Abstract
Compared to the conditional mean or median, conditional quantiles provide a more comprehensive picture of a variable in various scenarios. A semi-parametric quantile estimation method for a double threshold auto-regression with exogenous regressors and heteroskedasticity is considered, allowing representation of both asymmetry and volatility clustering. As such, GARCH dynamics with nonlinearity are added to a nonlinear time series regression model. An adaptive Bayesian Markov chain Monte Carlo scheme, exploiting the link between the quantile loss function and the asymmetric-Laplace distribution, is employed for estimation and inference, simultaneously estimating and accounting for nonlinear heteroskedasticity plus unknown threshold limits and delay lags. A simulation study illustrates sampling properties of the method. Two data sets are considered in the empirical applications: modelling daily maximum temperatures in Melbourne, Australia; and exploring dynamic linkages between financial markets in the US and Hong Kong.
Similar content being viewed by others
References
Bassett G, Koenker R (1982) An empirical quantile function for linear models with iid errors. J Am Stat Assoc 77:407–415
Bollerslev T (1986) Generalized autoregressive conditional heteroscedasticity. J Econom 31:307–327
Brooks C (2001) A double-threshold GARCH model for the French Franc/Deutschmark exchange rate. J Forecast 20:135–143
Cai Y (2010) Forecasting for quantile self-exciting threshold autoregressive time series models. Biometrika 97:199–208
Cai Y, Stander J (2008) Quantile-self exciting threshold time series models. J Time Ser Anal 29:187–202
Chen CWS, Gerlach RH, Lin AMH (2010) Falling and explosive, dormant, and rising markets via multiple-regime financial time series models. Appl Stoch Model Bus Ind 26:28–49
Chen CWS, Gerlach R, So MKP (2006) Comparison of non-nested asymmetric heteroscedastic models. Comput Stat Data Anal 51:2164–2178
Chen CWS, Gerlach R, Wei DCM (2009) Bayesian causal effects in quantiles: accounting for heteroscedasticity. Comput Stat Data Anal Spec Issue Comput Econom 53:1993–2007
Chen CWS, Lin S, Yu PLH (2012) Smooth transition quantile capital asset pricing models with heteroscedasticity. Comput Econ 40:19–48
Chen CWS, So MKP (2006) On a threshold heteroscedastic model. Int J Forecast 22:73–89
Chen C, Sato S (2008) Inhomogeneous jump-GARCH models with applications in financial time series analysis. In: COMPSTAT, Proceedings in computational statistics, pp 217–228
Engle RF (1982) Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation. Econometrica 50:987–1008
Engle RF, Manganelli S (2004) CAViaR: conditional autoregressive value at risk by regression quantiles. J Bus Econ Stat 22:367–381
Eun CS, Shim S (1989) International transmission of stock market movements. J Financ Quant Anal 24:241–256
Galvao AF Jr, Montes-Rojas G, Olmo J (2010) Threshold quantile autoregressive models. J Time Ser Anal 32:253–267
Gerlach R, Chen CWS, Chan NCY (2011) Bayesian time-varying quantile forecasting for value-at-risk in financial markets. J Bus Econ Stat 29:481–492
Giordani P, Kohn R, van Dijk D (2007) A unified approach to nonlinearity, structural change, and outliers. J Econom 137(1):112–133
Gourieroux C, Jasiak J (2008) Dynamic quantile models. J Econom 147:198–205
Hansen BE (1994) Autoregressive conditional density estimation. Int Econ Rev 35:705–730
Hastings WK (1970) Monte-Carlo sampling methods using Markov chains and their applications. Biometrika 57:97–109
Hyndman RJ, Bashtannyk DM, Grunwald GK (1996) Estimating and visualizing conditional densities. J Comput Graph Stat 5:315–336
Huang D, Yu B, Lu Z, Fabozzi F, Forcardi S, Fukushima M (2010) Index-Exciting CAViaR: a new empirical time-varying risk model. Stud Nonlinear Dyn Econom 14:1–24
Karolyi GA (1995) A multivariate GARCH model of international transmissions of stock returns and volatility: the case of the United States and Canada. J Bus Econ Stat 13:11–25
Koenker R (2000) Galton, Edgeworth, Frisch and prospects for quantile regression in econometrics. J Econom 95:347–374
Koenker R (2005) Quantile regression. Cambridge University Press, Cambridge
Koenker R, Bassett G (1978) Regression quantiles. Econometrica 46:33–50
Koenker R, Machado J (1999) Goodness of fit and related inference processes for quantile regression. J Am Stat Assoc 94:1296–1310
Koenker R, Xiao Z (2004) Unit root quantile autoregression inference. J Am Stat Assoc 99:775–87
Koenker R, Xiao Z (2006) Quantile autoregression. J Am Stat Assoc 101:980–990
Koenker R, Zhao Q (1996) Conditional quantile estimation and inference for ARCH models. Econom Theory 12:793–813
Li WK, Lam K (1995) Modelling asymmetry in stock returns by a threshold ARCH model. Statistician 44:333–341
Li CW, Li WK (1996) On a double-threshold autoregressive heteroscedastic time series model. J Appl Econom 11:253–274
McLeod AI, Li WK (1983) Diagnostic checking ARMA time series models using squared-residual autocorrelations. J Time Ser Anal 4:269–273
Metropolis N, Rosenbluth AW, Rosenbluth MN, Teller E (1953) Equations of state calculations by fast computing machines. J Chem Phys 21:1087–1091
Petruccelli JD, Woolford SW (1984) A threshold AR(1) model. J Appl Probab 21:270–286
Spiegelhalter DJ, Best NG, Carlin BP, Van der Linde A (2002) Bayesian measures of model complexity and fit. J R Stat Soc Ser B 64:583–639
Tong H (1978) On a threshold model. In: Chen CH (ed) Pattern recognition and signal. Sijtho& Noordho, Amsterdam
Tong H, Lim KS (1980) Threshold autoregression, limit cycles and cyclical data (with discussion). J R Stat Soc Ser B 42:245–292
Tsay RS (1989) Testing and modeling threshold autoregressive processes. J Am Stat Assoc 84:231–240
Tsay RS (1998) Testing and modeling multivariate threshold models. J Am Stat Assoc 93:1188–1202
Weiss A (1987) Estimating nonlinear dynamic models using least absolute error estimation. Econom Theory 7:46–68
Xiao Z, Koenker R (2009) Conditional quantile estimation for GARCH Models. J Am Stat Assoc 104:1696–1712
Yu K, Lu Z, Stander J (2003) Quantile regression: applications and current research area. Statistician 52:331–350
Yu K, Moyeed RA (2001) Bayesian quantile regression. Stat Probab Lett 54:437–447
Acknowledgments
We thank the editor, associate editor, and two anonymous referees for their insightful and helpful comments, which improved this paper. Cathy Chen is supported by the grants: NSC 99-2118-M-035 -001 -MY2 from the National Science Council (NSC) of Taiwan.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
Derivation of the standardized SL density and corresponding likelihood expression in (5) is given below. Consider that a random variable \(z\) follows a SL distribution with density
where \(\mu \) is the mode and \(\delta \) is a scale parameter. The mean and variance of a \(SL(0,\delta ,\tau )\) are
Suppose that the model error \(\varepsilon _t\) is equal to the standardized \(z\), i.e.
The density function for \(\varepsilon _t\) is then:
where \(|J|=|\frac{\partial z}{\partial \varepsilon _t}|=\frac{\sqrt{1-2\tau +2\tau ^2}}{\tau (1-\tau )}\delta \)
Moreover, since \(\varepsilon _t=(y_t-\acute{Y_{t-1}}\phi (\tau )-\acute{X_{t-1}}\psi (\tau ))/(\sqrt{h_t})\) and \(\frac{\partial \varepsilon _t}{\partial y_t}=(\sqrt{h_t})^{-1}\), the likelihood function for this model becomes as given in (5).
Rights and permissions
About this article
Cite this article
Chen, C.W.S., Gerlach, R. Semi-parametric quantile estimation for double threshold autoregressive models with heteroskedasticity. Comput Stat 28, 1103–1131 (2013). https://doi.org/10.1007/s00180-012-0346-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00180-012-0346-9