Chapter 5 Univariate time-series analysis. () Chapter 5 Univariate time-series analysis 1 / 59

Transcription

1 Chapter 5 Univariate time-series analysis () Chapter 5 Univariate time-series analysis 1 / 59

2 Time-Series Time-series is a sequence fx 1, x 2,..., x T g or fx t g, t = 1,..., T, where t is an index denoting the period in time in which x occurs. We shall treat x t as a random variable; hence, a time-series is a sequence of random variables ordered in time. Such a sequence is known as a stochastic process. The probability structure of a sequence of random variables is determined by the joint distribution of a stochastic process. The simplest possible probability model for such a joint distribution is: x t = α + ɛ t, ɛ t n.i.d. 0, σ 2 ɛ i.e., x t is normally independently distributed over time with constant variance and mean equal to α. In other words, x t is the sum of a constant and a white-noise process. If a white-noise process were a proper model for financial time-series, forecasting would not be very interesting as the best forecast for the moments of the relevant time series would be their unconditional moments. () Chapter 5 Univariate time-series analysis 2 / 59,

3 Better models The model: x t = α + ɛ t, ɛ t n.i.d. 0, σ 2 ɛ, ˆ α = 1 T ˆ 2 T 1 T x t, σɛ = x t ˆα T t=1 t=1 2 Reflect the traditional approach to portfolio allocation, but it does not reflect the data. At high frequency the variance is not constant and predictable, at low frequency returns are persistent and predictable. To construct more realistic models, we concentrate on univariate models first to consider then multivariate models. () Chapter 5 Univariate time-series analysis 3 / 59

4 Better models GER Stock Market Returns monthly ret monthly ret sim WN x 10 3 GER 10Y Bond YTM 5 10Y YTM 10Y YTM sim WN While the CER gives a plausible representation for the 1-month returns, the behaviour over time of the YTM of the 10-Year bond does not resemble at all that of the simulated data. () Chapter 5 Univariate time-series analysis 4 / 59

5 ARMA modelling A more general and more flexible class of models emerges when combinations of ɛ t are used to model x t. We concentrate on a class of models created by taking linear combinations of the white noise, the autoregressive moving average (ARMA) models: AR(1) : x t = ρx t 1 + ɛ t, MA(1) : x t = ɛ t + θɛ t 1, AR(p) : x t = ρ 1 x t 1 + ρ 2 x t ρ p x t p + ɛ t, MA(q) : x t = ɛ t + θ 1 ɛ t θ q ɛ t q, ARMA(p, q) : x t = ρ 1 x t ρ p x t p + θ 1 ɛ t θ q ɛ t q. () Chapter 5 Univariate time-series analysis 5 / 59

6 Analysing time-series models To illustrate empirically all fundamentals we consider a specific member of the ARMA family, the AR model with drift, x t = ρ 0 + ρ 1 x t 1 + ɛ t, (1) ɛ t n.i.d. 0, σ 2 ɛ. Given that each realization of our stochastic process is a random variable, the first relevant fundamental is the density of each observation. In particular, we distinguish between conditional and unconditional densities. () Chapter 5 Univariate time-series analysis 6 / 59

7 Conditional and Unconditional Densities The unconditional density is obtained under the hypothesis that no observation on the time-series is available, while conditional densities are based on the observation of some realization of random variables. In the case of time-series, we derive unconditional density by putting ourselves at the moment preceding the observation of any realization of the time-series. At that moment the information set contains only the knowledge of the process generating the observations. As observations become available, we can compute conditional densities. () Chapter 5 Univariate time-series analysis 7 / 59

8 Conditional Densities Consider the AR(1) model. The moments of the density of x t conditional upon x t 1 are immediately obtained from the relevant process: E (x t j x t 1 ) = ρ 0 + ρ 1 x t 1, Var (x t j x t 1 ) = σ 2 ɛ, Cov (x t j x t 1 ), x t j j x t j 1 = 0 for each j. To derive the moments of the density of x t conditional upon x t need to substitute x t 2 from (1) for x t 1 : 2, we E (x t j x t 2 ) = ρ 0 + ρ 0 ρ 1 + ρ 2 1 x t 2, Var (x t j x t 2 ) = σ 2 ɛ 1 + ρ 2 1, Cov (x t j x t 2 ), x t j j x t j 2 = ρ 1 σ 2 ɛ, for j = 1, Cov (x t j x t 2 ), x t j j x t j 2 = 0, for j > 1. () Chapter 5 Univariate time-series analysis 8 / 59

9 Unconditional Densities Unconditional moments are derived by substituting recursively from to express x t as a function of information available at time t 0, the moment before we start observing realizations of our process. E (x t ) = ρ ρ 1 + ρ 2 1 Var (x t ) = σ 2 ɛ 1 + ρ ρ ρt ρ2t 2 1 γ (j) = Cov x t, x t j = ρ j 1 Var (x t), ρ (j) = + ρ t 1 x 0,, Cov x t, x t p Var (xt ) Var (x t 1 ) = ρ j 1 Var (x t) p Var (xt ) Var (x t 1 ). j Note that γ (j) and ρ (j) are functions of j, known respectively as the autocovariance function and the autocorrelation function. () Chapter 5 Univariate time-series analysis 9 / 59

10 Insert Clicker 1 here () Chapter 5 Univariate time-series analysis 10 / 59

11 Stationarity A stochastic process is strictly stationary if its joint density function does not depend on time. More formally, a stochastic process is stationary if, for each j 1, j 2,..., j n, the joint distribution, f x t, x t+j1, x t+j2, x t+jn, does not depend on t. A stochastic process is covariance stationary if its two first unconditional moments do not depend on time, i.e. if the following relations are satisfied for each h, i, j: E (x t ) = E (x t+h ) = µ, E x 2 t = E x 2 t+h = µ 2, E x t+i x t+j = µ ij. () Chapter 5 Univariate time-series analysis 11 / 59

12 Insert Clicker 2 here () Chapter 5 Univariate time-series analysis 12 / 59

13 Stationarity In the case of our AR(1) process, the condition for stationarity is jρ 1 j < 1. When such a condition is satisfied, we have: Cov x t, x t E (x t ) = E (x t+h ) = ρ 0 1 ρ 1, Var (x t ) = Var (x t+h ) = σ2 ɛ 1 ρ 2, 1 j = ρ j 1 Var (x t). On the other hand, when jρ 1 j = 1, the process is obviously non-stationary: Cov x t, x t E (x t ) = ρ 0 t + x 0, Var (x t ) = σ 2 ɛt, j = σ 2 ɛ (t j). () Chapter 5 Univariate time-series analysis 13 / 59

14 General ARMA processes The Wold decomposition theorem warrants that any stationary stochastic process can be expressed as the sum of a deterministic and a stochastic moving-average component: x t = ɛ t + b 1 ɛ t 1 + b 2 ɛ t b n ɛ t n = 1 + b 1 L + b 2 L b n L n ɛ t = b(l)ɛ t, Represent the polynomial b(l) as the ratio of two polynomials of lower order: x t = b (L) ɛ t = a (L) c (L) ɛ t, c (L) x t = a (L) ɛ t. (2) This is an ARMA process. Stationary requires that the roots of c (L) lie outside the unit circle. Invertibility of the MA component require that the roots of a (L) lie outside the unit circle. () Chapter 5 Univariate time-series analysis 14 / 59

15 General ARMA processes Consider the simplest case, the ARMA(1,1) process: x t = c 1 x t 1 + ɛ t + a 1 ɛ t 1, (1 c 1 L) x t = (1 + a 1 L) ɛ t. The above equation is equivalent to: x t = 1 + a 1L 1 c 1 L ɛ t = (1 + a 1 L) = 1 + c 1 L + (c 1 L) ɛ t h 1 + (a 1 + c 1 ) L + c 1 (a 1 + c 1 ) L 2 + c 2 1 (a 1 + c 1 ) L i ɛ t. Which shows that the ratio of two finite lag polynomials allows us to model an infinite lag polynomial. () Chapter 5 Univariate time-series analysis 15 / 59

16 General ARMA processes We then have, Var (x t ) = = Cov (x t, x t 1 ) = = Hence, h i 1 + (a 1 + c 1 ) 2 + c 2 1 (a 1 + c 1 ) σ 2 ɛ " # 1 + (a 1 + c 1 ) 2 1 c 2 σ 2 ɛ, 1 h i (a 1 + c 1 ) + c 1 (a 1 + c 1 ) + c 2 1 (a 1 + c 1 ) +... " # (a 1 + c 1 ) + c 1 (a 1 + c 1 ) 2 1 c 2 σ 2 ɛ. 1 ρ (1) = Cov (x t,x t 1 ) Var (x t ) = (1 + a 1c 1 ) (a 1 + c 1 ) 1 + c a 1c 1. σ 2 ɛ () Chapter 5 Univariate time-series analysis 16 / 59

17 General ARMA processes For example, suppose c (L) x t = a (L) ɛ t and you want to find x t = d (L) ɛ t. Parameters in d (L) are most easily found by writing c (L) d (L) = a (L) and by matching terms in L j. For an illustration suppose a (L) = 1 + a 1 L, c (L) = 1 + c 1 L. Multiplying out d (L) we have (1 + c 1 L) 1 + d 1 L + d 2 L d n L n = 1 + a 1 L Matching powers of L, d 1 = a 1 c 1 c 1 d 1 + d 2 = 0 c 1 d 2 + d 3 = 0 c 1 d n 1 + d n = 0 x t = ɛ t + (a 1 c 1 ) ɛ t 1 c 1 (a 1 c 1 ) ɛ t ( c 1 ) n 1 (a 1 c 1 ) ɛ t n () Chapter 5 Univariate time-series analysis 17 / 59

18 Insert Clicker 3 here () Chapter 5 Univariate time-series analysis 18 / 59

19 Persistence and the linear model Persistence of time-series destroys one of the crucial properties for implementing valid estimation and inference in the linear model. In the context of the linear model y = Xβ + ɛ. The following property is required to implement valid estimation and inference E (ɛ j X) = 0. (3) Hypothesis (3) implies that E (ɛ i j x 1,...x i,..., x n ) = 0, (i = 1,..., n). Think of the simplest time-series model for a generic variable y: y t = a 0 + a 1 y t 1 + ɛ t. Clearly, if a 1 6= 0, then, although it is still true that E (ɛ t j y t 1 ) = 0, E (ɛ t 1 j y t 1 ) 6= 0 and (3) breaks down. () Chapter 5 Univariate time-series analysis 19 / 59

20 How serious is the problem? To assess intuitively the consequences of persistence, we construct a small Monte-Carlo simulation on the short sample properties of the OLS estimator of the parameters in an AR(1) process. A Monte-Carlo simulation is based on the generation of a sample from a known data generating process (DGP). First we generate a set of random numbers from a given distribution (here a normally independent white-noise disturbance) for a sample size of interest (say 200 observations) and then construct the process of interest (in our case, an AR(1) process). When a sample of observations on the process of interest is available, then we can estimate the relevant parameters and compare their fitted values with the known true value. the Monte-Carlo simulation is a sort of controlled experiment. To overcome the potential dependence of the set of random numbers drawn on the sequence of simulated white-noise residuals, the DGP is replicated many times. () Chapter 5 Univariate time-series analysis 20 / 59

21 From the figure we note that the estimate of a 1 is heavily biased in small samples, but the bias decreases as the sample gets larger, and disappears eventually. One can show analytically that the average of 2 the OLS estimate of a is a 1 () Chapter 5 Univariate. time-series analysis 21 / 59 How serious is the problem? We report the averages across replications in the following figure A1MEAN TRUEA1 Figure: Small sample bias

22 Asymptotic theory Stationary time-series feature time-independent distributions, as a consequence, the effect of any specific innovation disappears as time elapses. We show in this section that the intuition of the simple Monte-Carlo simulation can be extended and asymptotic theory can be used to perform valid estimation and inference when modelling stationary time-series. Consider a sequence fx T g of random variables with the associated sequence of distribution functions ff T g = F 1,..., F T, we give the following definitions of convergence for X T. () Chapter 5 Univariate time-series analysis 22 / 59

23 Convergence Given a random variable X with distribution function F, X T converges in distribution to X if the following equality is satisfied: lim T! P fx T < x 0 g = P fx < x 0 g, for all x 0, where the function F(x) is continuous. Given a random variable X with distribution function F, X T converges in probability to X if, for each ɛ > 0, the following relation holds: lim P fjx T Xj < ɛg = 1. T! Note that convergence in probability implies convergence in distribution. () Chapter 5 Univariate time-series analysis 23 / 59

24 Central limit theorem Given a sequence fx T g of identically and independently distributed random variables with mean µ and finite variance σ 2, define _ X = 1 T T i=1 ω = p T X i, _X σ µ ω converges in distribution to a standard normal. For any random variable X T, such that p lim X T = a, where a is a constant, given a function g() continuous at a, p lim g (X T ) = g (a).. () Chapter 5 Univariate time-series analysis 24 / 59

25 Cramer s theorem Given two random variables X T and Y T, such that Y T converges in distribution to Y and X T converges in probability to a constant a, the following relationships hold: X T + Y T converges in distribution to (a + Y); Y T /a T converges in distribution to (Y/a); Y T a T converges in distribution to (Y a). Note that all theorems introduced so far extend to vectors of random variables. () Chapter 5 Univariate time-series analysis 25 / 59

26 Mann-Wald Consider a vector z t (k 1) of random variables which satisfies the following property: p lim T 1 T z t zt 0 = Q, t=1 where Q is a positive definite matrix. Consider also a sequence ɛ t of random variables, identically and independently distributed with zero mean and finite variance, σ 2, for which finite moments of each order are defined. If E (z t ɛ t ) = 0, then r p lim T 1 T 1 z t ɛ t = 0, T t=1 T t=1 d z t ɛ t! N 0, σ 2 Q. () Chapter 5 Univariate time-series analysis 26 / 59

27 Insert Clicker 4 here () Chapter 5 Univariate time-series analysis 27 / 59

28 Models for stationary time-series Consider the following time-series model: y t = γy t 1 + ɛ t, where y t is a stationary variable and jγj < 1. As already shown, E (y t ɛ t i ) 6= 0 and the OLS estimator of γ is biased. By applying the Mann Wald result, we can derive the asymptotic distribution of the OLS estimator of γ, bγ: h bγ! d N γ, σ 2 Q 1i, and all the finite sample results available for cross-section can be extended to stationary time-series just by considering large-sample theory. () Chapter 5 Univariate time-series analysis 28 / 59

29 The Maximum Likelihood Method The likelihood function is the joint probability distribution of the data, treated as a function of the unknown coefficients The maximum likelihood estimator (MLE) consists of value of the coefficients that maximize the likelihood function The MLE selects the value of parameters to maximize the probability of drawing the data that have been effectively observed () Chapter 5 Univariate time-series analysis 29 / 59

30 MLE of an MA process Consider an MA process for a return r t+1 : r t+1 = θ 0 + ε t+1 + θ 1 ε t The time series of the residuals can be computed as ε t+1 = r t+1 θ 0 θ 1 ε t ε 0 = 0 If ε t+1 is normally distributed, than we have! 1 f (ε t+1 ) = (2πσ 2 ε ) 1/2 exp ε2 t+1 2σ 2 ε () Chapter 5 Univariate time-series analysis 30 / 59

31 MLE of an MA process If the ε t+1 are independent over time the likelihood function can be written as follows f (ε 1, ε 2,...ε t+1 ) = T Π i=1 f (ε i ) = Π T 1 exp ε2 i=1 (2πσ 2 1/2 ε ) i 2σ 2 ε! The MLE chooses θ 0, θ 1, σ 2 ε to maximize the probability that the estimated model has generated the observed data-set. The optimum is not always found analically, iterative search is the standard method. () Chapter 5 Univariate time-series analysis 31 / 59

32 MLE of an AR process Consider a vector x t containing observations on time-series variables at time t. A sample of T time-series observations on all the variables is represented as: 2 3 x 1 X 1. T = x T In general, estimation is performed by considering the joint sample density function, known also as the likelihood function, which can be expressed as D X 1 T j X 0, θ. The likelihood function is defined on the parameter space ˆ, given the observation of the observed sample X 1 T and of a set of initial conditions X 0. One can interpret such initial conditions as the pre-sample observations on the relevant variables (which are usually unavailable). () Chapter 5 Univariate time-series analysis 32 / 59

33 MLE of an AR process In case of independent observations the likelihood function can be written as the product of the density functions for each observation. However, this is not the relevant case for time-series, as time-series observations are in general sequentially correlated. In the case of time-series, the sample density is constructed using the concept of sequential conditioning. The likelihood function, conditioned with respect to initial conditions, can always be written as the product of a marginal density and a conditional density: Obviously, D X 1 T j X 0, θ = D (x 1 j X 0, θ) D X 2 T j X 1, θ. D X 2 T j X 0, θ = D (x 2 j X 1, θ) D X 3 T j X 2, θ, and, by recursive substitution: D X 1 T j X 0, θ = T D (x t j X t 1, θ). t=1 () Chapter 5 Univariate time-series analysis 33 / 59

34 MLE of an AR process Having obtained D X 1 T j X 0, θ, we can in theory derive D X 1 T, θ by integrating with respect to X 0 the density conditional on pre-sample observations. In practice this could be intractable analytically, as D (X 0 ) is not known. The hypothesis of stationarity becomes crucial at this stage, as stationarity restricts the memory of time-series and limits the effects of pre-sample observations to the first observations in the sample. This is why, in the case of stationary processes, one can simply ignore initial conditions. Clearly, the larger the sample, the better, as the weight of lost information becomes smaller. Moreover, note that even by omitting initial conditions, we have: D X 1 T T j X 0, θ = D (x 1 j X 0, θ) D (x t j X t 1, θ). Therefore, the likelihood function is separated in the product on T 1 conditional distributions and one unconditional distribution. In the case of non-stationarity, the unconditional distribution is undefined. On the other hand, in the case of stationarity, the DGP is completely () Chapter 5 Univariate time-series analysis 34 / 59 t=2

35 Insert Clicker 5 here () Chapter 5 Univariate time-series analysis 35 / 59

36 To give more empirical content to our case, let us consider again the case of the univariate first-order autoregressive process, X t j X t 1 N λx t 1, σ 2, (4) D X 1 T j λ, σ 2 = D X 1 j λ, σ 2 T t=2 D X t j X t 1, λ, σ 2. (5) From (5), the likelihood function clearly involves T 1 conditional densities and one unconditional density. The conditional densities are given by (4), the unconditional density can be derived only in the case of stationarity: x t = λx t 1 + u t, u t N.I.D 0, σ 2. () Chapter 5 Univariate time-series analysis 36 / 59

37 We can obtain by recursive substitution: x t = u t + λu t λ n 1 u 1 + λ n x 0. Only if jλj < 1, the effect of the initial condition disappears and we can write the unconditional density of x t as: D x t j λ, σ 2 σ = 2 N 0, 1 λ 2. Under stationarity we can derive the exact likelihood function: D X 1 T j λ, σ 2 = (2π) T 2 σ T 1 λ exp " 1 2σ 2 1 λ 2 x T t=2 (x t λx t 1 ) 2!# and estimates of the parameters of interest are derived by maximizing this function. Note that bλ cannot be derived analytically, using the exact likelihood function; but it requires conditioning the likelihood and operating a grid search. () Chapter 5 Univariate time-series analysis 37 / 59,

38 Putting ARMA models at work There are four main steps in the Box-Jenkins approach: PRE WHITENING: make sure that the time series is stationary.. MODEL SELECTION: Information criteria are a useful tool to this end. The Akaike s information criteria (AIC) and the.schwarz Bayesian Criterion (SBC) are the most commonly used criteria: AIC = 2 log(l) + 2(p + q) SBC = 2 log(l) + log(n)(p + q) MODEL CHECKING: residual tests. Make sure that residuals are not autocorrelated and check whether their distribution is normal, also ex-post evaluation technique based on RMSE and MAE are implemented (Diebold-Mariano, Giacomini-White). FORECASTING, the selected model is typically simulated forward after estimation of the estimation of parameters to produce forecasts for the variable of interests at the relevant horizon. () Chapter 5 Univariate time-series analysis 38 / 59

39 An Illustration To illustrate how ARMA model can be put at work consider the case of forecasting the YTM of 10-year German bonds from 1994:1 onward, given the availability of data over the period Estimation can be performed in MATLAB by using the appropriate specification in the GARCH procedure. The ML estimation delivers the following results: y 10 t = ( ) ( ) y10 t 1 + ( )0.5 (3.2616e 08) 0.5 ˆ u t () Chapter 5 Univariate time-series analysis 39 / 59

40 Realized YTM Simulated YTM Upper Bound Lower Bound An Illustration x GER10Y YTM Actual Predicted x 10 3 Forecasting 10Y YTM () Chapter 5 Univariate time-series analysis 40 / 59

41 Insert Clicker 6 here () Chapter 5 Univariate time-series analysis 41 / 59

42 Trends Forecasting using an ARMA models exploits two features of the data: mean-reversion and persistence. Unfortunately many financial time series do not feature mean reversion as they behave like non-stationary time series. Non-stationarity of time-series is a possible manifestation of a trend. Consider, for example, the random walk process with a drift:consider, for example, the random walk process with a drift: Recursive substitution yields x t = a 0 + x t 1 + ɛ t, ɛ t n.i.d. 0, σ 2 ɛ. t 1 x t = x 0 + a 0 t + ɛ t i, which shows that the non-stationary series contains both a deterministic (a 0 t) and a stochastic i=0 t 1 ɛ t i trend. () Chapter 5 Univariate time-series i=0 analysis 42 / 59

43 Integrated Series An easy way to make a non-stationary series stationary is differencing: x t = x t x t 1 = (1 L) x t = a 0 + ɛ t. In general, if a time-series needs to be differenced d times to become stationary, then it is integrated of order d or I(d). Our random walk is I(1). When the d-th difference of a time-series x, d x t, can be represented by an ARMA(p, q) model, we say that x t is an integrated moving-average process of order p, d, q and denote it as ARIMA(p, d, q). () Chapter 5 Univariate time-series analysis 43 / 59

44 Deterministic vs Stochastic Trends Compare the behaviour of an integrated process with that of a trend stationary process. Trend stationary processes feature only a deterministic trend: z t = α + βt + ɛ t. The z t process is non-stationary, but the non-stationarity is removed simply by regressing z t on the deterministic trend. Unlike this, for integrated processes like (??) the removal of the deterministic trend does not deliver a stationary time-series. Deterministic trends have no memory while integrated variables have an infinite one. Both integrated variable and deterministic trend exhibit systematic variations, but in the latter case the variation is predictable, whereas in the other one it is not. () Chapter 5 Univariate time-series analysis 44 / 59

45 Insert Clicker 7 here () Chapter 5 Univariate time-series analysis 45 / 59

46 Univariate decompositions of time-series Beveridge and Nelson (1981) provide an elegant way of decomposing a non-stationary time-series into a permanent and a temporary (cyclical) component by applying ARIMA methods. For any non-stationary time-series x t integrated of the first order, the Wold decomposition theorem could be applied to its first difference, to deliver the following representation: x t = µ + C (L) ɛ t, ɛ t n.i.d. 0, σ 2 ɛ, where C (L) is a polynomial of order q in the lag operator. Consider now the polynomial D (L), defined as: D (L) = C (L) C (1). () Chapter 5 Univariate time-series analysis 46 / 59

47 Univariate decompositions of time-series Given that C (1) is a constant, also D (L) will be of order q. Clearly, therefore, 1 is a root of D (L), and D (1) = 0, D (L) = C (L) (1 L), where C (L) is a polynomial of order q 1. By equating (??) to (??), we have: C (L) = C (L) (1 L) + C (1), () Chapter 5 Univariate time-series analysis 47 / 59

48 Univariate decompositions of time-series By integrating, we have: x t = µ + C (L) ɛ t + C (1) ɛ t. x t = C (L) ɛ t + µt + C (1) z t = C t + TR t, where z t is a process for which z t = ɛ t. C t is the cyclical component and TR t is the trend component made of a deterministic and a stochastic trend. Note that the trend component can be represented as: TR t = TR t 1 + µ + C (1) ɛ t. () Chapter 5 Univariate time-series analysis 48 / 59

49 decomposition of an IMA(1,1) process Consider the process: In this case: x t = ɛ t + θɛ t 1, 0 < θ < 1. C (L) = 1 + θl, C (1) = 1 + θ, C (L) = C (L) C (1) = 1 L θ. The Beveridge and Nelson decomposition gives the following result: x t = C t + TR t = θɛ t + (1 + θ) z t. () Chapter 5 Univariate time-series analysis 49 / 59

50 decomposition of an ARIMA(1,1) process Consider the process: x t = ρ x t 1 + ɛ t + θɛ t 1. Here: C (L) = 1 + θl 1 ρl, C (1) = 1 + θ 1 ρ, C (L) = C (L) C (1) 1 L = θ + ρ (1 ρ) (1 ρl), and the Beveridge and Nelson decomposition yields: x t = C t + TR t = θ + ρ (1 ρ) (1 ρl) ɛ t θ 1 ρ z t. () Chapter 5 Univariate time-series analysis 50 / 59

51 BN decomposition in practice The practical derivation of a Beveridge and Nelson decomposition for any ARIMA process is easily derived by applying a methodology suggested by Cuddington and Winters (1987). For any I(1) process, the stochastic trend can be represented as: TR t = TR t 1 + µ + C (1) ɛ t. The decomposition can then be applied in the following steps: 1 identify the appropriate ARIMA model and estimate ɛ t and all the parameters in µ and C (1); 2 given an initial value for TR 0, use (??) to generate the permanent component of the time-series; 3 generate the cyclical component as the difference between the observed value in each period and the permanent component. The above procedure gives the permanent component up to a constant. If the precision of this procedure is unsatisfactory, one can use further conditions to identify the decomposition more precisely. For example, one can impose the condition that the sample mean of () Chapter 5 Univariate time-series analysis 51 / 59

52 Nelson decomposition - The properties of the permanent and temporary components of an integrated time-series delivered by the Beveridge Nelson decomposition are worth some comments. the innovations in the permanent and the transitory components are perfectly negatively correlated; the trend component is more volatile than the actual time-series as the negative correlation between the permanent and the transitory components acts to smooth the original time-series. Note that in general the variance of innovations might have an economic interpretation and theory might suggest different patterns of correlations between innovations in the permanent and transitory components that do differ from a perfectly negative correlation. In general, different restrictions on the correlation between the trend and the cycle components lead to the identification of different stochastic trends for integrated time-series. () Chapter 5 Univariate time-series analysis 52 / 59

53 Nelson decomposition To see this point more explicitly we can compare the Beveridge Nelson trend with the trend extracted using an alternative technique which has been recently very successful in time-series analysis: the Hodrick Prescott filter. the Hodrick Prescott filter is derived by minimizing the following expression: T t=1 (x t TR t ) 2 + λ T 1 t=2 h (TR t+1 TR t ) 2 (TR t TR t 1 ) 2i. The penalty parameter λ controls the smoothness of the series, by controlling the ratio of the variance of the cyclical component and the variance of the series. The larger the λ, the smoother the TR t approaches a linear trend. In practical applications λ is set to 100 for annual data, 1600 for quarterly data and for monthly data. () Chapter 5 Univariate time-series analysis 53 / 59

54 Nelson decomposition Figure 8 reports the Beveridge Nelson trend and the Hodrick Prescott trend (with λ = 14400) for the data generated in the previous section. 40 BN trend HP trend IMA Trends () Chapter 5 Univariate time-series analysis 54 / 59

55 Insert Clicker 8 here () Chapter 5 Univariate time-series analysis 55 / 59

56 Asset Allocation with a simple TVER model: the SOP method Total stock market returns in local currency can be expressed as follows: 1 + H s t+1 P t+1 + D t+1 P t = P t+1 + D t+1 P t = P t+1 + D t+1 = P t+1 + D t+1 P t+1 P t P t P t P t+1 P t = 1 + D t+1 Pt+1 P t+1 P t taking logs: h s t+1 ' (p t+1 p t ) + D t+1 P t+1 () Chapter 5 Univariate time-series analysis 56 / 59

57 Asset Allocation with a simple TVER model In case of investment in a foreign currency we have a third term capturing exchange rate fluctuations. h s,l t+1 ' (p t+1 p t ) + D t+1 P t+1 + (e t+1 e t ) Asset allocation with the simplest TVER model can be applied by specifying univariate time series model to predict each part of the total return. The investor uses monthly data available over the period to find the tangency portfolio for an investment over the period The risky assets available for portfolio allocation are German, US and UK shares and the German 10-Year government bond. The risk free asset is the German short-term rate. () Chapter 5 Univariate time-series analysis 57 / 59

58 Asset Allocation with a simple TVER model German stock market returns: r GER t,t+36 = p GER t+36 p GER t E t p GER t+36 p GER t E t 36 j=1 36 D GER t+j j=1 P GER t+j D GER t+j P GER t+j! p GER t+36 p GER t = E t p GER t+36 p GER t + 36 i=j = β GER 0 + β GER 1 p GER t = E t 36 = 35 i=0 i=1 D GER t j P GER t j D GER t+j P GER t+j! D GER t+j P GER t+j + u 1,t u 2,t p GER t j j=1! () Chapter 5 Univariate time-series analysis 58 / 59

59 Asset Allocation with a simple TVER model US and UK stock market returns : p i t+36 r i t,t+36 = p i t E t p i t+36 p i t 36 D i t+i i=1 P i t+i p i t+36 p i t + = E t p i t+36 p i t = β i 0 + βi 1 p i t = 35 D i t j=0 P i t j j 36 D i t+j j=1 P i t+j + u i 1,t u i 2,t+36 β i 36 2 β i p i t 1 j=1 e i t+36 = e i t + u i 3,t+36, i = UK, US + e i t+36 e t The simple CER model is maintained for the German bond returns: j! () Chapter 5 Univariate time-series analysis 59 / 59