinese, Indian, Korean and Thai stock markets to the oil market during the Chinese stock market crash. This infers that volatility spillover is bidirectional between oil-India and oil-Korea during the Chinese stock market crash. It implies that the portfolio investors in Indian and Korean stock markets may not attain diversification benefits by holding oil assets within their stock portfolios during the Chinese stock market crash. 3.2: Optimal Weights and Hedge Ratio Portfolio Implications
Table 5 shows the optimal weights and hedge ratios for an oil-stock portfolio during the full sample, US crisis and Chinese stock market crash. The average optimal weight range is 0.68 for CHN/OIL to 0.90 for MYS/OIL during the full sample period, indicating that for a Chinaoil portfolio of $1, 68 cents should be invested in Chinese stocks and the remaining 32 cents in the oil market. The average hedge ratio range is 0.01 for PAK/OIL to 0.07 for IND/OIL during the full sample period, indicating that a $1 long position in Pakistani stocks can be hedged for 1 cent with a short position in oil assets. During the US financial crisis, the average optimal portfolio weight ranges from 0.62 for CHN/OIL to 0.92 for MYS/OIL. Furthermore, the average hedge ratio range is -0.01 for PHL/OIL to 0.20 for IND/OIL. During the Chinese stock market crash, the average optimal portfolio weights vary from 0.81 for CHN/OIL to 0.93 for MYS/OIL. However, the average hedge ratio range is -0.01 for MYS/OIL to 0.05 for KOR/OIL. Overall, all optimal weights suggest that investors and portfolio managers should allocate at least two-thirds of investment in emerging Asian stocks and remaining in oil assets for optimal oil-stock portfolios. Moreover, the weights of oil assets in oil-stock portfolios decrease during the Chinese stock market crash, compared to the weights in the full sample and US financial crisis. This suggests that portfolio investors should maintain less crude oil during Chinese crash as compare to US financial crisis in their portfolios. The results reveal that hedge ratios are mostly low for each portfolio, meaning few oil assets are required to minimize stockholder risk during normal and crisis periods. Moreover, hedge ratios are lowest during the Chinese stock market crash, compared to the full sample and US subprime crisis. This implies that less oil assets are needed to minimize the risk for stock investors during the Chinese market crash.
4. Conclusion
The main aim of this study is to examine the return and volatility transmission from oil to emerging Asian stock markets during the Chinese stock market crash of 2015. The VARAGARCH model results are used to compute the optimal weights and hedge ratios for an oilstock portfolio. The comprehensive analysis reveals that the association between oil and the stock markets varies across the emerging Asian markets and the various crises. Based on return linkages, the empirical results reveal a positive causal effect from crude oil price changes to the majority of Asian markets during the US subprime crisis. However, the crude oil price changes transmit to the Korean, Malaysian, Philippine and Taiwanese stock markets during the Chinese stock market crash. Based on volatility linkages, the empirical results show that volatility is transmitted from oil to the Korean, Taiwanese and Thai stock markets during the US subprime crisis period. The volatility spillover is insignificant from emerging Asian stock markets to the oil market during the US subprime crisis. Furthermore, volatility is only transmitted from oil to the Indian and Korean stock markets during the Chinese stock market crash. However, volatility is transmitted from a few emerging stock markets to the crude oil market during the Chinese stock market crash. This infers that volatility spillover is bidirectional between oil-India and oil-Korea during the Chinese stock market crash. It implies that the portfolio investors in Indian and Korean stock markets may not attain diversification benefits by holding oil asset within their stock portfolios during Chinese stock market crash. The study reveals that the weights of oil assets in oil-stock portfolios decrease during the Chinese stock market crash, compared to the full sample and the US financial crisis. This study has following policy implications for portfolio managers regarding the decision of asset allocations: portfolio investors should maintain less crude oil during Chinese crash as compare to US financial crisis in their oil-emerging Asian stock portfolios. Moreover,
portfolio managers should invest at least two-thirds of total investment in majority emerging Asian stocks, whereas, remaining proportion on oil assets during all periods. Hedge ratios are lowest during the Chinese crisis, which implies that less oil assets are required to minimize the portfolio risk of a stock investor during the Chinese crisis, compared to the full sample period and the US financial crisis. These findings are useful for international portfolio managers and policymakers in the countries concerned.
References: Ahmed, A. D., & Huo, R. (2019). Impacts of China's crash on Asia-Pacific financial integration: Volatility interdependence, information transmission and market comovement. Economic Modelling, 79, 28-46. Amin, S. B., & Alamgir, F. (2018). The Nexus between the Oil Price and Stock Market: Evidence from South Asia. USAEE Working Paper No. 18. Aydoğan, B., Tunç, G., & Yelkenci, T. (2017). The impact of oil price volatility on net-oil exporter and importer countries’ stock markets. Eurasian Economic Review, 7(2), 231-253. Bouri, E. I. (2013). Correlation and volatility of the MENA equity markets in turbulent periods, and portfolio implications. Economics Bulletin, 33(2), 1575-1593. Bouri, E. I. (2014). Israeli-Hezbollah war and global financial crisis in the Middle East and North African equity markets. Journal of Economic Integration, 1-19. Bouri, E. I. (2015). Return and volatility linkages between oil prices and the Lebanese stock market in crisis periods. Energy, 89, 365-371. Cai, X. J., Tian, S., Yuan, N., & Hamori, S. (2017). Interdependence between oil and East Asian stock markets: Evidence from wavelet coherence analysis. Journal of International Financial Markets, Institutions and Money, 48, 206-223. Dagher, L., & El Hariri, S. (2013). The impact of global oil price shocks on the Lebanese stock market. Energy, 63, 366-374. Ewing, B. T., & Malik, F. (2016). Volatility spillovers between oil prices and the stock market under structural breaks. Global finance journal, 29, 12-23. Fattouh, B. (2011). An anatomy of the crude oil pricing system. Oxford Institute for Energy Studies. Ferreira, P., Pereira, L., da Silva, M. F., & Pereira, H. B. (2019). Detrended correlation coefficients between oil and stock markets: The effect of the 2008 crisis. Physica A: Statistical Mechanics and its Applications, 517, 86-96.
Han, Q., & Liang, J. (2017). Index Futures Trading Restrictions and Spot Market Quality: Evidence from the Recent Chinese Stock Market Crash. Journal of Futures Markets, 37(4), 411-428. He, L. T. (2001). Time variation paths of international transmission of stock volatility—US vs. Hong Kong and South Korea. Global Finance Journal, 12(1), 79-93. Jang, H., & Sul, W. (2002). The Asian financial crisis and the co-movement of Asian stock markets. Journal of Asian Economics, 13(1), 94-104. Kroner, K. F., & Ng, V. K. (1998). Modeling asymmetric comovements of asset returns. The review of financial studies, 11(4), 817-844. Kroner, K. F., & Sultan, J. (1993). Time-varying distributions and dynamic hedging with foreign currency futures. Journal of financial and quantitative analysis, 28(4), 535-551. Li, Y., & Giles, D. E. (2015). Modelling volatility spillover effects between developed stock markets and Asian emerging stock markets. International Journal of Finance & Economics, 20(2), 155-177. Lieo, S., & Ziemba, W. T. (2016). The bond-stock earnings yield differential model: additional applications and other models for stock market crash prediction. Quantitative Finance Letters, 4(1), 26-34. Lin, B., Wesseh, P. K., & Appiah, M. O. (2014). Oil price fluctuation, volatility spillover and the Ghanaian equity market: Implication for portfolio management and hedging effectiveness. Energy Economics, 42, 172-182. Ling, S., & McAleer, M. (2003). Asymptotic theory for a vector ARMA-GARCH model. Econometric theory, 19(2), 280-310. Liu, X., An, H., Huang, S., & Wen, S. (2017). The evolution of spillover effects between oil and stock markets across multi-scales using a wavelet-based GARCH–BEKK model. Physica A: Statistical Mechanics and its Applications, 465, 374-383. Maghyereh, A. (2004). Oil price shocks and emerging stock markets. A generalized VAR approach. International Journal of Applied Econometrics and Quantitative Studies, 1(2), 27−40. Malik, F., & Hammoudeh, S. (2007). Shock and volatility transmission in the oil, US and Gulf equity markets. International Review of Economics & Finance, 16(3), 357-368. McAleer, M., Hoti, S., & Chan, F. (2009). Structure and asymptotic theory for multivariate asymmetric conditional volatility. Econometric Reviews, 28(5), 422-440. Nath Sahu, T., Bandopadhyay, K., & Mondal, D. (2014). An empirical study on the dynamic relationship between oil prices and Indian stock market. Managerial Finance, 40(2), 200215. Ready, R. C. (2017). Oil prices and the stock market. Review of Finance, 22(1), 155-176.
Wei, Y., Qin, S., Li, X., Zhu, S., & Wei, G. (2019). Oil price fluctuation, stock market and macroeconomic fundamentals: Evidence from China before and after the financial crisis. Finance Research Letters. Yang, J., Kolari, J. W., & Min, I. (2003). Stock market integration and financial crises: the case of Asia. Applied Financial Economics, 13(7), 477-486.
Appendix Table 2: Estimates of bivariate VAR-AGARCH for Oil and Asian stock markets during the full sample period CHN
OIL
IND
OIL
INDO
OIL
KOR
OIL
MYS
OIL
PAK
OIL
PHL
OIL
TAIW
OIL
THA
OIL
3.35e-04a (0.039) 0.051a (0.001)
7.20e-05 (0.777) 0.018 (0.361)
4.97e-04a (0.001) 0.137a (0.000)
9.10e-05 (0.715) 0.042b (0.029)
5.37e-04a (0.000) 0.151a (0.000)
8.57e-05 (0.714) 0.026 (0.242)
2.68e-04b (0.047) 0.073a (0.000)
1.57e-04 (0.554) 0.004a (0.042)
1.63e-04b (0.046) 0.188a (0.000)
1.29e-04 (0.662) 0.028 (0.469)
9.26e-04a (0.000) 0.165a (0.000)
1.11e-04 (0.673) -6.16e-03 (0.790)
3.56e-04b (0.031) 0.143a (0.000)
7.45e-05 (0.766) -0.016 (0.516)
2.33e-04 (0.102) 0.096 (0.119)
1.17e-04 (0.646) 0.027 (0.299)
6.09e-04a (0.000) 0.108a (0.000)
1.30e-04 (0.613) -0.014 (0.519)
0.023a (0.002)
0.063a (0.000)
3.4325e-03 (0.630)
0.060a (0.000)
0.016b (0.024)
0.062c (0.000)
0.034a (0.000)
0.060c (0.000)
0.024a (0.000)
0.062a (0.000)
0.016b (0.025)
0.065a (0.000)
0.027a (0.001)
0.061c (0.000)
0.039a (0.000)
0.063a (0.000)
0.028a (0.000)
0.071 (0.000)
Panel A: Mean Equation Constant
Panel B: Variance Equation Constant
Asymmetry
1.09e-06a (0.004) 0.068a (0.000)
6.18e-07 (0.225) -2.27e-04 (0.833)
2.57e-06a (0.000) 0.052a (0.000)
9.53e-07c (0.071) 3.32e-03 (0.120)
5.24e-06a (0.000) 0.070a (0.000)
8.36e-07 (0.198) 0.007a (0.000)
1.21e-06a (0.000) 0.032a (0.000)
1.34e-06a (0.006) 1.66e-03 (0.189)
8.70e-07a (0.000) 0.081a (0.001)
1.24e-06b (0.014) 1.42e-03a (0.003)
5.92e-06a (0.000) 0.096a (0.000)
9.75e-07c (0.079) -9.63e-04 (0.550)
8.82e-06a (0.000) 0.062a (0.000)
-7.77e-09 (0.992) -1.94e-03 (0.330)
1.07e-06a (0.000) 0.032a (0.000)
9.18e-07c (0.061) -9.62e-04 (0.392)
2.46e-06a (0.000) 0.068a (0.000)
3.93e-07 (0.358) -2.56e-03 (0.005)
-3.602e-04 (0.951) 0.920a (0.000) 7.91e-03 (0.226) 0.019b (0.025)
0.018a (0.001) 1.39e-03 (0.282) 0.953a (0.000) 0.051a (0.000)
0.021a (0.008) 0.868a (0.000) -0.014b (0.046) 0.118a (0.000)
0.016a (0.001) 3.72e-04 (0.839) 0.955a (0.000) 0.051a (0.000)
0.005 (0.654) 0.814a (0.000) -0.094b (0.093) 0.129a (0.000)
0.020a (0.000) 2.87e-04 (0.868) 0.954a (0.000) 0.047a (0.000)
0.027a (0.001) 0.922a (0.000) -0.015c (0.061) 0.079a (0.000)
0.019a (0.000) -1.42e-03 (0.304) 0.950a (0.000) 0.051a (0.000)
0.105a (0.000) 0.854a (0.000) -0.077a (0.002) 0.094a (0.000)
0.016a (0.001) -7.2e-05 (0.893) 0.956a (0.000) 0.047a (0.000)
0.019c (0.071) 0.788a (0.000) -0.010 (0.380) 0.148a (0.000)
0.018a (0.002) 4.79e-03 (0.197) 0.954a (0.000) 0.050a (0.000)
-2.44e-03 (0.744) 0.772a (0.000) 0.015 (0.326) 0.115a (0.000)
0.014a (0.003) 0.019a (0.000) 0.958a (0.000) 0.051a (0.000)
0.015 (0.143) 0.919a (0.000) 4.18e-04 (0.170) 0.078 (0.079)
0.018a (0.000) 2.02e-03 (0.141) 0.953c (0.052) 0.046a (0.000)
0.014 (0.217) 0.831a (0.000) -4.98e-03 (0.694) 0.143a (0.000)
0.012 (0.007) 0.010 (0.160) 0.963 (0.000) 0.044 (0.000)
Panel C: Constant Conditional Correlation 0.070a (0.000)
0.110a (0.000)
0.100a (0.000)
0.105a (0.000)
0.083a (0.000)
0.071a (0.000)
0.024 (0.115)
0.078a (0.000)
0.099a (0.000)
Panel D: Diagnostic Tests LogL
25950.9
26536.5
25332.9
26526.1
29316.5
26797.3
AIC
-8.593
-8.558
-8.971
-8.334
-10.126
-9.260
-9.332
-9.177
SIC
-8.199
-8.164
-8.577
-7.939
10.024
-8.866
-8.938
-8.783
JB
1468.694a (0.000)
503.265a (0.000)
723.688a (0.000)
425.896a (0.000)
1346.223a (0.000)
503.078a (0.000)
726.561a (0.000)
326.056a (0.000)
1601.079a (0.000)
301.083a (0.000)
2131.7132a (0.000)
26577.4
405.212a (0.000)
54819.475a (0.000)
26800.2
504.142a (0.000)
482.572a (0.000)
26780.8 -9.184 -8.790 441.107a (0.000)
13921.778a (0.000)
455.573a (0.000)
52.640a 5.379 17.070 5.733 12.535 4.789 9.591 5.246 15.176 5.768 54.140a 5.437 15.554 6.073 25.302b 5.264 35.047a 4.901 (0.000) (0.944) (0.147) (0.929) (0.404) (0.965) (0.652) (0.949) (0.232) (0.927) (0.000) (0.942) (0.213) (0.912) (0.013) (0.949) (0.000) (0.961) c 11.325 15.645 9.937 16.135 11.387 17.188 4.636 9.296 16.677 13.283 6.200 14.658 8.244 19.828 12.703 15.093 6.073 18.630c (12) (0.501) (0.208) (0.622) (0.185) (0.496) (0.143) (0.969) (0.678) (0.162) (0.349) (0.906) (0.261) (0.766) (0.070) (0.391) (0.236) (0.812) (0.098) Notes: The number of lags for VAR is decided using AIC and SIC criteria. JB, Q(12) and Q2(12) refer to the empirical statistics of Jarque -Bera test for normality, Ljung-Box Q statistics of order 12 for autocorrelation applied to the standardized residuals and squared a b c standardized residuals respectively. CHN, China; IND, India; INDO, Indonesia; KOR, Korea; MYS, Malaysia; PAK, Pakistan; PHL, Philippine; TAIW, Taiwan; THA, Thailand. Values in parentheses are the P-values. , , indicate the statistical significance at 1%, 5% and 10% respectively. Q(12)
Table 3: Estimates of bivariate VAR-AGARCH for Oil and Asian stock markets during the US Subprime Crisis period CHN
OIL
IND
OIL
INDO
OIL
KOR
OIL
MYS
OIL
PAK
OIL
PHL
OIL
TAIW
OIL
THA
OIL
7.00e-04 (0.382) 0.051 (0.186) 0.076b (0.027)
9.08e-04 (0.113) 0.143a (0.000) -0.023 (0.353)
9.71e-04 (0.233) 0.083c (0.056) 0.058 (0.113)
4.04e-04 (0.452) 0.133a (0.001) 0.054a (0.007)
9.9e-04 (0.220) 0.037 (0.480) 0.032b (0.042)
-9.4e-05 (0.836) 0.046 (0.241) 0.046b (0.037)
1.01e-03 (0.206) 0.040 (0.485) 0.075b (0.040)
3.85e-04 (0.140) 0.157a (0.000) 0.053a (0.000)
9.68e-04 (0.239) 0.013 (0.874) 0.071c (0.055)
-4.60e-04c (0.056) 0.228a (0.000) 1.7317e-03 (0.379)
1.58e-03b (0.021) -0.055 (0.254) 0.047a (0.000)
4.27e-04 (0.266) 0.131a (0.000) 0.078a (0.000)
1.24e-03c (0.088) 4.13e-03 (0.906) 0.027b (0.044)
1.50e-04 (0.757) 0.078b (0.039) 0.067a (0.001)
9.8011e-04 (0.498) 0.035 (0.522) 0.077b (0.032)
4.96e-04 (0.347) 0.081b (0.035) 0.045b (0.029)
1.19e-03 (0.146) 0.013 (0.805) 0.080b (0.030)
4.58e-06 (0.258) 0.011 (0.183) 1.27e-03 (0.929) -4.64e-03 (0.658) 0.957a (0.000) 0.060a (0.001)
4.90e-06 (0.112) 0.090a (0.000) 0.011 (0.612) 0.867a (0.000) 4.37e-03 (0.851) 0.097a (0.009)
9.01e-06c (0.068) 0.011 (0.197) 0.012 (0.433) -0.013 (0.318) 0.924a (0.000) 0.080b (0.022)
3.46e-05a (0.000) -0.048c (0.059) -1.59e-03 (0.953) 0.581a (0.000) 0.019 (0.721) 0.456a (0.000)
1.05e-05 (0.131) 7.84e-04 (0.872) 0.067b (0.011) 0.041a (0.002) 0.874a (0.000) 0.082b (0.024)
5.31e-06c (0.071) -0.030c (0.052) 0.046 (0.123) 0.862a (0.000) -0.054c (0.074) 0.205a (0.000)
8.93e-06b (0.016) 0.019a (0.007) 8.09e-03 (0.557) -2.96e-03 (0.672) 0.947a (0.000) 0.068a (0.004)
2.14e-06c (0.084) 0.102a (0.000) 0.032 (0.681) 0.777a (0.000) -0.030 (0.694) 0.175a (0.004)
1.07e-05b (0.026) 1.45e-04 (0.930) 0.015 (0.288) 2.89e-03 (0.257) 0.936a (0.000) 0.067a (0.009)
4.64e-05a (0.000) 0.115a (0.007) 0.055b (0.019) 0.550a (0.000) -0.057 (0.221) 0.506a (0.000)
4.35e-05a (0.000) -1.72e-03 (0.229) 0.293a (0.000) -3.76e-03 (0.230) 0.702a (0.000) 0.015a (0.000)
6.02e-05a (0.000) 0.074a (0.000) -9.63e-03b (0.035) 0.611a (0.000) 8.95e-03 (0.352) 0.235a (0.000)
1.26e-05a (0.000) 2.01e-03 (0.278) 0.079a (0.000) -5.24e-03 (0.141) 0.872a (0.000) 0.073a (0.000)
4.33e-06b (0.041) 0.025 (0.141) -0.046c (0.061) 0.911a (0.000) 0.059c (0.098) 0.077a (0.006)
1.21e-05b (0.057) 0.017b (0.012) 0.011 (0.441) -0.013c (0.065) 0.919c (0.000) 0.092c (0.005)
2.79e-05a (0.007) 0.124a (0.007) -0.080 (0.103) 0.538a (0.000) 0.196b (0.045) 0.266a (0.001)
-6.81e-06 (0.339) 6.29e-03 (0.131) 0.018 (0.241) 0.032 (0.102) 0.937a (0.000) 0.017b (0.045)
Panel A: Mean Equation Constant
-5.96e-06 (0.993) 0.041 (0.290) 0.070a (0.007)
Panel B: Variance Equation Constant
Asymmetry
1.58e-05a (0.002) 0.016 (0.407) 0.012 (0.443 0.863a (0.000) -6.53e-03 (0.734) 0.141a (0.000)
Panel C: Constant Conditional Correlation 0.166a (0.000)
0.242a (0.000)
0.184a (0.000)
0.190a (0.000)
0.206a (0.000)
0.055b (0.023)
0.070c (0.060)
0.217a (0.000)
0.256a (0.000)
Panel D: Diagnostic Tests LogL
3729.843
3792.372
3888.261
3951.892
4392.589
3990.862
3889.633
3929.026
3977.646
AIC
-8.904
-9.075
-9.143
-9.411
-10.469
-9.520
-9.361
-9.610
-9.667
SIC
-8.611
-8.782
-8.850
-9.118
-10.175
-9.227
-9.067
-9.317
-9.374
372.813a 683.462a 505.907a 707.356a 641.227a 636.453a 347.741a 695.302a 591.304a 655.213a 631.118a 680.070a 211.200a 665.809a 329.484a 698.102a 347.502a 673.404a (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Q(12) 12.910 5.531 13.998 5.525 12.347 4.892 12.056 5.469 11.343 6.007 14.387 5.199 11.364 5.345 14.639 4.925 14.713 5.483 (0.376) (0.938) (0.301) (0.938) (0.418) (0.961) (0.441) (0.940) (0.500) (0.916) (0.277) (0.951) (0.498) (0.945) (0.262) (0.960) (0.258) (0.940) a b 9.860 13.770 9.287 13.609 6.729 12.770 5.781 14.214 5.987 14.787 6.314 13.2410 25.369 13.438 6.094 14.412 7.670 14.045 (12) (0.628) (0.316) (0.678) (0.326) (0.875) (0.386) (0.927) (0.287) (0.917) (0.253) (0.899) (0.000) (0.013) (0.338) (0.911) (0.275) (0.810) (0.298) Notes: The number of lags for VAR is decided using AIC and SIC criteria. JB, Q(12) and Q2(12) refer to the empirical statistics of Jarque -Bera test for normality, Ljung-Box Q statistics of order 12 for autocorrelation applied to the standardized residuals and squared a b c standardized residuals respectively. CHN, China; IND, India; INDO, Indonesia; KOR, Korea; MYS, Malaysia; PAK, Pakistan; PHL, Philippine; TAIW, Taiwan; THA, Thailand. Values in parentheses are the P-values. , , indicate the statistical significance at 1%, 5% and 10% respectively. JB
Table 4: Estimates of bivariate VAR-AGARCH for Oil and Asian stock markets during the Chinese Stock market crash period CHN
OIL
IND
OIL
INDO
OIL
KOR
OIL
MYS
OIL
3.68e-04 (0.147) 0.153a (0.000) 0.017 (0.163)
1.95e-04 (0.766) 0.110 (0.220) 0.022 (0.537)
-4.38e-05 (0.865) 0.136b (0.001) 0.021 (0.117)
1.91e-05 (0.979) 2.51e-03 (0.976) 0.060c (0.098)
0.000 (0.531) 0.024 (0.629) 0.053a (0.000)
-0.001 (0.232) 0.020 (0.908) 0.087c (0.077)
0.000 (0.488) 0.081c (0.085) 0.035a (0.000)
-0.002 (0.195) -0.064 (0.755) 0.095c (0.070)
3.31e-06a (0.000) -0.082a (0.000) 0.257b (0.022) 0.856a (0.000) -0.233c (0.091) 0.202a (0.000)
2.62e-06 (0.365) 8.88e-04 (0.638) -0.011 (0.257) 6.92e-03b (0.030) 0.960a (0.000) 0.087a (0.000)
2.99e-06 (0.124) 0.053c (0.096) -0.018 (0.911) 0.764a (0.000) 0.143 (0.646) 0.173a (0.002)
9.57e-08 (0.986) -2.23e-03 (0.313) -6.66e-03 (0.772) 0.011 (0.150) 0.948a (0.000) 0.085 (0.129)
0.002 (0.427) -0.028 (0.504) 0.913b (0.020) -0.427 (0.116) 0.289b (0.045) 0.212a (0.001)
0.000 (0.721) 0.004 (0.270) 0.028 (0.294) 0.087a (0.007) 0.551b (0.017) 0.197b (0.022)
0.000 (0.822) 0.100a (0.009) 0.021 (0.977) 0.822a (0.000) 2.220 (0.386) -0.001 (0.977)
0.000 (0.367) -0.002 (0.122) 0.025 (0.553) 0.006 (0.176) 0.675a (0.005) 0.297c (0.056)
PAK
OIL
PHL
OIL
TAIW
OIL
THA
OIL
3.63e-04 (0.177) 0.289a (0.000) 0.016 (0.161)
2.64e-04 (0.706) -0.015 (0.845) 0.077b (0.046)
-4.52e-05 (0.898) 0.044 (0.291) 0.062a (0.000)
-7.25e-05 (0.918) 0.113 (0.125) 0.069b (0.049)
1.91e-05 (0.941) 0.094b (0.015) 0.057a (0.000)
4.58e-04 (0.453) -0.014 (0.880) 0.037b (0.021)
6.93e-05 (0.710) 0.158a (0.000) 0.015 (0.138)
9.58e-05 (0.879) 0.196 (0.524) 0.070c (0.053)
5.33e-06a (0.001) -0.011 (0.581) 0.012 (0.746) 0.835a (0.000) -0.033 (0.299) 0.264a (0.000)
4.99e-06 (0.110) 1.136e-03 (0.415) 2.41e-03 (0.877) -3.23e-03 (0.184) 0.957a (0.000) 0.073a (0.001)
5.33e-06 (0.255) 0.018 (0.687) 6.52e-03 (0.966) 0.845a (0.000) 0.138 (0.708) 0.103a (0.004)
-6.93e-06 (0.459) -1.91e-03 (0.541) -0.017 (0.142) 7.43e-03 (0.577) 0.968a (0.000) 0.078a (0.000)
5.91e-06a (0.004) -0.031 (0.150) 0.185c (0.097) 0.722a (0.000) -0.205 (0.310) 0.264a (0.000)
3.32e-06 (0.534) 9.49e-04 (0.696) -7.00e-03 (0.354) 7.94e-03 (0.164) 0.971a (0.000) 0.064a (0.000)
1.44e-06b (0.042) -0.036c (0.074) 0.086 (0.477) 0.777a (0.000) 0.193 (0.434) 0.266a (0.000)
6.41e-07 (0.758) -1.54e-03 (0.124) 5.65e-03 (0.642) 0.010b (0.015) 0.946a (0.000) 0.042b (0.025)
Panel A: Mean Equation Constant
0.000 (0.478) 0.049 (0.201) 0.032b (0.017)
3.73e-04 (0.579) 0.088 (0.139) 0.070b (0.043)
Panel B: Variance Equation Constant
Asymmetry
1.33e-06a (0.001) 0.014 (0.383) 0.026 (0.134) 0.962a (0.000) -0.017 (0.386) 0.033c (0.091)
3.00e-06 (0.220) 2.744e-03c (0.053) 0.027 (0.141) -5.31e-03b (0.014) 0.949a (0.000) 0.029 (0.166)
Panel C: Constant Conditional Correlation 0.086b (0.016)
0.100a (0.002)
0.097a (0.002)
0.185a (0.000)
0.161a (0.001)
0.086b (0.013)
0.083b (0.019)
0.098a (0.004)
0.107a (0.002)
Panel D: Diagnostic Tests LogL
4383.560
4641.700
4641.700
2502.320
2618.887
4568.196
4449.762
4645.519
4768.204
AIC
-9.939
-11.088
-11.215
-11.190
-11.603
-11.022
-11.059
-11.236
-11.491
SIC
-9.645
-10.795
-10.921
-10.790
-11.203
-10.728
-10.765
-10.942
-11.197
414.288a 466.084a 249.893a 331..790a 663.567a 855.230a 167.276a 198.407a 40.767a 68.018a 38.763a 54.888a 275.086a 122.360a 127.505a 152.307a 71.062a 82.969a (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Q(12) 17.168 8.073 13.966 9.565 15.588 9.431 10.556 10.627 9.491 7.131 12.165 9.165 9.654 8.827 12.048 9.824 5.066 10.443 (0.143) (0.779) (0.303) (0.654) (0.211) (0.666) (0.567) (0.561) (0.661) (0.849) (0.433) (0.689) (0.646) (0.718) (0.442) (0.631) (0.956) (0.577) 20.263c 13.811 4.153 3.920 10.745 16.411 12.726 9.883 6.358 18.003 8.133 14.740 3.033 15.671 8.341 3.014 24.439b 13.386 12) (0.062) (0.313) (0.981) (0.985) (0.551) (0.173) (0.389) (0.626) (0.897) (0.116) (0.775) (0.256) (0.995) (0.207) (0.758) (0.995) (0.018) (0.342) Notes: The number of lags for VAR is decided using AIC and SIC criteria. JB, Q(12) and Q2(12) refer to the empirical statistics of Jarque -Bera test for normality, Ljung-Box Q statistics of order 12 for autocorrelation applied to the standardized residuals and squared a b c standardized residuals respectively. CHN, China; IND, India; INDO, Indonesia; KOR, Korea; MYS, Malaysia; PAK, Pakistan; PHL, Philippine; TAIW, Taiwan; THA, Thailand. Values in parentheses are the P-values. , , indicate the statistical significance at 1%, 5% and 10% respectively. JB
Table 5: Optimal Weights and Hedge Ratios CHN/OIL
IND/OIL
KOR/OIL
INDO/OIL
MYS/OIL
PAK/OIL
PHL/OIL
THA/OIL
TAIW/OIL
0.68
0.74
0.74
0.75
0.90
0.74
0.75
0.76
0.76
0.05
0.07
0.07
0.06
0.03
0.01
0.04
0.06
0.05
0.62
0.66
0.77
0.76
0.92
0.70
0.70
0.74
0.75
0.14
0.20
0.12
0.11
0.08
0.01
-0.01
0.14
0.14
0.81
0.91
0.91
0.89
0.93
0.85
0.85
0.94
0.89
0.04
0.03
0.05
0.04
-0.01
0.04
0.04
0.03
-0.00
Full Sample Period
US Subprime Crisis
Chinese Stock Market Crash
Note:
and
refer to the optimal weights and hedge ratios respectively.