Cointegration: A First Look

Size: px
Start display at page:

Download "Cointegration: A First Look"

Transcription

1 Cointegration: A First Look Professor: Alan G. Isaac Last modified: 2006 Apr 01 Contents 1 Deterministic and Stochastic Trends GDP A Difficulty Error Correction 8 3 Cointegration ECMs VECMs An Alternative Representation Application Cointegration Deterministic and Stochastic Trends Time trend decomposition is a common practice. One traditional practice is to decompose output into two components: a non-stationary growth trend and a stationary cyclical 1

2 page 2 component. 1 There is no general agreement on the appropriate way to detrend data. 2 The traditional practice is close to Hooker s (1901) suggestion to focus on deviations from trend. For example, output is often regressed on (a polynomial in) time; residuals are interpreted as the cyclical component. Nelson and Plosser (JME,1982) argue that this is a misspecification, that secular movement in output is better modeled as the result of a random walk. Implications: fluctuations in output are largely due to permanent real shocks, and a large fraction of output fluctuations cannot be explained by money induced transitory fluctuations. Consider the simple example discussed by Nelson and Plosser (JME,1982). If y t = α + βt + u t (1) where u t is a stationary shock (the cyclical component). Then, although we may exploit any autocorrelation in u t for SR forecasting, our LR forecasts are just α + βt: current and past events don t affect our LR expectations. Such a process is called trend stationary (TS) since it is stationary once we remove a time trend. In contrast, if (1 L)y t = β + u t (2) then y t = y 0 + βt + t u j (3) so that current and past shocks are relevant even for our LR forecasts. Note that although just removing the trend doesn t render y t stationary, the series is stationary after differencing. Such data generating processes (DGPs) are called integrated or difference stationary (DS). 3 Note that in contrast to the bounded variance of the forecast error for a TS process, the 1 We will call a random variable stationary when its mean and auto-covariance are independent of time. 2 See the discussion in Stock and Watson, JEP A DGP with a spectrum that is finite but non-zero at all frequencies is said to be integrated of order zero or I(0). A stationary DGP is an example of an I(0) process. If a series needs to be differenced d times to become I(0), we say it is integrated of order d or I(d). Thus a difference stationary process is I(1). i=1

3 page 3 forecast error of a DS process increases without bound (since we must allow for a new permanent shock every period). In our simple example, if u t is white noise then the t period ahead forecast has forecast error variance σu 2 for the TS process but tσu 2 for the DS process. 1.1 GDP In the trend stationary case, we sometimes refer to β as the natural rate of growth of GNP. If output follows a TS process, then the natural rate of growth will give us a good prediction of future GNP. As a result we can refer to α + βt as natural real GNP. In contrast, if output follows a DS process, then predictions of future output that are based on the natural rate of growth deteriorate as the forecast horizon increases. There is no natural level of GNP. Advocates of the DS characterization of GNP, the real business cycle (RBC) theorists, emphasize supply side rather than demand side shocks to the level of output. Technology shocks tend to be emphasized as positive permanent shocks, but any effect on factor supplies should also be included (e.g., demographic shocks, changes in preferences, regulatory shocks, tax shocks). Comment: RBC theorist often motivate autocorrelation in u t (persistence of the shock) by including the time to build new capital and the consumption smoothing (permanent income) behavior of consumers. Comment: the RBC framework was motivated partly by the observation of procyclical real wages, but it is also compatible with countercyclical wage behavior (given, say, a labor supply shock.) 1.2 A Difficulty Problem: we can get the wrong econometric answers if we treat a DS process as if it were TS. 4 For example, Shiller (JPE,1979) rejects bond market efficiency under the TS assumption 4 The intuition for this turns on the convergence of the moments of an integrated variable to random variables, in contrast to stationary variables whose moments converge to constants that are the population counterparts. See Pagan and Wickens p and Also, see the example constructed by Stock and

4 page 4 for SR interest rates but not under the DS assumption. This suggests that we should not use the TS assumption without testing it. Consider a special case of our simple example: u t iid(0, σ 2 ). Both the TS and the DS case can be written y t = α + βt + u t /(1 φl) (4) where φ is zero or one. Multiplying by (1 φl), y t = [α(1 φ) + φβ] + φy t 1 + β(1 φ)t + u t (5) So we can run the regression y t = µ + ρy t 1 + γt + u t (6) and test the null hypothesis ρ = 1, γ = 0 in order to test the null of φ = 1, i.e., that y t is DS. Problem: the t-ratios are not t-distributed under the null: standard t-tests are biased toward rejecting the null. Solution: Fuller (1976) and Evans and Savin (Econometrica 1981,1984) provide tabulations of the t-ratio distribution under the null. Also see Dickey and Fuller (Econometrica 1981) for likelihood ratio statistics and for results on higher order AR representations. Comment: Nelson and Plosser (1982) suggest some natural but informal additional diagnostics based on sample autocorrelations. Comment: For the special case we have considered, it may be more natural to just examine the Durbin-Watson (DW) statistic from the regression of y on a constant (y t = c + u t ). Sargan and Bhargava (Econometrica,1983) tabulate the appropriate critical values for this test, which is invariant to the presence of a trend in the true model and is the uniformly most powerful invariant test of a simple random walk null against a stationary AR(1) alternative. However the Sargan and Bhargava DW test is only applicable in this special case. Watson J.Econ.Persp.,1988.

5 page 5 Comment: Caution! Inclusion of an intercept or trend, actual drift in the series, and autocorrelation in u t all of these changes affect the distribution of ˆρ. Comment: Autocorrelation in u t is most often handled by the augmented Dickey-Fuller (ADF) test: estimate by OLS y t = µ (1 ρ)y t 1 + p θ i y t i + e t (7) i=1 The original Dickey-Fuller tables tabulate results for the test statistic T (ˆρ 1). Unfortunately, procedures for determining the appropriate number of lags remain ad hoc. Note that the presence of a unit root in a series implies that, with probability one, economically absurd values for many real variables must eventually be realized. Further, as indicated by Perron (1989), unit root tests are sensitive to missing regressors: mean shifts do to missing regressors may be interpreted as unit roots. A big puzzle in this context is the blithe reliance on unit root tests where acceptance of the null hypothesis of the existence of a unit root is based on low Type I error. As emphasized by R.A. Fisher, failure to reject should not be equated with acceptance. (Nelson and Plosser do acknowledge this.) Current practice becomes all the more puzzling given the universal acknowledgement of the low power of the tests against the more believable autoregressive alternatives. The real reason for using a DS null is probably because it would natural to test a simple null that the process is TS by looking for a unit root in moving average component of the innovation to the differenced series. (E.g., from the simple example, estimate y t = β + u t φu t 1 and test the null of φ = 1.) Unfortunately, testing for a unit MA root is known to be problematic. 5 It may also be worth noting that current unit roots tests rule out fractional 5 Plosser and Schwert,1977,J.Econometrics; but see the exact MLE estimation suggested by Campbell and Mankiw, QJE 102(4), Nov.1987.

6 page 6 differencing a priori (see F. Sowell, Econometrica 58(2), March 1990, p ). Despite these many difficulties, while some economists such as Cochrane (JPE,1988) continue to dispute the Nelson and Plosser story, many economists have accepted it. Econometric considerations arise when regressors are DS. Suppose y t = x t β + z t γ + u t (8) where z t is stationary (and can include a constant) but x t is DS. If x t includes a drift (so that E x t 0) then standard t-ratios for β and γ should be checked against a standardized normal table. 6 However, if a trend is included in the regression then the Dickey-Fuller tables remain relevant for β. Related to this last point, if x t is without drift then the t-ratios for β lose their general usefulness: significance testing for integrated regressors is a problem. Further problems arise if there is more than one DS regressor with drift, unless they are detrended (see Park and Phillips,1988, in Econometric Theory vol 4). An exception to the problem: if it is possible to rewrite the regression equation so that the coefficients of interest are on mean zero stationary variables the F -statistic and t-statistic have the usual asymptotic distributions (Stock and West, JME,1988). Note that Stock and West claim that this result does not depend on being able to rewrite the entire regression in this fashion! Caution! This blessing is a curse in the presence of cointegration: then, in the absence of strict exogeneity, the stationary regressor will typically be correlated with the error term and the parameter estimate will be inconsistent. (Stock and Watson 1988,p.166). That is, simultaneous equations bias (etc.) is relevant once again see below. Granger and Newbold (J.Econometrics,1974) joined Yule (J.Roy.Stat.Soc., 1926) in warning that regressions involving trending or integrated variables can yield spurious (misleadingly strong) correlations, and they urged differencing as a propadeutic. (For example, 6 Roughly: see West, Econometrica 56(6),1988; of course this relies on the asymptotic distribution. West s results apply for univariate x t, i.e., a single DS regressor. He also shows (p.1408) that the results extend the the case of several DS variables if those variables cointegrate. The results also extend to the IV case.

7 page 7 regression of a DS variable on another independent DS variable yields parameter estimates that converge to random variables.)

8 page 8 2 Error Correction Suppose we relate two difference stationary variables as follows: y t = β x t + ε t (9) where ε t is white noise. Noting that y t = t y s + y 0 (10) i=1 is an identity, we can substitute for y t to get t y t = (β x t + ε s ) + y 0 i=1 = y 0 + β(x t x 0 ) + = (y 0 βx 0 ) + βx t + t i=1 ε s t i=1 ε s (11) There is no long-run equilibrium relationship between y and x. Davidson et al. and Hendry and Mizon (EJ,1978) worried about this implied loss of information about any LR steady state solution. Error correction models (ECMs) were introduced basically in order to recover this information. Suppose for example that we relate two difference stationary variables as follows: y t = β x t γ(y t 1 αx t 1 ) + ε t (12) The new term is called the error correction term. The basic ECM derives from classical control theory and has the form y t = β y t + γ(y t 1 y t 1 ) + δs t 1 + e t (13)

9 page 9 where y is the control variable, y t is the target, and S t 1 = δ (y t j y t j ) (14) Note that y = y in the long run steady state. In practice, S t 1 is generally dropped from the model and the target is expressed as a linear function of exogenous variables (yt = x t β). As Nickell (Oxford Bulletin of Economics and Statistics,1985) notes, the minimization of discounted costs of deviation from target and of adjustment introduces lagged values of y, where the number of lags depends on the data generating process for y. 3 Cointegration This sections draws on N. Mark section 2.6. Consider I(1) processes: q t and f t. We say these two series are cointegrated if they have a stationary linear combination. Otherwise they are not cointegrated. We will model q t and f t as a sum of nonstationary and stationary components. q t = ξ 1 t + z 1 t (15) f t = ξ 2 t + z 2 t (16) The components are two stationary contributions, z 1 t and z 2 t, and two random walks: ξ 1 t = ξ 1 t 1 + u 1 t (17) ξ 2 t = ξ 2 t 1 + u 2 t (18) (Here u 1 t and u 2 t are independent white noise processes.)

10 page 10 Now consider possible linear combinations of our two variables: q t βf t = (ξ 1 t βξ 2 t ) + (z 1 t βz 2 t ) (19) This can only be stationary if ξ 1 t βξ 2 t is stationary. That is, we need q t and f t to share a common stochastic trend. 3.1 ECMs Interest in ECMs has surged recently as a result of developments in the unit root literature. In particular, the Granger Representation Theorem (Granger, UCSC disc.paper 83-13,1983) established an equivalence ECMs and what is called cointegration: ECMs imply cointegration, and driftless cointegrated variables have an ECM representation. 7 Cointegration exists when a linear combination of DS variables is stationary. Recall our example y t = x t β + z t γ + u t where z t is stationary but x t is DS. In this case y t and x t are cointegrated, since (1 β)(y t x t ) is stationary. We call (1 β) the cointegrating vector. The good (and interesting) news is that since the variance in x t will dominate the variance in the stationary variables z t, it is acceptable to simply regress y t on x t in order to get a consistent estimate of β. In fact, the resulting estimate converges to β very fast a result known as the super-consistency of the cointegrating vector. 8 It may also be viewed as good news that this super-consistency is preserved even if you regress x t on y t : considerations of endogeneity and of errors in variables become relatively unimportant. (I believe this is a major source of the favor 7 More precisely, if the components of x t are I(1) and αx t is I(0), then there are finite order lag polynomials A(L) and d(l) such that A(L)(1 L)x t = γαx t + d(l)ɛ t 8 In a multiple regression, this super-consistency requires that no cointegrating relationship exists among the regressors. See Hendry (1986, Ox.Bul.Ec.Stat.,p.208).

11 page 11 cointegration receives in the profession: it supports single equation techniques by minimizing the importance of simultaneous equations bias. See Stock and Watson s JEP, 1988 p.168 comment on Haavelmo s problem.) The bad news is that, despite this super-consistency, small sample bias can be very large (see Banerjee et al., Oxford Bulletin of Economics and Statistics,1986). In order to get a feel for the link between cointegration and ECMs, recall our example y t = x t β + z t γ + u t Subtracting y t 1 from each side yields, with y t = x t β, y t = x t β + z t γ y t 1 + u t = x t β + z t γ (y t 1 x t 1 β) + u t = y t + (y t 1 y t 1 ) + z t γ + u t Granger and Engle (1987,p.263) provide a more interesting example. 9 The usefulness of this transformation is clear: the regression now involvles only I(0) variables and so our usual regression theory applies. Note that the close links between ECMs and cointegration has led some economists to mistakenly interpret a cointegrating vector as characterizing long run economic equilibrium. As noted by Hall and Henry (1988), The term equilibrium has many meanings in economics and its use in the cointegration literature is rather different from most definitions of equilibrium. Within the cointegration literature all that is meant by equilibrium is that it is an observed relationship which has, on average, been maintained by a set of variables for a long period. It implies none of the usual theoretical implications of market clearing or full 9 We can also write y t = x t β + z t γ (y t 1 x t 1 β z t 1 γ) + u t of course; but z t is already I(0) by assumption. Our example is close to that in Phillips 1988 Cowles Foundation Discussion Paper #893, as presented by Pagan and Wickens, except that we did not presume that yt should be dumped into the error term just because it is I(0).

12 page 12 employment and neither does it imply that the system is at rest. See Pagan and Wickens (EJ,1989,p.1003), Swamy (J.Econometrics,1989), or even Granger (Ox.Bul.Ec.Stat., 1986, p.216) for for discussion of this error. 3.2 VECMs This section draws on N. Mark ch.2 Let y t = (q t, f t ). Suppose we have the following DGP: y t = A 1 y t 1 + A 2 y t 2 + u t (20) Subtract y t 1 from both sides: y t = (A 1 I)y t 1 + A 2 y t 2 + u t (21) We can rewrite this as y t = (A 1 + A 2 I)y t 1 A 2 y t 1 + u t (22) or y t = Ry t 1 A 2 y t 1 + u t (23) This is just the vector analog of our Dickey-Fuller discussion. If y t I(1), then y t is stationary. One way to achive this is R = 0, by direct analogy to our Dickey-Fuller discussion. In this case y t follows a standard vector autoregression. But with two variables, we can also potentially find linear combinations of q t and f t that are stationary. For example, if q t βf t is stationary, then this becomes the error correction

13 page 13 term, and we can write y t = r [ ] 11 1 β y t 1 A 2 y t 1 + u t r 21 = r 11 βr 11 y t 1 A 2 y t 1 + u t r 21 βr 21 (24) Note that the matrix is singular. Letting ζ t = (1 R = r 11 βr 11 (25) r 21 βr 21 β)y t be the error correction term, we now have y t = r 11 ζ t 1 A 2 y t 1 + u t (26) r 21 Equation (26) is called the vector error correction representation of y t. Note that is looks like a VAR in first differences, except for the addtional error correction term. The presence of the error correction term suggests that a simple VAR in differences would be misspecified: it wd have an omitted variable. 3.3 An Alternative Representation Suppose the cointegrating vector is given by theory. Then we can form the error correction term without estimation. Multiply (26) by the cointegrating vector to get ζ t = (r 11 βr 21 )ζ t 1 (1 β)a 2 y t 1 + (1 β)u t (27)

14 page 14 which implies ζ t = (1 + r 11 βr 21 )ζ t 1 (1 β)a 2 y t 1 + (1 β)u t (28) This relationship between stationary variables looks like it could be included in the system (26). But since we produced this as a linear combination of the equations in (26), this clearly adds no new information to that system. But we can use it to replace either one of the equations in (26). For example, we can use f t = (q t ζ t )/β to substitute everywhere for f t in (26) and (28) to get a two equation system in q t and ζ t. ζ t ζ t 1 ζ t 2 q t = Â1 q t 1 + Â2 q t u t (29) 1 β Or, we can use q t = (βf t + ζ t ) to substitute everywhere for q t in (26) and (28) to get a two equation system in ζ t and f t. ζ t = Ã1 ζ t 1 + Ã2 ζ t 2 f t f t 1 f t β u t (30) 0 1 So, why would we bother to do this? It allows us to use the simpler VAR estimation framework, with all its standard results, rather than relying on the more complicated VECM estimation framework. And in particular, the forecasting algebra for the VAR representation is simple Application Nelson Mark presents a reestimation of the MacDonald and Taylor (1993) application of these observations to the monetary approach model. (This in turn derives from earlier work by? on the present value model.)

15 page 15 Recall our core implication of the monetary model: s deviates from m to the extent that there is expected depreciation. s t = m t + λ(s e t+1 s t ) (31) Mark argues as follows: s is nonstationary, but s e t+1 s t is stationary, therefore s and m must be cointegrated. s t m t = λ(s e t+1 s t ) (32) What is more, the cointegrating vector is known to be (1 1). Defining ζ t = s t m t, our earlier work suggests the VAR m t = ζ t maxlag j=1 A j m t 1 + v t (33) ζ t 1 This can be estimated as a VAR, and the cross equation restrictions implied by the rational expectations hypothesis can be tested. Mark estimates the model using US and German data from to , and he uses a Wald test of the cross equation restrictions. Note that he imposes values for λ: he does not estimate these! But for robustness he tries several different values. His conclusion: the restrictions are rejected for reasonable values of λ. 3.4 Cointegration Testing for cointegration just involves testing that the residuals from a cointegrating regression do not contain a unit root. Thus when cointegration tests are applied to a pre-specified model, they are just another specification test, testing the econometric model s assumption of a stationary error term. 10 One approach is to just run the DF test on the residuals, but 10 In this light, it is easy to understand Pagan and Wickens suggestion (p.1009) that there seems little to be gained in running the cointegrating regression rather than running the entire prespecified regression from the beginning. (The stationary variables coefficients will be consistently estimated and their test

16 page 16 Figure 1: Theoretical and Actual Spreads

17 page 17 the bias in the estimate of the cointegrating vector leads too often to the rejection of the null of no cointegration. The most popular test is the cointegrating regression Durbin-Watson test (CRDW), which relies on the Durbin-Watson statistic, but only the bounds of the distribution are available (as opposed to its use in testing individual series for integration). Engle and Granger (Econometrica,1987) suggest that a better procedure is to apply the ADF test to the residuals of the cointegrating regression. Engle and Yoo (J.Econometrics,1987) suggest changes in the critical values of the ADF in order to account for the dependence of these residuals on estimated parameters. These unit root tests of course retain the weaknesses noted above, making it perhaps surprising that so many researchers find cointegration. It seems to me, however, that this is an easily explained artifact of the autocorrelation in the time series data. High autocorrelation makes it difficult to reject the null that a series contains a unit root, but the same autocorrelation in multiple series makes it relatively easy to get lower autocorrelation in the residuals of a cointegrating regression. References Chow, Gregory C. (1987). Money and Price Level Determination in China. Journal of Comparative Economics 11, Fuller, Wayne A. (1976). Introduction to Statistical Time Series (1st ed.). Wiley-Interscience. MacDonald, Ronald and Mark P. Taylor (1993). The Monetary Approach to the Exchange Rate: Rational Expectations, Long-Run Equilibrium, and Forecasting. International Monetary Fund Staff Papers 40, Nelson, Charles R. and Charles I. Plosser (1982). Trends and Random Walks in Macroecostatistics will have the usual large sample properties: Stock and Watson,JEP 1988,p.167.) On the other hand, cointegrating regressions have been viewed by some as an aid to specification search, preliminary to the formulation of an ECM. Since cointegrating vectors are not unique in fact, both sign and magnitude are completely up for grabs this approach appears to have obvious dangers.

18 page 18 nomic Time Series: Some Evidence and Implications. Journal of Monetary Economics 10, Engle, Robert F. and Granger, C.W.J., 1987, Cointegration and Error Correction: Representation, Estimation, and Testing, Econometrica 55(2), Hooker, R.H., 1901, Correlation of the Marriage Rate with Trade, Journal of the Royal Statistical Society 64, p , cited by Hendry, 1986, Ox.Bul.Ec.Stat. 48(3). Perron, Pierre, 1989, The Great Crash, the Oil Price Shock and the Unit Root Hypothesis, Econometrica 57, Stock, James H., and Watson, Mark W., Variable Trends in Economic Time Series, JEP 2(3), Summer 1988,

10) Time series econometrics

10) Time series econometrics 30C00200 Econometrics 10) Time series econometrics Timo Kuosmanen Professor, Ph.D. 1 Topics today Static vs. dynamic time series model Suprious regression Stationary and nonstationary time series Unit

More information

Cointegration, Stationarity and Error Correction Models.

Cointegration, Stationarity and Error Correction Models. Cointegration, Stationarity and Error Correction Models. STATIONARITY Wold s decomposition theorem states that a stationary time series process with no deterministic components has an infinite moving average

More information

E 4101/5101 Lecture 9: Non-stationarity

E 4101/5101 Lecture 9: Non-stationarity E 4101/5101 Lecture 9: Non-stationarity Ragnar Nymoen 30 March 2011 Introduction I Main references: Hamilton Ch 15,16 and 17. Davidson and MacKinnon Ch 14.3 and 14.4 Also read Ch 2.4 and Ch 2.5 in Davidson

More information

ECON 4160, Spring term Lecture 12

ECON 4160, Spring term Lecture 12 ECON 4160, Spring term 2013. Lecture 12 Non-stationarity and co-integration 2/2 Ragnar Nymoen Department of Economics 13 Nov 2013 1 / 53 Introduction I So far we have considered: Stationary VAR, with deterministic

More information

7. Integrated Processes

7. Integrated Processes 7. Integrated Processes Up to now: Analysis of stationary processes (stationary ARMA(p, q) processes) Problem: Many economic time series exhibit non-stationary patterns over time 226 Example: We consider

More information

7. Integrated Processes

7. Integrated Processes 7. Integrated Processes Up to now: Analysis of stationary processes (stationary ARMA(p, q) processes) Problem: Many economic time series exhibit non-stationary patterns over time 226 Example: We consider

More information

9) Time series econometrics

9) Time series econometrics 30C00200 Econometrics 9) Time series econometrics Timo Kuosmanen Professor Management Science http://nomepre.net/index.php/timokuosmanen 1 Macroeconomic data: GDP Inflation rate Examples of time series

More information

Lecture 8a: Spurious Regression

Lecture 8a: Spurious Regression Lecture 8a: Spurious Regression 1 Old Stuff The traditional statistical theory holds when we run regression using (weakly or covariance) stationary variables. For example, when we regress one stationary

More information

Lecture 8a: Spurious Regression

Lecture 8a: Spurious Regression Lecture 8a: Spurious Regression 1 2 Old Stuff The traditional statistical theory holds when we run regression using stationary variables. For example, when we regress one stationary series onto another

More information

E 4160 Autumn term Lecture 9: Deterministic trends vs integrated series; Spurious regression; Dickey-Fuller distribution and test

E 4160 Autumn term Lecture 9: Deterministic trends vs integrated series; Spurious regression; Dickey-Fuller distribution and test E 4160 Autumn term 2016. Lecture 9: Deterministic trends vs integrated series; Spurious regression; Dickey-Fuller distribution and test Ragnar Nymoen Department of Economics, University of Oslo 24 October

More information

1 Regression with Time Series Variables

1 Regression with Time Series Variables 1 Regression with Time Series Variables With time series regression, Y might not only depend on X, but also lags of Y and lags of X Autoregressive Distributed lag (or ADL(p; q)) model has these features:

More information

Econometrics. Week 11. Fall Institute of Economic Studies Faculty of Social Sciences Charles University in Prague

Econometrics. Week 11. Fall Institute of Economic Studies Faculty of Social Sciences Charles University in Prague Econometrics Week 11 Institute of Economic Studies Faculty of Social Sciences Charles University in Prague Fall 2012 1 / 30 Recommended Reading For the today Advanced Time Series Topics Selected topics

More information

11/18/2008. So run regression in first differences to examine association. 18 November November November 2008

11/18/2008. So run regression in first differences to examine association. 18 November November November 2008 Time Series Econometrics 7 Vijayamohanan Pillai N Unit Root Tests Vijayamohan: CDS M Phil: Time Series 7 1 Vijayamohan: CDS M Phil: Time Series 7 2 R 2 > DW Spurious/Nonsense Regression. Integrated but

More information

BCT Lecture 3. Lukas Vacha.

BCT Lecture 3. Lukas Vacha. BCT Lecture 3 Lukas Vacha vachal@utia.cas.cz Stationarity and Unit Root Testing Why do we need to test for Non-Stationarity? The stationarity or otherwise of a series can strongly influence its behaviour

More information

Time series: Cointegration

Time series: Cointegration Time series: Cointegration May 29, 2018 1 Unit Roots and Integration Univariate time series unit roots, trends, and stationarity Have so far glossed over the question of stationarity, except for my stating

More information

CHAPTER 21: TIME SERIES ECONOMETRICS: SOME BASIC CONCEPTS

CHAPTER 21: TIME SERIES ECONOMETRICS: SOME BASIC CONCEPTS CHAPTER 21: TIME SERIES ECONOMETRICS: SOME BASIC CONCEPTS 21.1 A stochastic process is said to be weakly stationary if its mean and variance are constant over time and if the value of the covariance between

More information

Darmstadt Discussion Papers in Economics

Darmstadt Discussion Papers in Economics Darmstadt Discussion Papers in Economics The Effect of Linear Time Trends on Cointegration Testing in Single Equations Uwe Hassler Nr. 111 Arbeitspapiere des Instituts für Volkswirtschaftslehre Technische

More information

It is easily seen that in general a linear combination of y t and x t is I(1). However, in particular cases, it can be I(0), i.e. stationary.

It is easily seen that in general a linear combination of y t and x t is I(1). However, in particular cases, it can be I(0), i.e. stationary. 6. COINTEGRATION 1 1 Cointegration 1.1 Definitions I(1) variables. z t = (y t x t ) is I(1) (integrated of order 1) if it is not stationary but its first difference z t is stationary. It is easily seen

More information

ECON 4160, Lecture 11 and 12

ECON 4160, Lecture 11 and 12 ECON 4160, 2016. Lecture 11 and 12 Co-integration Ragnar Nymoen Department of Economics 9 November 2017 1 / 43 Introduction I So far we have considered: Stationary VAR ( no unit roots ) Standard inference

More information

G. S. Maddala Kajal Lahiri. WILEY A John Wiley and Sons, Ltd., Publication

G. S. Maddala Kajal Lahiri. WILEY A John Wiley and Sons, Ltd., Publication G. S. Maddala Kajal Lahiri WILEY A John Wiley and Sons, Ltd., Publication TEMT Foreword Preface to the Fourth Edition xvii xix Part I Introduction and the Linear Regression Model 1 CHAPTER 1 What is Econometrics?

More information

Nonstationary Time Series:

Nonstationary Time Series: Nonstationary Time Series: Unit Roots Egon Zakrajšek Division of Monetary Affairs Federal Reserve Board Summer School in Financial Mathematics Faculty of Mathematics & Physics University of Ljubljana September

More information

Introduction to Algorithmic Trading Strategies Lecture 3

Introduction to Algorithmic Trading Strategies Lecture 3 Introduction to Algorithmic Trading Strategies Lecture 3 Pairs Trading by Cointegration Haksun Li haksun.li@numericalmethod.com www.numericalmethod.com Outline Distance method Cointegration Stationarity

More information

Topic 4 Unit Roots. Gerald P. Dwyer. February Clemson University

Topic 4 Unit Roots. Gerald P. Dwyer. February Clemson University Topic 4 Unit Roots Gerald P. Dwyer Clemson University February 2016 Outline 1 Unit Roots Introduction Trend and Difference Stationary Autocorrelations of Series That Have Deterministic or Stochastic Trends

More information

Inflation Revisited: New Evidence from Modified Unit Root Tests

Inflation Revisited: New Evidence from Modified Unit Root Tests 1 Inflation Revisited: New Evidence from Modified Unit Root Tests Walter Enders and Yu Liu * University of Alabama in Tuscaloosa and University of Texas at El Paso Abstract: We propose a simple modification

More information

Financial Time Series Analysis: Part II

Financial Time Series Analysis: Part II Department of Mathematics and Statistics, University of Vaasa, Finland Spring 2017 1 Unit root Deterministic trend Stochastic trend Testing for unit root ADF-test (Augmented Dickey-Fuller test) Testing

More information

ARDL Cointegration Tests for Beginner

ARDL Cointegration Tests for Beginner ARDL Cointegration Tests for Beginner Tuck Cheong TANG Department of Economics, Faculty of Economics & Administration University of Malaya Email: tangtuckcheong@um.edu.my DURATION: 3 HOURS On completing

More information

On the robustness of cointegration tests when series are fractionally integrated

On the robustness of cointegration tests when series are fractionally integrated On the robustness of cointegration tests when series are fractionally integrated JESUS GONZALO 1 &TAE-HWYLEE 2, 1 Universidad Carlos III de Madrid, Spain and 2 University of California, Riverside, USA

More information

Non-Stationary Time Series, Cointegration, and Spurious Regression

Non-Stationary Time Series, Cointegration, and Spurious Regression Econometrics II Non-Stationary Time Series, Cointegration, and Spurious Regression Econometrics II Course Outline: Non-Stationary Time Series, Cointegration and Spurious Regression 1 Regression with Non-Stationarity

More information

Advanced Econometrics

Advanced Econometrics Based on the textbook by Verbeek: A Guide to Modern Econometrics Robert M. Kunst robert.kunst@univie.ac.at University of Vienna and Institute for Advanced Studies Vienna May 2, 2013 Outline Univariate

More information

Time Series Methods. Sanjaya Desilva

Time Series Methods. Sanjaya Desilva Time Series Methods Sanjaya Desilva 1 Dynamic Models In estimating time series models, sometimes we need to explicitly model the temporal relationships between variables, i.e. does X affect Y in the same

More information

Testing for non-stationarity

Testing for non-stationarity 20 November, 2009 Overview The tests for investigating the non-stationary of a time series falls into four types: 1 Check the null that there is a unit root against stationarity. Within these, there are

More information

A TIME SERIES PARADOX: UNIT ROOT TESTS PERFORM POORLY WHEN DATA ARE COINTEGRATED

A TIME SERIES PARADOX: UNIT ROOT TESTS PERFORM POORLY WHEN DATA ARE COINTEGRATED A TIME SERIES PARADOX: UNIT ROOT TESTS PERFORM POORLY WHEN DATA ARE COINTEGRATED by W. Robert Reed Department of Economics and Finance University of Canterbury, New Zealand Email: bob.reed@canterbury.ac.nz

More information

Economics 308: Econometrics Professor Moody

Economics 308: Econometrics Professor Moody Economics 308: Econometrics Professor Moody References on reserve: Text Moody, Basic Econometrics with Stata (BES) Pindyck and Rubinfeld, Econometric Models and Economic Forecasts (PR) Wooldridge, Jeffrey

More information

Vector error correction model, VECM Cointegrated VAR

Vector error correction model, VECM Cointegrated VAR 1 / 58 Vector error correction model, VECM Cointegrated VAR Chapter 4 Financial Econometrics Michael Hauser WS17/18 2 / 58 Content Motivation: plausible economic relations Model with I(1) variables: spurious

More information

Regression with time series

Regression with time series Regression with time series Class Notes Manuel Arellano February 22, 2018 1 Classical regression model with time series Model and assumptions The basic assumption is E y t x 1,, x T = E y t x t = x tβ

More information

Econ 423 Lecture Notes: Additional Topics in Time Series 1

Econ 423 Lecture Notes: Additional Topics in Time Series 1 Econ 423 Lecture Notes: Additional Topics in Time Series 1 John C. Chao April 25, 2017 1 These notes are based in large part on Chapter 16 of Stock and Watson (2011). They are for instructional purposes

More information

Questions and Answers on Unit Roots, Cointegration, VARs and VECMs

Questions and Answers on Unit Roots, Cointegration, VARs and VECMs Questions and Answers on Unit Roots, Cointegration, VARs and VECMs L. Magee Winter, 2012 1. Let ɛ t, t = 1,..., T be a series of independent draws from a N[0,1] distribution. Let w t, t = 1,..., T, be

More information

Lecture 6a: Unit Root and ARIMA Models

Lecture 6a: Unit Root and ARIMA Models Lecture 6a: Unit Root and ARIMA Models 1 2 Big Picture A time series is non-stationary if it contains a unit root unit root nonstationary The reverse is not true. For example, y t = cos(t) + u t has no

More information

Nonsense Regressions due to Neglected Time-varying Means

Nonsense Regressions due to Neglected Time-varying Means Nonsense Regressions due to Neglected Time-varying Means Uwe Hassler Free University of Berlin Institute of Statistics and Econometrics Boltzmannstr. 20 D-14195 Berlin Germany email: uwe@wiwiss.fu-berlin.de

More information

Lecture 5: Unit Roots, Cointegration and Error Correction Models The Spurious Regression Problem

Lecture 5: Unit Roots, Cointegration and Error Correction Models The Spurious Regression Problem Lecture 5: Unit Roots, Cointegration and Error Correction Models The Spurious Regression Problem Prof. Massimo Guidolin 20192 Financial Econometrics Winter/Spring 2018 Overview Stochastic vs. deterministic

More information

Cointegration. Example 1 Consider the following model: x t + βy t = u t (1) x t + αy t = e t (2) u t = u t 1 + ε 1t (3)

Cointegration. Example 1 Consider the following model: x t + βy t = u t (1) x t + αy t = e t (2) u t = u t 1 + ε 1t (3) Cointegration In economics we usually think that there exist long-run relationships between many variables of interest. For example, although consumption and income may each follow random walks, it seem

More information

Covers Chapter 10-12, some of 16, some of 18 in Wooldridge. Regression Analysis with Time Series Data

Covers Chapter 10-12, some of 16, some of 18 in Wooldridge. Regression Analysis with Time Series Data Covers Chapter 10-12, some of 16, some of 18 in Wooldridge Regression Analysis with Time Series Data Obviously time series data different from cross section in terms of source of variation in x and y temporal

More information

Non-Stationary Time Series and Unit Root Testing

Non-Stationary Time Series and Unit Root Testing Econometrics II Non-Stationary Time Series and Unit Root Testing Morten Nyboe Tabor Course Outline: Non-Stationary Time Series and Unit Root Testing 1 Stationarity and Deviation from Stationarity Trend-Stationarity

More information

Economic modelling and forecasting. 2-6 February 2015

Economic modelling and forecasting. 2-6 February 2015 Economic modelling and forecasting 2-6 February 2015 Bank of England 2015 Ole Rummel Adviser, CCBS at the Bank of England ole.rummel@bankofengland.co.uk Philosophy of my presentations Everything should

More information

Econometrics of Panel Data

Econometrics of Panel Data Econometrics of Panel Data Jakub Mućk Meeting # 6 Jakub Mućk Econometrics of Panel Data Meeting # 6 1 / 36 Outline 1 The First-Difference (FD) estimator 2 Dynamic panel data models 3 The Anderson and Hsiao

More information

Empirical Market Microstructure Analysis (EMMA)

Empirical Market Microstructure Analysis (EMMA) Empirical Market Microstructure Analysis (EMMA) Lecture 3: Statistical Building Blocks and Econometric Basics Prof. Dr. Michael Stein michael.stein@vwl.uni-freiburg.de Albert-Ludwigs-University of Freiburg

More information

Moreover, the second term is derived from: 1 T ) 2 1

Moreover, the second term is derived from: 1 T ) 2 1 170 Moreover, the second term is derived from: 1 T T ɛt 2 σ 2 ɛ. Therefore, 1 σ 2 ɛt T y t 1 ɛ t = 1 2 ( yt σ T ) 2 1 2σ 2 ɛ 1 T T ɛt 2 1 2 (χ2 (1) 1). (b) Next, consider y 2 t 1. T E y 2 t 1 T T = E(y

More information

ECONOMETRICS II, FALL Testing for Unit Roots.

ECONOMETRICS II, FALL Testing for Unit Roots. ECONOMETRICS II, FALL 216 Testing for Unit Roots. In the statistical literature it has long been known that unit root processes behave differently from stable processes. For example in the scalar AR(1)

More information

Non-Stationary Time Series and Unit Root Testing

Non-Stationary Time Series and Unit Root Testing Econometrics II Non-Stationary Time Series and Unit Root Testing Morten Nyboe Tabor Course Outline: Non-Stationary Time Series and Unit Root Testing 1 Stationarity and Deviation from Stationarity Trend-Stationarity

More information

Prof. Dr. Roland Füss Lecture Series in Applied Econometrics Summer Term Introduction to Time Series Analysis

Prof. Dr. Roland Füss Lecture Series in Applied Econometrics Summer Term Introduction to Time Series Analysis Introduction to Time Series Analysis 1 Contents: I. Basics of Time Series Analysis... 4 I.1 Stationarity... 5 I.2 Autocorrelation Function... 9 I.3 Partial Autocorrelation Function (PACF)... 14 I.4 Transformation

More information

Introduction to Eco n o m et rics

Introduction to Eco n o m et rics 2008 AGI-Information Management Consultants May be used for personal purporses only or by libraries associated to dandelon.com network. Introduction to Eco n o m et rics Third Edition G.S. Maddala Formerly

More information

Non-Stationary Time Series and Unit Root Testing

Non-Stationary Time Series and Unit Root Testing Econometrics II Non-Stationary Time Series and Unit Root Testing Morten Nyboe Tabor Course Outline: Non-Stationary Time Series and Unit Root Testing 1 Stationarity and Deviation from Stationarity Trend-Stationarity

More information

Stationarity and Cointegration analysis. Tinashe Bvirindi

Stationarity and Cointegration analysis. Tinashe Bvirindi Stationarity and Cointegration analysis By Tinashe Bvirindi tbvirindi@gmail.com layout Unit root testing Cointegration Vector Auto-regressions Cointegration in Multivariate systems Introduction Stationarity

More information

This chapter reviews properties of regression estimators and test statistics based on

This chapter reviews properties of regression estimators and test statistics based on Chapter 12 COINTEGRATING AND SPURIOUS REGRESSIONS This chapter reviews properties of regression estimators and test statistics based on the estimators when the regressors and regressant are difference

More information

Cointegration and Tests of Purchasing Parity Anthony Mac Guinness- Senior Sophister

Cointegration and Tests of Purchasing Parity Anthony Mac Guinness- Senior Sophister Cointegration and Tests of Purchasing Parity Anthony Mac Guinness- Senior Sophister Most of us know Purchasing Power Parity as a sensible way of expressing per capita GNP; that is taking local price levels

More information

Christopher Dougherty London School of Economics and Political Science

Christopher Dougherty London School of Economics and Political Science Introduction to Econometrics FIFTH EDITION Christopher Dougherty London School of Economics and Political Science OXFORD UNIVERSITY PRESS Contents INTRODU CTION 1 Why study econometrics? 1 Aim of this

More information

Title. Description. var intro Introduction to vector autoregressive models

Title. Description. var intro Introduction to vector autoregressive models Title var intro Introduction to vector autoregressive models Description Stata has a suite of commands for fitting, forecasting, interpreting, and performing inference on vector autoregressive (VAR) models

More information

Unit Root and Cointegration

Unit Root and Cointegration Unit Root and Cointegration Carlos Hurtado Department of Economics University of Illinois at Urbana-Champaign hrtdmrt@illinois.edu Oct 7th, 016 C. Hurtado (UIUC - Economics) Applied Econometrics On the

More information

Lecture 6: Dynamic Models

Lecture 6: Dynamic Models Lecture 6: Dynamic Models R.G. Pierse 1 Introduction Up until now we have maintained the assumption that X values are fixed in repeated sampling (A4) In this lecture we look at dynamic models, where the

More information

Trending Models in the Data

Trending Models in the Data April 13, 2009 Spurious regression I Before we proceed to test for unit root and trend-stationary models, we will examine the phenomena of spurious regression. The material in this lecture can be found

More information

Chapter 2: Unit Roots

Chapter 2: Unit Roots Chapter 2: Unit Roots 1 Contents: Lehrstuhl für Department Empirische of Wirtschaftsforschung Empirical Research and undeconometrics II. Unit Roots... 3 II.1 Integration Level... 3 II.2 Nonstationarity

More information

This is a repository copy of The Error Correction Model as a Test for Cointegration.

This is a repository copy of The Error Correction Model as a Test for Cointegration. This is a repository copy of The Error Correction Model as a Test for Cointegration. White Rose Research Online URL for this paper: http://eprints.whiterose.ac.uk/9886/ Monograph: Kanioura, A. and Turner,

More information

Stationarity and cointegration tests: Comparison of Engle - Granger and Johansen methodologies

Stationarity and cointegration tests: Comparison of Engle - Granger and Johansen methodologies MPRA Munich Personal RePEc Archive Stationarity and cointegration tests: Comparison of Engle - Granger and Johansen methodologies Faik Bilgili Erciyes University, Faculty of Economics and Administrative

More information

EC408 Topics in Applied Econometrics. B Fingleton, Dept of Economics, Strathclyde University

EC408 Topics in Applied Econometrics. B Fingleton, Dept of Economics, Strathclyde University EC408 Topics in Applied Econometrics B Fingleton, Dept of Economics, Strathclyde University Applied Econometrics What is spurious regression? How do we check for stochastic trends? Cointegration and Error

More information

Defence Spending and Economic Growth: Re-examining the Issue of Causality for Pakistan and India

Defence Spending and Economic Growth: Re-examining the Issue of Causality for Pakistan and India The Pakistan Development Review 34 : 4 Part III (Winter 1995) pp. 1109 1117 Defence Spending and Economic Growth: Re-examining the Issue of Causality for Pakistan and India RIZWAN TAHIR 1. INTRODUCTION

More information

ECON3327: Financial Econometrics, Spring 2016

ECON3327: Financial Econometrics, Spring 2016 ECON3327: Financial Econometrics, Spring 2016 Wooldridge, Introductory Econometrics (5th ed, 2012) Chapter 11: OLS with time series data Stationary and weakly dependent time series The notion of a stationary

More information

Nonstationary time series models

Nonstationary time series models 13 November, 2009 Goals Trends in economic data. Alternative models of time series trends: deterministic trend, and stochastic trend. Comparison of deterministic and stochastic trend models The statistical

More information

11. Further Issues in Using OLS with TS Data

11. Further Issues in Using OLS with TS Data 11. Further Issues in Using OLS with TS Data With TS, including lags of the dependent variable often allow us to fit much better the variation in y Exact distribution theory is rarely available in TS applications,

More information

Lecture 7: Dynamic panel models 2

Lecture 7: Dynamic panel models 2 Lecture 7: Dynamic panel models 2 Ragnar Nymoen Department of Economics, UiO 25 February 2010 Main issues and references The Arellano and Bond method for GMM estimation of dynamic panel data models A stepwise

More information

Trends and Unit Roots in Greek Real Money Supply, Real GDP and Nominal Interest Rate

Trends and Unit Roots in Greek Real Money Supply, Real GDP and Nominal Interest Rate European Research Studies Volume V, Issue (3-4), 00, pp. 5-43 Trends and Unit Roots in Greek Real Money Supply, Real GDP and Nominal Interest Rate Karpetis Christos & Varelas Erotokritos * Abstract This

More information

THE IMPACT OF REAL EXCHANGE RATE CHANGES ON SOUTH AFRICAN AGRICULTURAL EXPORTS: AN ERROR CORRECTION MODEL APPROACH

THE IMPACT OF REAL EXCHANGE RATE CHANGES ON SOUTH AFRICAN AGRICULTURAL EXPORTS: AN ERROR CORRECTION MODEL APPROACH THE IMPACT OF REAL EXCHANGE RATE CHANGES ON SOUTH AFRICAN AGRICULTURAL EXPORTS: AN ERROR CORRECTION MODEL APPROACH D. Poonyth and J. van Zyl 1 This study evaluates the long run and short run effects of

More information

Multivariate Time Series

Multivariate Time Series Multivariate Time Series Fall 2008 Environmental Econometrics (GR03) TSII Fall 2008 1 / 16 More on AR(1) In AR(1) model (Y t = µ + ρy t 1 + u t ) with ρ = 1, the series is said to have a unit root or a

More information

Time Series Econometrics 4 Vijayamohanan Pillai N

Time Series Econometrics 4 Vijayamohanan Pillai N Time Series Econometrics 4 Vijayamohanan Pillai N Vijayamohan: CDS MPhil: Time Series 5 1 Autoregressive Moving Average Process: ARMA(p, q) Vijayamohan: CDS MPhil: Time Series 5 2 1 Autoregressive Moving

More information

MA Advanced Econometrics: Spurious Regressions and Cointegration

MA Advanced Econometrics: Spurious Regressions and Cointegration MA Advanced Econometrics: Spurious Regressions and Cointegration Karl Whelan School of Economics, UCD February 22, 2011 Karl Whelan (UCD) Spurious Regressions and Cointegration February 22, 2011 1 / 18

More information

Department of Economics, UCSD UC San Diego

Department of Economics, UCSD UC San Diego Department of Economics, UCSD UC San Diego itle: Spurious Regressions with Stationary Series Author: Granger, Clive W.J., University of California, San Diego Hyung, Namwon, University of Seoul Jeon, Yongil,

More information

Econometría 2: Análisis de series de Tiempo

Econometría 2: Análisis de series de Tiempo Econometría 2: Análisis de series de Tiempo Karoll GOMEZ kgomezp@unal.edu.co http://karollgomez.wordpress.com Segundo semestre 2016 IX. Vector Time Series Models VARMA Models A. 1. Motivation: The vector

More information

Financial Econometrics

Financial Econometrics Financial Econometrics Long-run Relationships in Finance Gerald P. Dwyer Trinity College, Dublin January 2016 Outline 1 Long-Run Relationships Review of Nonstationarity in Mean Cointegration Vector Error

More information

Identifying the Monetary Policy Shock Christiano et al. (1999)

Identifying the Monetary Policy Shock Christiano et al. (1999) Identifying the Monetary Policy Shock Christiano et al. (1999) The question we are asking is: What are the consequences of a monetary policy shock a shock which is purely related to monetary conditions

More information

Time Series Analysis. James D. Hamilton PRINCETON UNIVERSITY PRESS PRINCETON, NEW JERSEY

Time Series Analysis. James D. Hamilton PRINCETON UNIVERSITY PRESS PRINCETON, NEW JERSEY Time Series Analysis James D. Hamilton PRINCETON UNIVERSITY PRESS PRINCETON, NEW JERSEY & Contents PREFACE xiii 1 1.1. 1.2. Difference Equations First-Order Difference Equations 1 /?th-order Difference

More information

Lecture 2: Univariate Time Series

Lecture 2: Univariate Time Series Lecture 2: Univariate Time Series Analysis: Conditional and Unconditional Densities, Stationarity, ARMA Processes Prof. Massimo Guidolin 20192 Financial Econometrics Spring/Winter 2017 Overview Motivation:

More information

A nonparametric test for seasonal unit roots

A nonparametric test for seasonal unit roots Robert M. Kunst robert.kunst@univie.ac.at University of Vienna and Institute for Advanced Studies Vienna To be presented in Innsbruck November 7, 2007 Abstract We consider a nonparametric test for the

More information

Stochastic Trends & Economic Fluctuations

Stochastic Trends & Economic Fluctuations Stochastic Trends & Economic Fluctuations King, Plosser, Stock & Watson (AER, 1991) Cesar E. Tamayo Econ508 - Economics - Rutgers November 14, 2011 Cesar E. Tamayo Stochastic Trends & Economic Fluctuations

More information

13. Time Series Analysis: Asymptotics Weakly Dependent and Random Walk Process. Strict Exogeneity

13. Time Series Analysis: Asymptotics Weakly Dependent and Random Walk Process. Strict Exogeneity Outline: Further Issues in Using OLS with Time Series Data 13. Time Series Analysis: Asymptotics Weakly Dependent and Random Walk Process I. Stationary and Weakly Dependent Time Series III. Highly Persistent

More information

Multivariate Time Series: Part 4

Multivariate Time Series: Part 4 Multivariate Time Series: Part 4 Cointegration Gerald P. Dwyer Clemson University March 2016 Outline 1 Multivariate Time Series: Part 4 Cointegration Engle-Granger Test for Cointegration Johansen Test

More information

Introductory Workshop on Time Series Analysis. Sara McLaughlin Mitchell Department of Political Science University of Iowa

Introductory Workshop on Time Series Analysis. Sara McLaughlin Mitchell Department of Political Science University of Iowa Introductory Workshop on Time Series Analysis Sara McLaughlin Mitchell Department of Political Science University of Iowa Overview Properties of time series data Approaches to time series analysis Stationarity

More information

When Do Wold Orderings and Long-Run Recursive Identifying Restrictions Yield Identical Results?

When Do Wold Orderings and Long-Run Recursive Identifying Restrictions Yield Identical Results? Preliminary and incomplete When Do Wold Orderings and Long-Run Recursive Identifying Restrictions Yield Identical Results? John W Keating * University of Kansas Department of Economics 334 Snow Hall Lawrence,

More information

Consider the trend-cycle decomposition of a time series y t

Consider the trend-cycle decomposition of a time series y t 1 Unit Root Tests Consider the trend-cycle decomposition of a time series y t y t = TD t + TS t + C t = TD t + Z t The basic issue in unit root testing is to determine if TS t = 0. Two classes of tests,

More information

CHAPTER III RESEARCH METHODOLOGY. trade balance performance of selected ASEAN-5 countries and exchange rate

CHAPTER III RESEARCH METHODOLOGY. trade balance performance of selected ASEAN-5 countries and exchange rate CHAPTER III RESEARCH METHODOLOGY 3.1 Research s Object The research object is taking the macroeconomic perspective and focused on selected ASEAN-5 countries. This research is conducted to describe how

More information

Introduction to Modern Time Series Analysis

Introduction to Modern Time Series Analysis Introduction to Modern Time Series Analysis Gebhard Kirchgässner, Jürgen Wolters and Uwe Hassler Second Edition Springer 3 Teaching Material The following figures and tables are from the above book. They

More information

Multivariate Time Series Analysis and Its Applications [Tsay (2005), chapter 8]

Multivariate Time Series Analysis and Its Applications [Tsay (2005), chapter 8] 1 Multivariate Time Series Analysis and Its Applications [Tsay (2005), chapter 8] Insights: Price movements in one market can spread easily and instantly to another market [economic globalization and internet

More information

A Guide to Modern Econometric:

A Guide to Modern Econometric: A Guide to Modern Econometric: 4th edition Marno Verbeek Rotterdam School of Management, Erasmus University, Rotterdam B 379887 )WILEY A John Wiley & Sons, Ltd., Publication Contents Preface xiii 1 Introduction

More information

On Perron s Unit Root Tests in the Presence. of an Innovation Variance Break

On Perron s Unit Root Tests in the Presence. of an Innovation Variance Break Applied Mathematical Sciences, Vol. 3, 2009, no. 27, 1341-1360 On Perron s Unit Root ests in the Presence of an Innovation Variance Break Amit Sen Department of Economics, 3800 Victory Parkway Xavier University,

More information

Stationary and nonstationary variables

Stationary and nonstationary variables Stationary and nonstationary variables Stationary variable: 1. Finite and constant in time expected value: E (y t ) = µ < 2. Finite and constant in time variance: Var (y t ) = σ 2 < 3. Covariance dependent

More information

Multivariate Time Series: VAR(p) Processes and Models

Multivariate Time Series: VAR(p) Processes and Models Multivariate Time Series: VAR(p) Processes and Models A VAR(p) model, for p > 0 is X t = φ 0 + Φ 1 X t 1 + + Φ p X t p + A t, where X t, φ 0, and X t i are k-vectors, Φ 1,..., Φ p are k k matrices, with

More information

Stationarity Revisited, With a Twist. David G. Tucek Value Economics, LLC

Stationarity Revisited, With a Twist. David G. Tucek Value Economics, LLC Stationarity Revisited, With a Twist David G. Tucek Value Economics, LLC david.tucek@valueeconomics.com 314 434 8633 2016 Tucek - October 7, 2016 FEW Durango, CO 1 Why This Topic Three Types of FEs Those

More information

A Gaussian IV estimator of cointegrating relations Gunnar Bårdsen and Niels Haldrup. 13 February 2006.

A Gaussian IV estimator of cointegrating relations Gunnar Bårdsen and Niels Haldrup. 13 February 2006. A Gaussian IV estimator of cointegrating relations Gunnar Bårdsen and Niels Haldrup 3 February 26. Abstract. In static single equation cointegration regression models the OLS estimator will have a non-standard

More information

Co-integration and Error-Correction Modeling of Agricultural Output. A Case of Groundnut

Co-integration and Error-Correction Modeling of Agricultural Output. A Case of Groundnut Co-integration and Error-Correction Modeling of Agricultural Output. A Case of Groundnut Ngbede,Samson Ochoche,Akintola,Joseph Olatunji 1. National Horticultural Research Institute,Mbato p m b 1076 Okigwe

More information

Department of Economics, UCSB UC Santa Barbara

Department of Economics, UCSB UC Santa Barbara Department of Economics, UCSB UC Santa Barbara Title: Past trend versus future expectation: test of exchange rate volatility Author: Sengupta, Jati K., University of California, Santa Barbara Sfeir, Raymond,

More information

Econometrics and Structural

Econometrics and Structural Introduction to Time Series Econometrics and Structural Breaks Ziyodullo Parpiev, PhD Outline 1. Stochastic processes 2. Stationary processes 3. Purely random processes 4. Nonstationary processes 5. Integrated

More information

Time Series Analysis. James D. Hamilton PRINCETON UNIVERSITY PRESS PRINCETON, NEW JERSEY

Time Series Analysis. James D. Hamilton PRINCETON UNIVERSITY PRESS PRINCETON, NEW JERSEY Time Series Analysis James D. Hamilton PRINCETON UNIVERSITY PRESS PRINCETON, NEW JERSEY PREFACE xiii 1 Difference Equations 1.1. First-Order Difference Equations 1 1.2. pth-order Difference Equations 7

More information