Loading [MathJax]/jax/output/SVG/fonts/TeX/fontdata.js
Research article Special Issues

Forecasting turbulence in the Asian and European stock market using regime-switching models

  • An early warning system to timely forecast turbulences in the Asian and European stock market is proposed. To ensure comparability, the model is constructed analogously to the early warning system for the US stock market presented by Hauptmann et al. (2014). Based on the time series of discrete monthly returns of the Nikkei 225 and the EuroStoxx 50, filtered probabilities are estimated by two successive Markov-switching models with two regimes each. The market is thus separated in three states: calm, turbulent positive and turbulent negative. Subsequently, a forecasting model using logistic regression and economic input factors is selected. In an empirical asset management case study it is illustrated that the investment performance is improved when considering the signals of the established warning system. Moreover, the US, Asian and European model are compared and interdependencies are highlighted.

    Citation: Janina Engel, Markus Wahl, Rudi Zagst. Forecasting turbulence in the Asian and European stock market using regime-switching models[J]. Quantitative Finance and Economics, 2018, 2(2): 388-406. doi: 10.3934/QFE.2018.2.388

    Related Papers:

    [1] Albert A. Agyemang-Badu, Fernando Gallardo Olmedo, José María Mella Márquez . Conditional macroeconomic and stock market volatility under regime switching: Empirical evidence from Africa. Quantitative Finance and Economics, 2024, 8(2): 255-285. doi: 10.3934/QFE.2024010
    [2] Md Akther Uddin, Mohammad Enamul Hoque, Md Hakim Ali . International economic policy uncertainty and stock market returns of Bangladesh: evidence from linear and nonlinear model. Quantitative Finance and Economics, 2020, 4(2): 236-251. doi: 10.3934/QFE.2020011
    [3] Zhenghui Li, Hao Dong, Zhehao Huang, Pierre Failler . Asymmetric Effects on Risks of Virtual Financial Assets (VFAs) in different regimes: A Case of Bitcoin. Quantitative Finance and Economics, 2018, 2(4): 860-883. doi: 10.3934/QFE.2018.4.860
    [4] Melike E. Bildirici, Mesut M. Badur . The effects of oil prices on confidence and stock return in China, India and Russia. Quantitative Finance and Economics, 2018, 2(4): 884-903. doi: 10.3934/QFE.2018.4.884
    [5] Cemile Özgür, Vedat Sarıkovanlık . An application of Regular Vine copula in portfolio risk forecasting: evidence from Istanbul stock exchange. Quantitative Finance and Economics, 2021, 5(3): 452-470. doi: 10.3934/QFE.2021020
    [6] Haoyu Wang, Dejun Xie . Optimal profit-making strategies in stock market with algorithmic trading. Quantitative Finance and Economics, 2024, 8(3): 546-572. doi: 10.3934/QFE.2024021
    [7] Ana Sousa, Ana Braga, Jorge Cunha . Impact of macroeconomic indicators on bankruptcy prediction models: Case of the Portuguese construction sector. Quantitative Finance and Economics, 2022, 6(3): 405-432. doi: 10.3934/QFE.2022018
    [8] Anton Velinov . On the importance of testing structural identification schemes and the potential consequences of incorrectly identified models. Quantitative Finance and Economics, 2018, 2(1): 591-611. doi: 10.3934/QFE.2018.1.106
    [9] Lucjan T. Orlowski, Monika Sywak . Wavering interactions between commodity futures prices and us dollar exchange rates. Quantitative Finance and Economics, 2019, 3(2): 221-243. doi: 10.3934/QFE.2019.2.221
    [10] Fangzhou Huang, Jiao Song, Nick J. Taylor . The impact of business conditions and commodity market on US stock returns: An asset pricing modelling experiment. Quantitative Finance and Economics, 2022, 6(3): 433-458. doi: 10.3934/QFE.2022019
  • An early warning system to timely forecast turbulences in the Asian and European stock market is proposed. To ensure comparability, the model is constructed analogously to the early warning system for the US stock market presented by Hauptmann et al. (2014). Based on the time series of discrete monthly returns of the Nikkei 225 and the EuroStoxx 50, filtered probabilities are estimated by two successive Markov-switching models with two regimes each. The market is thus separated in three states: calm, turbulent positive and turbulent negative. Subsequently, a forecasting model using logistic regression and economic input factors is selected. In an empirical asset management case study it is illustrated that the investment performance is improved when considering the signals of the established warning system. Moreover, the US, Asian and European model are compared and interdependencies are highlighted.


    1. Introduction

    Financial crises such as Black Monday (1987), the Gulf War aftermath (1990), the Asian financial crisis (1997), the Russian financial crisis (1998), the bursting of the dotcom bubble (2000), the financial crisis (2007-2009) and the European debt crisis (since 2010) emphasize the need for early warning systems. For this reason, there exists abundant literature regarding the identification and prediction of crises. Various approaches include binary classification tree models (Duttagupta and Cashin, 2008), signal extraction (Demirgüc-Kunt and Detragiache, 2005 and Kaminsky and Reinhart, 1996), logit models (Barrell et al., 2010 and Li et al., 2015), logit and binomial trees (Davis and Karim, 2008b), logit and signal extraction (Davis and Karim, 2008a), probit models (Kamin et al., 2001 and Meichle et al., 2011) and Markov-switching models (Hamilton, 1989; Diebold et al., 1989; , 2002; Abiad, 2003; Maheu and McCurdy, 2000 and Hauptmann and Zagst, 2011). The prediction of turning points of time series was particularly treated in Wecker (1979) and Lahiri and Wang (1994).

    In our paper Hauptmann et al. (2014), we developed an early warning system for the US stock market based on two Markov-switching models with two states each. In a comparison, this approach turned out to be more stable than the use of one Markov-switching model with three states. We now aim at developing early warning systems for the Asian and European stock market using the approach from Hauptmann et al. (2014). Furthermore, a detailed comparison of all three markets and models is drawn, shedding light on the underlying dependence structure. To the best of our knowledge, there are two main aspects that, in combination, separate our work from other approaches. First, in contrast to models using purely binary variables, the amount of incorporated information is increased by considering filtered probabilities. Second, we include macroeconomic variables, which as for example shown in Chen (2009), comprise a better predictive power than stock market movements when forecasting recessions.

    The remainder of this article is structured as follows. Section 2 introduces the underlying Markov-switching models. Section 3 analyses interdependencies between the US, Asian and European market as well as the filtered probabilities. Subsequently in Section 4, logistic regression models are developed to predict the results of the Markov-switching models. The performance of the established warning system is illustrated in Section 5. Finally, Section 6 concludes.


    2. Markov-switching models

    Since the methodology of the construction of the early warning system is basically the same as for the US model proposed by Hauptmann et al. (2014), this section will only give a short summary of the relevant aspects. For a more detailed explanation as well as the results for the US market, the reader is kindly referred to Hauptmann et al. (2014). As e.g. Timmermann (2000) has shown, Markov-switching models are highly suitable for applications to economic time series, as they are able to comprise typical characteristics thereof such as fat tails, asymmetries, autocorrelation and volatility clustering. In the following, let (rt)t=1,T refer to the observed monthly returns (of the Nikkei 225 when considering the Asian setting or the EuroStoxx 50 for the European setting). Based on empirical results, we assume normally distributed returns with regime-dependent mean μStR and regime-dependent volatility σSt>0. I.e.

    rt=μSt+σStϵt,with ϵtN(0,1) iid. t{1,,T}, (1)

    where (St)t=1,,T with St{0,1} for all t{1,,T} denotes the unobservable state process with two regimes. St is supposed to be given by a time-homogeneous Markov chain with fixed transition probabilities p,q[0,1], i.e.

    (p1p1qq), (2)

    where p:=P(St=0St1=0) and q:=P(St=1St1=1) and the initial distribution (δ,1δ) with δ:=P(S1=0)[0,1]. Hence, the model is completely determined by the vector

    θ=(p,q,μ0,μ1,σ0,σ1,δ). (3)

    Since the underlying state process (St)t=1,,T cannot be observed, the parameter vector θ is estimated by an expectation maximization (EM) algorithm for incomplete data as formulated by Dempster et al. (1977). More precisely, we use the Baum-Welch algorithm developed by Baum et al. (1970) for Markov-switching models. An advantage of this algorithm is that, due to the Gaussian setting, the optimal solution of the M-step of the EM algorithm can be analytically determined, which results in reduced calculation effort.

    Let ˆθt be the estimate of θt, derived from the information available up to time t, which is given by rt=(rt,,rt). Then the filtered probabilities for the two states j{0,1} are defined by

    pjt:=P(St=jrt;ˆθt). (4)

    Regarding the monthly returns of the Nikkei 225, the characteristics of the two regimes are summarized in Table 1.*

    *The analysis regarding the Asian market was chosen to focus on the years after the Japanese bubble economy, since it marked a massive change in the Japanese market with very different economic characteristics.

    Table 1. Sample characteristics of the monthly returns of the Nikkei 225 derived from a single two-state Markov-switching model based on the time period from 01/1992 to 10/2015.
    calm turbulent
    Mean (ann.) 0.1854 −0.1136
    Standard deviation (ann.) 0.1291 0.2202
     | Show Table
    DownLoad: CSV

    The Markov-switching model identifies two regimes with different volatilities, leading us to the classification of a calm and turbulent market. In the following, let St=0 denote a calm and St=1 a turbulent market state. Since the turbulent market state is of special interest, we decided to further analyse this regime through a second Markov-switching model. Therefore let T={1,,T} be the complete time period under consideration and TD={tT:P(St=1rt;ˆθt)>0.5} (D indicates Distress) the turbulent market periods. Furthermore, let (SDt)tTD denote the regime process restricted to the turbulent phases. Then the filtered probabilites of the second Markov-switching model are given by j{0,1}

    pD,jt:=P(SDt=jSt=1;rt;ˆθt). (5)

    The second filtered probabilities essentially represent conditional probabilities, since they are conditioned on the characterization of a turbulent market. Therefore we define

    pGt:=p0t,pYt:=p1tpD,0tandpRt:=p1tpD,1t, (6)

    where G, Y and R stand for Green, Yellow and Red, referring to the traffic light system introduced in Hauptmann et al. (2014). Note that pGt+pYt+pRt=1 holds.

    The empirical results of all three states of the Nikkei 225 are given in Table 2. The denomination of a turbulent negative and a turbulent positive market state stems from the difference in the mean.

    Table 2. Sample characteristics of the monthly returns of the Nikkei 225 derived from a step-by-step combination of two two-state Markov-switching models based on the time period from 01/1992 to 10/2015.
    calmturbulent negativeturbulent positive
    Mean (ann.)0.1854-0.46390.5867
    Std. dev. (ann.)0.12910.07840.3501
     | Show Table
    DownLoad: CSV

    The corresponding results for the European market are presented in Table 3 and Table 4. Note that the extreme values in Table 4 are due to the short time period that exhibits only 27 turbulent negative and 21 turbulent positive states.

    Table 3. Sample characteristics of the monthly returns of the EuroStoxx 50 derived from a single two-state Markov-switching model based on the time period from 01/1987 to 10/2014.
    calmturbulent
    Mean (ann.)0.2080-0.2848
    Standard deviation (ann.)0.14850.2119
     | Show Table
    DownLoad: CSV
    Table 4. Sample characteristics of the monthly returns of the EuroStoxx 50 derived from a step-by-step combination of two two-state Markov-switching models based on the time period from 01/1987 to 10/2014.
    calmturbulent negativeturbulent positive
    Mean (ann.)0.2080-0.57580.9475
    Std. dev. (ann.)0.14850.06100.2927
     | Show Table
    DownLoad: CSV

    3. Comparison of market turbulences

    In this section we compare the US, Asian and European market according to the filtered probabilities for market turbulences.

    Note that a comparison of the second filtered probabilities is only possible in months which are marked as turbulent by the first models in all three markets. Moreover, a reliable prediction of the second regression model requires a sufficiently large training data set. For these reasons, reliable and available data points start only in mid-2004, leaving just about 50 turbulent months which is too small for a data set to permit a meaningful comparison. Therefore, we will focus the analysis on the first filtered probabilities for now.

    The scatterplots of the considered indices (see Figure 1) reveal a clear dependency. The 'lost decade' of Japan is the only phase that does not indicate any similarity of the markets, which makes sense since the Japanese economy was internationally less integrated at that time. It is interesting to see whether this dependency can also be found in the filtered probabilities. Pearson's linear correlation coefficient and Kendall's tau are given in Table 5 for different periods in time. Overall the results of both correlation coefficients embody the same findings. During the financial crisis positive dependence is highest. In addition, all three markets move very similar within the US real estate upswing period. The European debt crisis is closely related to the US market, while the Asian market behaves rather different. Furthermore, the fact that the Asian market was very loosely tied to the US and the European market during its 'lost decade' is also reflected in the correlation coefficients.

    Figure 1. Scatterplots of the Nikkei 225, the S & P 500 and the EuroStoxx 50 from 05/1995-09/2014. Japan's lost decade is marked in blue (05/1995-12/1999). The burst of the dotcom bubble is marked in orange (01/2000-10/2002). The US real estate upswing is marked in green (11/2002-07/2007). The financial crisis is marked in red (08/2007-02/2009). The European debt crisis is marked in magenta (02/2009-09/2014).
    Table 5. Pearson's linear correlation coefficient and Kendall's tau for the filtered probabilities of the turbulent state with respect to the US, Asian and European market for different periods in time.
    Pearson's linear correlation coefficientAsia - USAsia - EuropeEurope - US
    Lost Decade in Japan (05/95-12/99)0.28760.30710.6727
    Burst of Dotcom Bubble (01/00-10/02)0.26350.05030.7656
    US Real Estate Upswing (11/02-07/07)0.73020.76950.7496
    Financial Crisis (08/07 - 02/09)0.93540.84130.8517
    European Debt Crisis (03/09-09/14)0.48510.51130.7848
    Kendall's tauAsia - USAsia - EuropeEurope - US
    Lost Decade in Japan (05/95-12/99)0.09220.26100.0506
    Burst of Dotcom Bubble (01/00-10/02)0.0553-0.00890.6364
    US Real Estate Upswing (11/02-07/07)0.44740.54890.6353
    Financial Crisis (08/07-02/09)0.60230.68420.7661
    European Debt Crisis (03/09-09/14)0.38850.37040.6852
     | Show Table
    DownLoad: CSV

    Figure 2 visualizes the dependence structure by exhibiting the corresponding copula plots (generated via the empirical distribution function).

    Figure 2. Scatterplots of the copulas of the filtered probabilities for the turbulent state (derived via the empirical distribution function) with respect to the US, Asian and European market from 05/1995-09/2014. Japan's lost decade is marked in blue (05/1995-12/1999). The burst of the dotcom bubble is marked in orange (01/2000-10/2002). The US real estate upswing is marked in green (11/2002-07/2007). The financial crisis is marked in red (08/2007-02/2009). The European debt crisis is marked in magenta (02/2009-09/2014).

    All three scatter plots reveal that, with the exception of the Japanese lost decade and the burst of the dotcom bubble regarding the Asian market, the filtered probabilities are clearly concordant. Moreover, months of crises seem to gather in the upper right quadrant. In fact, the more severe the crisis the closer are the points to the upper right corner. For example during the burst of the dotcom bubble, the EuroStoxx 50 was hit harder than the S & P 500 and the Nikkei 225 (as one can see by simply comparing the indices). Therefore, the orange dots spread on the upper part of the European axis, while regarding the US and the Asian axis some dots can also be found on the lower part. The financial crisis, as the biggest here considered downturn, is exclusively located in the upper right corner. With respect to Japan's lost decade the Nikkei 225 was highly volatile, while the S & P 500 and the EuroStoxx 50 stayed very calm. Accordingly the blue dots spread along the Asian axis, while gathering on the lower half of the US and European axis. In addition, it was tested whether the US market as a globally leading economy is one month ahead of the Asian and European market. The correlations of the filtered probabilities of the US market at time t1 and the Asian/European market at time t are displayed in Table 6 for different periods in time. Comparing these linear coefficients with the ones regarding all markets at the same points in time (see Table 5), an advance of one month of the US market cannot be confirmed.

    Table 6. Correlation coefficient for the filtered probabilities of the turbulent state of the US market at time t1 and the Asian/European market at time t for different periods in time.
    Pearson's linear correlation coefficientAsiat - USt1Europet - USt1
    Lost Decade in Japan (05/95-12/99)0.28040.4714
    Burst of Dotcom Bubble (01/00-10/02)0.18690.3462
    US Real Estate Upswing (11/02-07/07)0.69170.6988
    Financial Crisis (08/07-02/09)0.74640.6326
    European Debt Crisis (03/09-09/14)0.43820.7147
    Kendall's tauAsiat - USt1Europet - USt1
    Lost Decade in Japan (05/95-12/99)0.11300.0143
    Burst of Dotcom Bubble (01/00-10/02)-0.00890.3262
    US Real Estate Upswing (11/02-07/07)0.46240.5589
    Financial Crisis (08/07-02/09)0.21640.3801
    European Debt Crisis (03/09-09/14)0.31190.5636
     | Show Table
    DownLoad: CSV

    4. Logistic regression

    In order to forecast the filtered probabilities p1t and pD,1t, as derived in Section 2, we develop a linear regression model for each of them. To end up with an ordinary linear regression, the probabilities are first transformed with the logistic function, i.e. for j=1 and tT or respectively for j=(D,1) and tTD we define

    yjt:=ln(pjt1pjt). (7)

    Then, the relation between the transformed response variable yjt and a set of covariates xjt is given by

    yjt+1:=βjxjt+ϵjt. (8)

    Furthermore, the regression model is extended by an ARMA(p,q) process to capture time-dependency, possibly inherent in the error term of Equation (8). This leads to

    yjt+1:=βjxjt+pl=1ϕlϵjtl+qk=1θkδjtk+δjt, (9)

    where ϕl and θk denote the coefficients of the ARMA(p,q) process, ϵjt the error terms of the regression models and δjt the error term of the ARMA(p,q) process.

    For further details, see e.g. Hamilton (1994) or Brockwell and Davis (1991).

    Regarding the Asian model, a set of 37 different economic factors was tested as explanatory variables, as well as various transformations thereof.§ The set covers various economic aspects such as interest rates (e.g. MUTAN, TIBOR, LIBOR, Government and Corporate bonds), stock market data (e.g. historical volatility of the Nikkei 225, 10 days momentum), factors reflecting the Japanese economy (e.g. monetary base, GDP, debt level, unemployment rate, purchasing manager's index, producer price index, consumer price index) plus indicators of the global economic situation (e.g. OECD CLI, oil price, exchange rates, Ifo Weltwirtschaftsklima Index). To account for interdependencies between the variables, bivariate products are also admitted as additional factors. All time series are shifted to their date of publication, such that for the regression of yjt+1 only data available at time period t is used. Model selection is based on the AIC criterion (cf. Akaike, 1974) and the Bonferroni correction (see e.g. Shao, 2003) to the significance level α=0.05. Because of the large number of tested factors and products thereof, the resulting model is heavily overfitted with a high adjusted R2 but also an enormous number of covariates. In order to identify a well performing model with a reasonable number of economically interpretable factors, the data set was reduced to a minimal set of relevant factors based on knowledge from the US model. Precisely, the starting set comprised the Nikkei 225 Volatility Index, the Corporate Bond Spread, the Termspread 5Y-1Y and the Termspread 10Y-1Y. Afterwards, this data set was successively increased by the variable leading to the highest increase of the adjusted R2 and suiting the combination of included economic factors best. This procedure was stopped when no significant improvement could anymore be achieved by model extension. The resulting model is reported in Table 7. The adjusted R2 of this model is 53.17%.

    §See Appendix A for a list of all significant factors.

    PCH stands for percentage change.

    Table 7. Summary of the estimated Asian regression model of the transformed turbulent filtered probabilities y1t+1. Significance at the level of 5% (1% and 0.1% resp.) is denoted by '*' ('**' and '***' resp.)
    estimatestd. errorp value
    (Intercept)-1.10***0.230.000004
    Monetary Base (in PCH)-0.10***0.030.000860
    Nikkei 225 10 Days Momentum-0.0006***1.521e-040.000206
    OECD CLI (in PCH)-8.26***1.260.000000
    TANKAN (small enterprises manufacturing)-0.11***0.010.000000
    Trade Balance (in PCH)0.0004***0.00010.000150
    Corporate Bond Spread * Mutan3.00***0.370.000000
    Corporate Bond Spread5.99***1.020.000000
    * Spread TIBOR 1Y - Gov. Bond 1Y
    TANKAN (small enterprises manufacturing)0.08***0.010.000000
    * Termspread 5Y-1Y
    Corporate Bond Spread0.12***0.020.000000
    * Nikkei 225 Volatility Index
    Mutan * Asia CLI (in PCH)-9.06***1.650.000000
    OECD CLI (in PCH) * Mutan5.37***1.000.000000
    Spread TIBOR 1Y - Gov. Bond 1Y-16.20***3.450.000004
    * Asia CLI (in PCH)
    Corporate Bond Spread8.68***2.550.000776
    * Asia CLI (in PCH)
    OECD CLI (in PCH)-14.43**4.450.001348
    * Spread LIBOR 6M - TIBOR 6M
    Nikkei 225 1M Volatility Index0.04***0.0080.000003
    * Termspread 10Y-1Y
    Nikkei 225 10 Days Momentum-0.0008**0.00020.001233
    * Spread TIBOR 1Y - Gov. Bond 1Y
    OECD CLI (in PCH)-0.01**0.0030.005425
    * Trade Balance (in PCH)
     | Show Table
    DownLoad: CSV

    Table 8 summarizes the second regression model which forecasts the turbulent negative probabilities.|| The adjusted R2 of this model is 53.23%. In summary, both models contain factors covering sentiment, liquidity, leading economic indicators and market confidence.

    ||PCH stands for percentage change.

    Table 8. Summary of the estimated Asian regression model of the transformed turbulent negative filtered probabilities yD,1t+1. Significance at the level of 5% (1% and 0.1% resp.) is denoted by '*' ('**' and '***' resp.)
    estimatestd. errorp value
    (Intercept)-43.77***6.400.000000
    Consumer Confidence Index-3.16**1.010.002058
    Producer Price Index (in PCH)-23.13***5.930.000141
    Termspread 5Y-1Y33.68***6.170.000002
    Termspread 10Y-1Y6.87***2.030.000885
    OECD CLI (in PCH)0.04***0.010.000000
    * Trade Balance (in PCH)
    Termspread 5Y-1Y * US Dollar/JPY-0.28***0.040.000000
    US Dollar/JPY * PMI0.06***0.000.000000
    Consumer Confidence Index9.49***2.260.000044
    * Corporate Bond Spread
    Termspread 10Y-1Y-12.94***3.300.000129
    * Corporate Bond Spread
    US Dollar/JPY * Corporate Bond Spread0.19***0.050.000149
    Termspread 10Y-1Y * Brent0.15***0.040.000185
    Corporate Bond Spread * Asia CLI-8.31***2.380.000607
    Producer Price Index (in PCH) * Brent0.06***0.020.000372
    US Dollar/JPY * Asia CLI0.03**0.010.001102
    Producer Price Index (in PCH)0.16**0.050.002158
    * US Dollar/JPY
    US Dollar/JPY8.457e-06**3.032e-060.005943
    * Nikkei 225 10 Days Momentum
    Termspread 5Y-1Y * Brent-0.19*0.070.011189
     | Show Table
    DownLoad: CSV

    The European logistic regression models were constructed in a similar way and are presented in Table 9 and Table 10.** The adjusted R2 of the first European model is 62.09% and of the second 67.99%.

    **See Appendix B for a list of all significant factors.

    Table 9. Summary of the estimated European regression model of the transformed turbulent filtered probabilities yD,1t+1. Significance at the level of 5% (1% and 0.1% resp.) is denoted by '*' ('**' and '***' resp.)
    estimatestd. errorp value
    (Intercept)-0.400.310.199096
    Termspread 10Y-6M1.82***0.250.000000
    Termspread 5Y-1M-3.63***0.500.000000
    EONIA-3.93***0.520.000000
    Consumer Confidence Index0.10***0.020.000000
    EuroStoxx 50 10 Days Momentum-0.003***0.00080.000666
    VSTOXX-0.04**0.010.002299
    EONIA * CPI0.04***0.010.000000
    VSTOXX * Corporate Bond Spread0.03***0.010.000000
    EONIA * OECD CLI-0.13***0.030.000007
    Corporate Bond Spread * LIBOR Spread-0.50***0.130.000189
    Corporate Bond Spread0.002***0.00060.000730
    * EuroStoxx 50 10 Days Momentum
    Termspread 5Y-1M * VSTOXX0.05***0.010.000101
     | Show Table
    DownLoad: CSV
    Table 10. Summary of the estimated European regression model of the transformed turbulent negative filtered probabilities yD,1t+1. Significance at the level of 5% (1% and 0.1% resp.) is denoted by '*' ('**' and '***' resp.)
    estimatestd. errorp value
    (Intercept)13.87***2.450.000001
    VSTOXX-0.32***0.070.000010
    EONIA * LIBOR Spread4.00***0.700.000001
    LIBOR * EuroStoxx 50 10 Days Momentum0.01***0.0020.000007
    OECD CLI * EuroStoxx 50 1M Volatility0.16***0.030.000000
    VSTOXX * Ifo Geschäftsklima Index-0.02***0.0040.000022
    EONIA * LIBOR-OIS 3M-0.03**0.010.002730
     | Show Table
    DownLoad: CSV

    The first regression models for both the Asian and the European market contain an ARMA(p, q) extension. To fit the parameters, several values for p and q were tested using the Akaike information criterion (AIC) and Bayesian information criterion (BIC). As in Hauptmann et al. (2014), we find that based on these criteria, the choice of p=1, q=0 leads to a better fit than models of higher order. Also as in Hauptmann et al. (2014), no ARMA(p, q) extension is performed for the second regression models.

    Comparing the covariates of the US, Asian and European model, we find that a broad range of economic aspects are reflected in the models. Concerning interest rates, all three models contain termspreads (displaying the form of the yield curve) and corporate bond spreads. In the first model for the European market, the term spreads can be identified measuring the concavity of the term structure in the same way as in the US model (see Hauptmann et al., 2014). In both the model for the European market and for the US market, the corporate bond spreads are included in the first model, but they are not included in the second model. The liquidity in the inter-bank market is reflected by the MUTAN, EONIA and LIBOR in the different markets. Information about the respective stock markets are included via implicit and historical volatility as well as the momentum. While the historical volatility in the US model is both a measure of the current volatility of the market and the market movements from the recent past, it is substituted by the forward looking implied volatility to capture the expected market movements and the momentum to capture the past trading days in the European model and the Asian model. The global economic situation is covered in all three models by the OECD CLI. Additionally, the Major Five Asia CLI accounts especially for the regional influences on the Japanese economy. Other specific factors that reflect the local economies are included by the Trade Balance, the YEN/US Dollar exchange rate, the Producer Price Index, the Consumer Confidence Index and the TANKAN in the Asian model as well as the Consumer Confidence Index and the ifo Geschäftsklima Index in the European model. In the forecasting models, we observe that less input factors are needed to forecast the US stock market compared to the European and Asian market. This can be interpreted by the leading role of the US economy which influences the European and Asian economies as studied in Section 3 using the example of the real estate upswing and subsequent financial crisis originating in the US. On the other hand, the presence of indicators of financial stability such as the LIBOR spread and LIBOR-OIS spread in the European model can be interpreted by the big influence of the financial crisis and European debt crisis as well as the risk emerging from the greater heterogenity of the economies. Finally, the existence of regional economic indicators in the Asian model mentioned above can be interpreted by greater regional characteristics such as Japan's lost decade, which is not reflected in the other markets.


    5. Prediction results

    The estimation results of the two Asian regression models derived in Section 4 are illustrated in Figure 3, whereas the forecast results of the European model can be found in Figure 4.

    Figure 3. One-month-ahead prediction of the two Asian regression models derived in Section 4 from 09/1993–10/2015.
    Figure 4. One-month-ahead prediction of the two European regression models derived in Section 4 from 09/2003–09/2014.

    As can be seen, calm and bullish months are mostly colored in green and yellow while most bearish months are marked red. To measure the quality of the out-of-sample prediction results, various forecast accuracy measures of the filtered probabilities (p derived by the Markov-switching models) and the estimates (ˆp=1/(1+exp(y)) derived by the logistic regression models) are calculated. Besides the root mean squared error (RMSE), defined as

    RMSE=ni=1(piˆpi)2n,

    we also consider the mean absolute error and the median absolute error. The results are shown in Table 11.

    Table 11. Various measures of forecast accuracy of the one-month-ahead prediction of the regression models proposed for the Asian and European market.
    AsiaEurope
    RMSE0.37560.3469
    Mean absolute error0.26780.2682
    Median absolute error0.17070.2023
    Time period09/93-10/1509/03-09/14
     | Show Table
    DownLoad: CSV

    Furthermore, the performance of all three forecasting models is illustrated by a simple asset management case study. We consider the options to invest in the Nikkei 225, the EuroStoxx 50, the S & P 500 and a risk-free asset. For simplicity we take the 1-month US Treasury Constant Maturity Rate†† as risk-free asset and neglect transaction costs. The investment strategy S consists of investing the relative proportion of the forecasted green probability FCI,G plus the forecasted yellow FCI,Y probability in each index I under the condition that the forecasted probability of red FCI,R is not the largest. I.e. the investment ratio invested in index I{Nikkei 225, EuroStoxx 50, S&P 500} at time t equals

    ††The 1-month US Treasury Constant Maturity Rate is available at the Federal Reserve Bank of St. Louis, https://research.stlouisfed.org/fred2/series/DGS1MO.

    VIt={FCI,Gt+FCI,Yt3if max{FCI,G,FCI,Y,FCI,R}FCI,R0if max{FCI,G,FCI,Y,FCI,R}=FCI,R (10)

    and the investment in the risk-free asset rf is given by

    Vrft=1(VNikkei 225t+VS&P 500t+VEuroStoxx 50t). (11)

    The results of the investment strategy are summarized in Table 12 and illustrated in Figure 5. For comparison the results of a Buy-and-Hold-Strategy of the risk-free asset and a 1/n portfolio with respect to the Nikkei 225, the S & P 500 and the EuroStoxx 50 are also listed. Including the information from the warning systems in a simple strategy already increases the return to 5.58% (in contrast to 3.85% for the 1/n portfolio and to 1.41% for the risk-free investment) while reducing the volatility and especially the downside risk compared to the 1/n portfolio essentially. Furthermore, Figure 5 reveals a substantial investment protection by the strategy during the global downturn of the financial crisis.

    Table 12. Comparison of different investment strategies based on the period 08/2004-09/2014 and with an initial portfolio value of 100.
    Risk-free investment1/n portfolioStrategy S
    Terminal value115.32147.76173.68
    Return (% p.a.)1.413.855.58
    Volatility (% p.a.)5.2515.599.85
    mod. Sharpe ratio (p.m.)a-0.06970.1326
    Omega measure Ω (p.m.)b-1.20551.4129
    95%-VaR (p.m.)0-0.0897-0.0454
    95%-CVaR (p.m.)0-0.1094-0.0631
    Maximum drawdown0-0.5544-0.3089
    Note: [a] The modified Sharpe ratio measures the expected excess return of a strategy in terms of the 1-month US Treasury Constant Maturity Rate divided by the standard deviation of the excess returns. [b] The Omega measure denotes the ratio between the expected upside and the expected downside of the excess returns, where the upside (downside) is defined as the positive (negative) excess returns.
     | Show Table
    DownLoad: CSV
    Figure 5. Plot of the development of all three investment strategies compared in Table 12.

    6. Conclusions

    An early warning system for the Asian and European stock market was developed, analogously to the US forecasting model proposed by Hauptmann et al. (2014). The model is based on two Markov-switching models with two regimes each. The thus derived filtered probabilities separate the market in a calm, turbulent positive and turbulent negative market while maintaining the dependence structure of the underlying stock indices. In general, a strong linear dependency of the stock markets can be observed in periods of no locally restricted crisis. Furthermore, logistic regression models using economic covariates were constructed to forecast the filtered probabilities. Via an empirical asset management case study it was shown that the integration of the signs of the warning systems leads to a significantly higher return and reduced volatility and downside risk.


    Acknowledgments

    Janina Engel greatfully acknowledges the support of the Joint Research Centre of the European Commission. Markus Wahl and Rudi Zagst greatfully acknowledge the support of Allianz Global Investors for this research.


    Conflict of Interest

    All authors declare no conflict of interest.


    Appendix


    A. Underlying data set for the Asian model

    In the following, all economic factors that turned out to be significant for the Asian model are given as well as their data sources. If reasonable, some transformations such as the absolute and relative change or the spread between two different variables etc. were also considered.‡‡

    ‡‡For a complete list of all tested economic factors, please contact us.

    TANKAN

    Bank of Japan (http://www.stat-search.boj.or.jp/index_en.html#)

    US Dollar/JPY

    Bank of Japan (http://www.stat-search.boj.or.jp/index_en.html#)

    Mutan ON Call Rate JPY

    Reuters (JPONMU = RR)

    TIBOR-OIS Spread 6M

    Reuters (JPYTIB3-OS6M = R)

    OECD Composite Leading Indicator (OECD + Major 6 NME)

    OECD Statistics (http://stats.oecd.org/index.aspx?datasetcode=MEI_CLI)

    Major Five Asia Composite Leading Indicator

    OECD Statistics (http://stats.oecd.org/index.aspx?datasetcode=MEI_CLI)

    Producer Price Index, All commodities

    Bank of Japan (http://www.stat-search.boj.or.jp/index_en.html#)

    Brent Crude Spot FOB Sullom Voe North SeaUSD

    Reuters (BRT-)

    Corporate Bonds - Government Bond 4Y

    Reuters (Merrill Lynch Japan Corporate Index - JPGOV4YZ = R)

    Nikkei 225 10 Days Momentum

    Federal Reserve Bank of St. Louis

    (https://research.stlouisfed.org/fred2/series/NIKKEI225#)

    (https://research.stlouisfed.org/fred2/series/NIKKEI225#)

    Spread TIBOR 1Y - Government Bond 1Y

    Reuters (TIJPY1YD = - JPGOV1YZ = R)

    Termspread Government Bonds 5Y - 1Y

    Reuters (TIJPY5YD = - JPGOV1YZ = R)

    Termspread Government Bonds 10Y - 1Y

    Reuters (TIJPY10YD = - JPGOV1YZ = R)

    Consumer Confidence Index (CCI)

    OECD Statistics

    (https://data.oecd.org/leadind/consumer-confidence-index-cci.htm#)

    Nikkei Volatility Index

    Reters (.JNIV)

    Monetary Base

    Bank of Japan (http://www.stat-search.boj.or.jp/index_en.html#)

    Real Trade Balance

    Bank of Japan (http://www.stat-search.boj.or.jp/index_en.html#)


    B. Underlying data set for the European model

    In the following, all economic factors that turned out to be significant for the European model are given as well as their data sources. If reasonable, some transformations such as the absolute and relative change or the spread between two different variables etc. were also considered.

    Termspread Government Bonds 10Y - 6M

    ECB Statistical Data Warehouse

    (http://sdw.ecb.europa.eu/browseTable.do?node=qview&SERIES_KEY=165.YC.B.U2.EUR.4F.G_N_A.SV_C_YM.SR_10Y&start=01-01-1999&end=01-10-)

    (http://sdw.ecb.europa.eu/browseTable.do?node=qview&SERIES_KEY=165.YC.B.U2.EUR.4F.G_N_A.SV_C_YM.SR_6M&start=01-01-1999&end=01-10-)

    Termspread Government Bonds 5Y - 1Y

    ECB Statistical Data Warehouse

    (http://sdw.ecb.europa.eu/browseTable.do?node=qview&SERIES_KEY=165.YC.B.U2.EUR.4F.G_N_A.SV_C_YM.SR_5Y&start=01-01-1999&end=01-10-)

    (http://sdw.ecb.europa.eu/browseTable.do?node=qview&SERIES_KEY=165.YC.B.U2.EUR.4F.G_N_A.SV_C_YM.SR_1Y&start=01-01-1999&end=01-10-)

    EURO STOXX 50 Volatility (VSTOXX)

    STOXX (https://www.stoxx.com/index-details?symbol=V2TX

    Corporate Bond Spread

    iBoxx Eur Corporate BBB Index - iBoxx Eur Corporate AAA Index

    Reuters (.IBBEU005E) - (.IBBEU0057)

    LIBOR Spread

    LIBOR 3M - Euro Yield Curve 3M

    global-rates

    (http://www.global-rates.com/interest-rates/libor/european-euro/1989.aspx

    ECB Statistical Data Warehouse

    (http://sdw.ecb.europa.eu/browseTable.do?node=qview&SERIES_KEY=165.YC.B.U2.EUR.4F.G_N_A.SV_C_YM.SR_3M&start=01-01-1999&end=01-10-)

    OECD Composite Leading Indicator (OECD + Major 6 NME)

    OECD Statistics (http://stats.oecd.org/index.aspx?datasetcode=MEI_CLI)

    Consumer Surveys, Confidence indicator, SA

    Reuters ('EUCONS = ECI')

    EuroStoxx 50 10 Days Momentum

    STOXX (https://www.stoxx.com/index-details?symbol=SX5E)

    EuroStoxx 50 1 Month Volatility

    STOXX (https://www.stoxx.com/index-details?symbol=SX5E)

    Euro Zone CPI, All Items

    Reuters (aXZCPI)

    EONIA

    ECB Statistical Data Warehouse

    (http://sdw.ecb.europa.eu/quickview.do?SERIES_KEY=198.EON.D.EONIA_TO.RATE)

    Ifo Geschäftsklima Index

    CES ifo Group

    (http://www.cesifo-group.de/de/ifoHome/facts/Time-series-and-Diagrams/Zeitreihen.html)

    LIBOR-OIS Spread 1M and 3M

    Reuters (EURL-O1M = R) (EURL-O3M = R)


    [1] Abiad A (2003) Early-warning systems: a survey and a regime-switching approach. IMF Work Paper 32.
    [2] Akaike H (1974) A new look at the statistical model identification. IEEE Trans Autom Control 19: 716–723. doi: 10.1109/TAC.1974.1100705
    [3] Barrell R, Davis P, Karim D, et al. (2010) Bank regulation, property prices and early warning systems for banking crises in OECD countries. J Bank Financ 34: 2255–2264. doi: 10.1016/j.jbankfin.2010.02.015
    [4] Baum L, Petrie T, Soules G, et al. (1970) A maximization technique occuring in the statistical analysis of probabilitistic functions of Markov chains. Ann Math Stat 41: 164–171. doi: 10.1214/aoms/1177697196
    [5] Brockwell P, Davis R (1991) Time Series: Theory and Method. 2 Eds., New York: Springer.
    [6] Chen S (2009) Predicting the bear stock market: Macroeconomic variablesas leading indicators. J Bank Financ 33: 211–223. doi: 10.1016/j.jbankfin.2008.07.013
    [7] Davis P, Karim D (2008) Comparing early warning systems for banking crises. J Financ Stabil 4: 89–120. doi: 10.1016/j.jfs.2007.12.004
    [8] Davis P, Karim D (2008) Could early warning systems have helped to predict the sub-prime crisis? Natl Inst Econ Rev 206: 35–47. doi: 10.1177/0027950108099841
    [9] Demirg¨uc-Kunt A, Detragiache E (2005) Cross-country empirical studies of systemic bank distress: a survey. IMF Working Paper 05.
    [10] Dempster A, Laird N, Rubin D (1977) Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc Ser 39: 1–38.
    [11] Diebold F, Lee J, Weinbach G (1989) Regime switching with time-varying transition probabilities. Feder Reserv Bank Philad Work Paper 93: 183–302.
    [12] Duttagupta R, Cashin P (2008) The anatomy of banking crises. IMF Working Paper 08.
    [13] Hamilton J (1989) A new approach to the economic analysis of nonstationary time series and the business cycle. Econometrica 57: 357–384. doi: 10.2307/1912559
    [14] Hamilton J (1994) Time Series Analysis. Princeton: Princeton University Press.
    [15] Hauptmann J, Zagst R (2011) Systemic risk. In: Wu, D. W., Quant Financ Risk Management , Berlin: Springer, 18: 321–338.
    [16] Hauptmann J, Hoppenkamps A, Min A, et al. (2014) Forecasting market turbulence using regimeswitching models. Financ Mark Portf Management 28: 139–164. doi: 10.1007/s11408-014-0226-0
    [17] Kamin S, Schindler J, Samuel S (2001) The contribution of domestic and external factors to emerging market devaluation crises: an early warning systems approach. FRB Int Financ Discuss Paper 711.
    [18] Kaminsky G, Reinhart C (1996) The Twin Crises: The Causes of Banking and Balance-of-Payments Problems. FRB Int Financ Discuss Paper 544.
    [19] Lahiri K, Wang JG (1994) Predicting cyclical turning points with leading index in a markov switching model. J Forecasting 13: 245–263. doi: 10.1002/for.3980130302
    [20] Li WX, Chen C, French J (2015) Toward an early warning system of financial crises: What can indexfutures and options tell us? Q Rev Econ Financ 55: 87–99. doi: 10.1016/j.qref.2014.07.004
    [21] Maheu M, McCurdy T (2000) Identifying bull and bear markets in stock returns. J Bus Econ Stat 18: 100–112.
    [22] Martinez-Peria M (2002) A regime-switching approach to the study of speculative attacks: a focus on EMS crises. Empir Econom 27: 299–334. doi: 10.1007/s001810100102
    [23] Meichle M, Ranaldo A, Zanetti A (2011) Do financial variables help predict the state of the business cycle in small open economies? Evidence from Switzerland. Financ Mark Portf Manag 25: 435– 453. doi: 10.1007/s11408-011-0173-y
    [24] Shao J (2003) Mathematical Statistics. 2 Eds., New York: Springer.
    [25] Timmermann A (2000) Moments of Markov switching models. J Econ 96: 75–111. doi: 10.1016/S0304-4076(99)00051-2
    [26] Wecker WE (1979) Predicting the Turning Points of a Time Series. J Bus 52: 35–50. doi: 10.1086/296032
  • This article has been cited by:

    1. Oscar V. De la Torre-Torres, Francisco Venegas-Martínez, Mᵃ Isabel Martínez-Torre-Enciso, Enhancing Portfolio Performance and VIX Futures Trading Timing with Markov-Switching GARCH Models, 2021, 9, 2227-7390, 185, 10.3390/math9020185
    2. Oscar V. De la Torre-Torres, Evaristo Galeana-Figueroa, José Álvarez-García, A Test of Using Markov-Switching GARCH Models in Oil and Natural Gas Trading, 2019, 13, 1996-1073, 129, 10.3390/en13010129
    3. Oscar V. De la Torre-Torres, Dora Aguilasocho-Montoya, José Álvarez-García, Biagio Simonetti, Using Markov-switching models with Markov chain Monte Carlo inference methods in agricultural commodities trading, 2020, 24, 1432-7643, 13823, 10.1007/s00500-019-04629-5
    4. Oscar V. De la Torre-Torres, Evaristo Galeana-Figueroa, José Álvarez-García, Markov-Switching Stochastic Processes in an Active Trading Algorithm in the Main Latin-American Stock Markets, 2020, 8, 2227-7390, 942, 10.3390/math8060942
    5. Md Akther Uddin, Mohammad Enamul Hoque, Md Hakim Ali, International economic policy uncertainty and stock market returns of Bangladesh: evidence from linear and nonlinear model, 2020, 4, 2573-0134, 236, 10.3934/QFE.2020011
    6. Mengna Zhou, Jizheng Yi, Jieqiong Yang, Yi Sima, Characteristic Representation of Stock Time Series Based on Trend Feature Points, 2020, 8, 2169-3536, 97016, 10.1109/ACCESS.2020.2995958
    7. Lean Yu, Shaodong Liang, Rongda Chen, Kin Keung Lai, Predicting monthly biofuel production using a hybrid ensemble forecasting methodology, 2019, 01692070, 10.1016/j.ijforecast.2019.08.014
    8. Oscar V. De la Torre-Torres, Dora Aguilasocho-Montoya, María de la Cruz del Río-Rama, A Two-Regime Markov-Switching GARCH Active Trading Algorithm for Coffee, Cocoa, and Sugar Futures, 2020, 8, 2227-7390, 1001, 10.3390/math8061001
    9. Oscar V. De la Torre-Torres, Evaristo Galeana-Figueroa, José Álvarez-García, A Markov-Switching VSTOXX Trading Algorithm for Enhancing EUR Stock Portfolio Performance, 2021, 9, 2227-7390, 1030, 10.3390/math9091030
    10. Leticia Bollain-Parra, Oscar V. De la Torre-Torres, Dora Aguilasocho-Montoya, María de la Cruz del Río-Rama, 2021, 978-1-80071-071-9, 141, 10.1108/978-1-80071-070-220211009
    11. Oscar V. De la Torre-Torres, Evaristo Galeana-Figueroa, María de la Cruz Del Río-Rama, José Álvarez-García, Using Markov-Switching Models in US Stocks Optimal Portfolio Selection in a Black–Litterman Context (Part 1), 2022, 10, 2227-7390, 1296, 10.3390/math10081296
    12. Oscar V. De la Torre-Torres, Dora Aguilasocho-Montoya, José Álvarez-García, Testing an Algorithm with Asymmetric Markov-Switching GARCH Models in US Stock Trading, 2021, 13, 2073-8994, 2346, 10.3390/sym13122346
    13. Ender Baykut, Ercan Özen, 2022, 978-1-80262-972-9, 203, 10.1108/S1569-37592022000108B042
    14. Daniel Page, David McClelland, Christo Auret, Pure quantile portfolios on the Johannesburg stock exchange, 2023, 11, 2332-2039, 10.1080/23322039.2023.2231662
    15. Oscar V. De la Torre-Torres, María de la Cruz del Río-Rama, Álvarez-García José, Non-Commodity Agricultural Price Hedging with Minimum Tracking Error Portfolios: The Case of Mexican Hass Avocado, 2024, 14, 2077-0472, 1692, 10.3390/agriculture14101692
    16. Sufang An, Xiangyun Gao, Feng An, Tao Wu, Early warning of regime switching in a complex financial system from a spillover network dynamic perspective, 2025, 28, 25890042, 111924, 10.1016/j.isci.2025.111924
    17. Adhmir Renan Voltolini Gomes, Nelson Hein, Adriana Kroenke, Desenvolvimento da turbulência ambiental sistêmica, 2025, 24, 2237-7662, e3554, 10.16930/2237-766220253554
  • Reader Comments
  • © 2018 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(5470) PDF downloads(1268) Cited by(17)

Article outline

Figures and Tables

Figures(5)  /  Tables(12)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog