# zbMATH — the first resource for mathematics

On stationarity and ergodicity of the bilinear model with applications to GARCH models. (English) Zbl 1224.62063
The author deals with the process $$\{y_t,\;t=1,2,\dots\}$$ which is a solution of the so-called subdiagonal bilinear model introduced by C. W. J. Granger and A. P. Andersen [Angewandte Statistik und Ökonometrie. Heft 8. Göttingen: Vandenhoeck & Ruprecht. 94 S. (1978; Zbl 0379.62074)]: $y_t=a_0+\sum_{i=1}^{p}a_iy_{t-i}+ \sum_{i=1}^{q}b_i\varepsilon_{t-i}+\sum_{i=1}^{Q}\sum_{j=1}^{P}c_{i,j}y_{t-i-j}\varepsilon_{t-i},$ where $$\{\varepsilon_{t}\}$$ are i.i.d. random variables. This bilinear time series can be represented as $$y_t = Z_{1,t-1}+b_0\varepsilon_{t}$$, where the process $$Z_t = (Z_{1,t},\dots, Z_{n,T})'\in\mathbb R^n$$, $$n =\max\{p,P + q,P + Q\}$$, solves the random coefficient autoregressive (RCA) equation $$Z_t=A_tZ_{t-1}+B_t$$, where the random matrix $$A_t$$ is of the form $$A_t=A_0+A_1\varepsilon_t$$ and the random vector $$B_t=B_0+B_1\varepsilon_t+B_2\varepsilon^2_t$$. For a sequence of i.i.d. matrices $$\{A_t\}$$ with $$E[\log^+(| | A_t| | )]<\infty$$ the associated Lyapunov exponent $$\gamma$$ is defined by $$\gamma=\lim_{t\to\infty}t^{-1}E[\log\| A_t\cdots A_1\| ]$$. Under the condition that the errors $$\{\varepsilon_{t}\}$$ are i.i.d. random variables with $$E[\log^+(\varepsilon^2_{t})]<\infty$$, the author proved that $$\gamma<0$$ is a sufficient condition for the above bilinear model to have a unique stationary solution. If, furthermore, the model is irreducible, see P. Bougerol and N. Picard, Ann. Probab. 20, No. 4, 1714–1730 (1992; Zbl 0763.60015), then $$\gamma<0$$ is a necessary condition for stationarity. For two special subcases of the general subdiagonal bilinear model precise necessary and sufficient conditions for stationarity are given. These results are used to give necessary and sufficient conditions for the existence of stationary solutions for two classes of GARCH$$(p,q)$$ models that can be written in the form of a bilinear time-series model that in turn can be written as an RCA model. These two classes include, among others, the linear GARCH model [R.F. Engle, Econometrica 50, 987–1007 (1982; Zbl 0491.62099); T. Bollerslev, J. Econom. 31, 307–327 (1986; Zbl 0616.62119)], Power GARCH [R. F. Engle and T. Bollerslev, Econ. Rev. 5, 1–50 (1986; Zbl 0619.62105)], LGARCH [P. M. Robinson, J. Econ. 47, No. 1, 67–84 (1991; Zbl 0734.62070)], Log- GARCH, EGARCH [D. B. Nelson, Econometrica 59, No. 2, 347–370 (1991; Zbl 0722.62069)], Asymmetric/GJR GARCH, T-GARCH [S. Taylor, Modelling financial time series. Chichester: Wiley (1986; Zbl 1130.91345); J.-M. Zakoïan, J. Econ. Dyn. Control 18, No. 5, 931–955 (1994; Zbl 0875.90197)]. Simple conditions are given for each of these specifications to have stationary solutions. Since these conditions are not only sufficient but also necessary, they are the weakest possible.

##### MSC:
 62M10 Time series, auto-correlation, regression, etc. in statistics (GARCH) 60G10 Stationary stochastic processes
Full Text:
##### References:
 [1] Alsmeyer, On the Harris recurrence of iterated random Lipschitz functions and related convergence rate results, Journal of Theoretical Probability 16 pp 217– (2003) · Zbl 1022.60067 [2] Bera, ARCH models: properties and estimation, Journal of Economic Surveys 7 pp 305– (1993) [3] Bollerslev, Generalized autoregressive conditional heteroscedasticity, Journal of Econometrics 31 pp 307– (1986) · Zbl 0616.62119 [4] Bougerol, Strict stationarity of generalized autoregressive processes, Annals of Probability 20 pp 1714– (1992) · Zbl 0763.60015 [5] Bougerol, Stationarity of GARCH processes and of some nonnegative time series, Journal of Econometrics 52 pp 115– (1992) · Zbl 0746.62087 [6] Brandt, The stochastic equation Yn+1 = AnYn+Bn with stationary coefficients, Advances in Applied Probability 18 pp 211– (1986) [7] Brockwell, Time Series: Theory and Methods (1991) · Zbl 0709.62080 · doi:10.1007/978-1-4419-0320-4 [8] Carrasco, Mixing and moment properties of various GARCH and stochastic volatility models, Econometric Theory 18 pp 17– (2002) · Zbl 1181.62125 [9] Cline, A note on a simple markov bilinear stochastic process, Statistics and Probability Letters 56 pp 283– (2002) · Zbl 0999.60067 [10] Ding, A long memory property of stock market returns and a new model, Journal of Empirical Finance 1 pp 83– (1993) [11] Duan, Augmented GARCH(p, q) process and its diffusion limit, Journal of Econometrics 79 pp 97– (1997) · Zbl 0898.62141 [12] Elton, A multiplicative ergodic theorem for lipschitz maps, Stochastic Processes and Their Applications 34 pp 39– (1990) · Zbl 0686.60028 [13] Engle, Autoregressive conditional heteroscedasticity with estimates of United Kingdom inflation, Econometrica 50 pp 987– (1982) · Zbl 0491.62099 [14] Engle, Modelling the persistence of conditional variances, Econometric Reviews 5 pp 1– (1986) · Zbl 0619.62105 [15] Engle, Measuring and testing the impact of news on volatility, Journal of Finance 48 pp 1749– (1993) [16] Franq, Mixing properties of a general class of GARCH(1,1) models without moment assumptions, Econometric Theory 22 pp 815– (2006) [17] Geweke, Modelling the persistence of conditional variances: comment, Econometric Reviews 5 pp 57– (1986) [18] Giraitis, LARCH, leverage and long memory, Journal of Financial Econometrics 2 pp 177– (2004) [19] Glosten, Relationship between the expected value and the volatility of the nominal excess return on stocks, Journal of Finance 48 pp 1779– (1993) [20] Granger, An Introduction to Bilinear Time Series Models (1978) · Zbl 0379.62074 [21] Guegan, Probabilistic properties of the $$\beta$$-ARCH model, Statistica Sinica 4 pp 71– (1994) [22] Hall, Inference in ARCH and GARCH models, Econometrica 71 pp 285– (2003) · Zbl 1136.62368 [23] Hentschel, All in the Family: nesting symmetric and asymmetric GARCH models, Journal of Financial Economics 39 pp 71– (1995) [24] Kazakevicius, On stationarity in the ARCH() model, Econometric Theory 18 pp 1– (2002) [25] Kingman, Subadditive ergodic theory, Annals of Probability 1 pp 883– (1973) · Zbl 0311.60018 [26] Kristensen (2007) [27] Liu, On stationarity and asymptotic inference of bilinear time series models, Statistica Sinica 2 pp 479– (1992) · Zbl 0826.60017 [28] Meitz, Ergodicity, mixing and existence of moments of a class of markov models with applications to GARCH and ACD models, Econometric Theory 24 pp 1291– (2008) · Zbl 1284.62566 [29] Meyn, Asymptotic behavior of stochastic systems possessing markovian realizations, SIAM Journal of Control and Optimization 29 pp 535– (1991) · Zbl 0725.60070 [30] Mittnik, stationarity of stable power-GARCH processes, Journal of Econometrics 106 pp 97– (2002) [31] Nelson, Stationarity and persistence in the GARCH(1,1) model, Econometric Theory 6 pp 318– (1990) [32] Nelson, Conditional heteroskedasticity in asset returns: a new approach, Econometrica 59 pp 347– (1991) · Zbl 0722.62069 [33] Pantula, Modelling the persistence of conditional variances: comment, Econometric Reviews 5 pp 71– (1986) [34] Peng, Least absolute deviations estimation for ARCH and GARCH models, Biometrika 90 pp 967– (2003) · Zbl 1436.62439 [35] Pham, Bilinear Markovian representation of bilinear models, Stochastic Processes and Their Applications 20 pp 295– (1985) · Zbl 0588.62162 [36] Pham, The mixing property of bilinear and generalised random coefficient autoregressive models, Stochastic Processes and Their Applications 23 pp 291– (1986) · Zbl 0614.60062 [37] Pham, Dimension estimation and models pp 191– (1993) [38] Robinson, Testing for strong serial correlation and dynamic conditional heteroskedasticity in multiple regression, Journal of Econometrics 47 pp 67– (1991) · Zbl 0734.62070 [39] Schwert, Why does stock market volatility change over time, Journal of Finance 44 pp 1115– (1989) [40] Straumann, Quasi-maximum-likelihood estimation in conditionally heteroscedastic time series: a stochastic recurrence equations approach, Annals of Statistics 34 pp 2449– (2006) · Zbl 1108.62094 [41] Taylor, Modelling financial time series (1986) [42] Tong, A note on a Markov bilinear stochastic process in discrete time, Journal of Time Series Analysis 2 pp 279– (1981) · Zbl 0548.60071 [43] Zakoian, Threshold heteroskedastic models, Journal of Economic Dynamics and Control 18 pp 931– (1994) · Zbl 0806.90018
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.