Linear dynamic panel data models

Size: px
Start display at page:

Download "Linear dynamic panel data models"

Transcription

1 Linear dynamic panel data models Laura Magazzini University of Verona L. Magazzini (UniVR) Dynamic PD 1 / 67

2 Linear dynamic panel data models Dynamic panel data models Notation & Assumptions One of the advantage of panel data is that allows the study of dynamics For a randomly drawn cross section observation, the basic linear multivariate dynamic panel data model can be written as (t = 2,3,...,T) y it = ρy it 1 +x itβ +c i +u it or, alternatively as y it = w itγ +c i +u it where w it = (y it 1,x it ) and γ = (ρ,β ) As usual u it denotes the idiosyncratic error term c i captures individual heterogeneity More complicated dynamic structures which include additional lags of the dependent variables and/or a distributed lag structure for the variables in x can be accomodted in this framework L. Magazzini (UniVR) Dynamic PD 2 / 67

3 Linear dynamic panel data models Dynamic modeling Dynamic models are of interest in a wide range of economic applications including Euler equations for household consumption Adjustment cost models for firms factor demands (e.g. labor demand, investment); Empirical models of economic growth Consistent estimates of other parameters of interest (e.g. estimation of dynamic production functions) Panel data is now widely used to estimate dynamic models Obvious advantage over cross-section data in this context: we cannot estimate dynamic models from observations at a single point in time Advantages over aggregate time series: (i) possibility that underlying microeconomic dynamics may be obscured by aggregation biases; (ii) the scope to investigate heterogeneity in adjustment dynamics between different types of individuals L. Magazzini (UniVR) Dynamic PD 3 / 67

4 Linear dynamic panel data models FE or RE? Is it possible to apply FE or RE? Obviouly no: the model violates both the orthogonality condition and the strict exogeneity assumption The regressor y it 1 is correlated both with c i and u it 1 : y it 1 = ρy it 2 +x it 1β +c i +u it 1 L. Magazzini (UniVR) Dynamic PD 4 / 67

5 Linear dynamic panel data models The idea of internal instruments Suppose that u it is uncorrelated over time and consider the first difference of our model y it y it 1 = ρ(y it 1 y it 2 )+(x it x it 1 ) β+(c i c i )+(u it u it 1 ) y it = ρ y it 1 + x itβ + u it OLS is inconsistent: y it 1 is correlated with u it You need to apply IV estimation Can you think of a valid instrument? Values of y it 2 are valid instruments in the equation estimated in first differences (Anderson & Hsiao, 1981, 1982) L. Magazzini (UniVR) Dynamic PD 5 / 67

6 Linear dynamic panel data models Sequential exogeneity E(u it w it,w it 1,...,w i2,c i ) = 0 w it (t = 1,2,...T) are sequentially exogenous conditional on unobserved heterogeneity c i One implication of sequential exogeneity is with t = 2,3,...,T and s t E(w is u it ) = 0 In words, the idiosyncratic error in each time period is assumed to be uncorrelated only with past and present values of the explanatory variables but is allowed to be correlated with future values of the explanatory variables L. Magazzini (UniVR) Dynamic PD 6 / 67

7 Linear dynamic panel data models Show that sequential exogeneity implies E(w is u it ) = 0 with t = 2,3,...,T and s t E(w isu it ) = E[E(w isu it w is )] = E[w ise(u it w is )] = E[w ise(e(u it w it,w it 1,...,w i2,c i ) w is )] = E[w ise(0 w is )] = 0 L. Magazzini (UniVR) Dynamic PD 7 / 67

8 Linear dynamic panel data models Another useful implication of sequential exogeneity is E(u it u it j ) = 0 for j > 0 In words, under sequential exogeneity, the idiosyncratic error terms are serially uncorrelated In fact, by the Law of Total Expectations E(u it u it j ) = E[E(u it u it j w it,w it 1,w i2,c i )] Since u it j is a linear combination of the variables in the conditioning set E[E(u it u it j w it,w it 1,w i2,c i )] = E[u it j E(u it w it,w it 1,w i2,c i )] = 0 L. Magazzini (UniVR) Dynamic PD 8 / 67

9 The first order AR panel data model y it = ρy it 1 +v it v it = c i +u it ρ < 1 t = 2,...,T Standard panel data estimation methods such as within group (WG) and first differences (FD) allow for unrestricted correlation between unobserved heterogeneity and past, present and future values of the right hand side variables However, in order to achieve consistency they require that all regressors are strictly exogenous This assumption is clearly violated in the case of the lagged dependent variable since u it cannot be orthogonal (and therefore not correlated) to y it,y it 1,...,y it L. Magazzini (UniVR) Dynamic PD 9 / 67

10 Inconsistent of pooled OLS, WG, FD Under sequential exogeneity it can be shown that pooled OLS, WG and FD estimators are all inconsistent for fixed T There is a difference however since under appropriate stability conditions ( ρ < 1 in the AR(1) model) the inconsistency in the WG estimator but not in the pooled OLS or in the FD estimator is of order T 1 This in turn implies that the asymptotic bias of the WG estimator but not of the OLS and FD estimators goes to zero as T goes to infinity L. Magazzini (UniVR) Dynamic PD 10 / 67

11 Inconsistency of pooled OLS For the pooled OLS estimator we can write ˆρ OLS = ( N T i=1 t=2 y 2 it 1 ) 1 ( N i=1 t=2 ) T y it 1 y it To study the asymptotic properties of OLS, we substitute y it = ρy it 1 +v it to get: ˆρ OLS = ρ+ ( N T i=1 t=2 y 2 it 1 ) 1 ( N i=1 t=2 ) T y it 1 v it L. Magazzini (UniVR) Dynamic PD 11 / 67

12 Inconsistency of pooled OLS The asymptotic bias of the simple pooled OLS estimator for ρ is given by plim(ˆρ OLS ρ) = (1 ρ) σ 2 c σ 2 u σc 2 σu ρ 1+ρ > 0 where E(c 2 i ) = σ2 c and E(u 2 it ) = σ2 u It follows that the pooled OLS estimator is biased upwards and its asymptotic bias does not depend on T The consistency of the pooled OLS estimator cannot be rescued by allowing the time dimension to grow without bounds L. Magazzini (UniVR) Dynamic PD 12 / 67

13 Inconsistency of WG For the WG regression, we can write ˆρ WG = ( N T i=1 t=2 ÿ 2 it 1 with ÿ it = y it ȳ i and ȳ i = T y it/t ) 1 ( N i=1 t=2 ) T ÿ it 1 ÿ it As in the previous case, to analyse the asymptotic properties of ˆρ WG we substitute ÿ it = ρÿ it 1 +ü it : ˆρ WG = ρ+ ( N T i=1 t=2 ÿ 2 it 1 ) 1 ( N i=1 t=2 ) T ÿ it 1 ü it L. Magazzini (UniVR) Dynamic PD 13 / 67

14 Inconsistency of WG Nickell (1981) derives the analytical expression for the asymptotic bias ( ) plim(ˆρ WG ρ) = 1 1+ρ T T 2ρ (1 ρ)(t 1) Nickell also provides a simple approximation 1 ρ T 1 ρ ( 1 1 T ) 1 ρ T 1 ρ plim(ˆρ WG ρ) 1+ρ T 1 for reasonably large values of T This makes it clear that when ρ > 0, the WG estimator is biased downwards Furthermore, even with T = 10, which is the order of magnitude of most sets of panel data, if ρ = 0.5, than the asymptotic bias is.167 If T also goes to infinity, the asymptotic bias converges to 0 so that the WG estimator is consistent for ρ if both N and T L. Magazzini (UniVR) Dynamic PD 14 / 67

15 Suggestion for applied work The fact that the pooled OLS estimator is biased upward and the WG estimator is biased downward can be used to assess the performance of consistent estimators in applications If the estimate produced by a consistent estimator lies outside the interval defined by the pooled OLS (upper bound) and the WG (loeer bound) estimates, the chosen consistent estimator is likely to have finite sample problems L. Magazzini (UniVR) Dynamic PD 15 / 67

16 Inconsistency of FD For the FD estimator, we can write ˆρ FD = ( N T i=1 t=2 y 2 it 1 ) 1 ( N i=1 t=2 ) T y it 1 y it with y it = y it y it 1 To study the asymptotic properties of ˆρ FD, we substitute y it = ρ y it 1 + u it to get: ˆρ FD = ρ+ ( N T i=1 t=2 y 2 it 1 ) 1 ( N i=1 t=2 ) T y it 1 u it L. Magazzini (UniVR) Dynamic PD 16 / 67

17 Inconsistency of FD The asymptotic bias of the FD estimator is given by: plim(ˆρ FD ρ) = 1+ρ 2 < 0 and therefore the FD estimator is biased downwards and its asymptotic bias does not depend on T L. Magazzini (UniVR) Dynamic PD 17 / 67

18 How can ρ be consistently estimated? The simplest linear dynamic panel data model is the first order autoregressive panel data model: y it = ρy it 1 +c i +u it with ρ < 1 and t = 2,...,T In order to consistently estimate ρ: 1. Use the First Difference transformation to remove the individual effect y it = ρ y it 1 + u it 2. Exploit the sequential exogeneity assumption to find instruments for y it 1 (and, therefore, moment conditions) 3. ρ can be consistently estimated by using the GMM estimator L. Magazzini (UniVR) Dynamic PD 18 / 67

19 Assumption: sequential exogeneity E(u it y it 1,y it 2,...,y i1,c i ) = E(u it y t 1 i,c i ) = 0 Recall that the seq.exo. assumption implies lack of autocorrelation in u it Lagged u is are linear combinations of the variables in the conditioning set (e.g. u it 1 = y it 1 ρy it 2 c i ) How can this assumption be exploited to find instruments for y it 1? L. Magazzini (UniVR) Dynamic PD 19 / 67

20 Anderson & Hsiao (1981, 1982) One simple possibility suggested by Anderson and Hsiao (1981, 1982) is to use y it 2 as instrument for y it 1 Relevance: Is y it 2 correlated with y it 1? Exogeneity: Is y it 2 uncorrelated with u it 1? That is, E[y it 2 u it ] = 0? They also proposed an alternative where y it 2 is used instead as instrument Both these IV estimators thus exploit one moment condition in estimation, respectively E[y it 2 u it ] = E[y it 2 ( y it ρ y it 1 )] = 0 and E[ y it 2 u it ] = E[ y it 2 ( y it ρ y it 1 )] = 0 L. Magazzini (UniVR) Dynamic PD 20 / 67

21 Arellano & Bond (1991) Arellano and Bond (1991) proposed to use the entire set of instruments (or at least a larger subset) in a GMM procedure Under seq.exo. values of y lagged twice or more are valid instruments for y it 1 in the first differecing version of the equation of interest y it = ρ y it 1 + u it So, at a generic time t, we can use y i1,...,y it 2 as potential instruments for y it 1 L. Magazzini (UniVR) Dynamic PD 21 / 67

22 Arellano & Bond (1991) For example, when T = 5, we can write Moment conditions for t = 3 Moment conditions for t = 4 Moment conditions for t = 5 E[y i1 ( y i3 ρ y i2 )] = 0 E[y i2 ( y i4 ρ y i3 )] = 0 E[y i1 ( y i4 ρ y i3 )] = 0 E[y i3 ( y i5 ρ y i4 )] = 0 E[y i2 ( y i5 ρ y i4 )] = 0 E[y i1 ( y i5 ρ y i4 )] = 0 L. Magazzini (UniVR) Dynamic PD 22 / 67

23 Practical suggestion The first difference GMM approach generates moment conditions prolifically, with the instrument count quadratic in T T (T 1)(T 2) Since the number of elements in the estimated variance matrix is quadratic in the instrument count, it is quartic in T A finite sample may lack adequate information to estimate such a large matrix In applications it is common (and wise) practice to use only less distant lags as instruments (t 2,t 3 or t 2,t 3,t 4) L. Magazzini (UniVR) Dynamic PD 23 / 67

24 Extensions This method extends easily to models with limited moving-average serial correlation in u it For instance, if u it is itself MA(1), then the first differenced error term u it is MA(2) As a consequence, y it 2 would not be a valid instrumental variable in these first-differenced equations but y it 3 and longer lags remain available as instruments In this case ρ is identified provided T > 4 L. Magazzini (UniVR) Dynamic PD 24 / 67

25 Asymptotic Distribution Under the assumptions Sequential exogeneity Conditional homoschedasticity Time series homoschedasticity the first difference one-step GMM estimator is consistent, efficient and asymptotically normal If homoschedasticity is not satisfied, the one-step estimator is no longer efficient (still, it remains consistent) A heteroschedasticity robust estimator of the asymptotic variance matrix for the one-step estimator is also available Under the assumption of sequential exogeneity, the first difference two-step GMM estimator is efficient and asymptotically normally distributed L. Magazzini (UniVR) Dynamic PD 25 / 67

26 Testing the Assumptions Autocorrelation test by Arellano & Bond The consistency of the GMM-FD estimator relies on the absence of first (and higher) order serial correlation in the disturbances E(u it u it 1 ) = 0 t = 3,4,...,T Since the model is estimated in first differences it is useful to rewrite this condition as E( u it u it 2 ) = 0 t = 5,4,...,T Lack of autocorrelation in u it implies that u it is autocorrelated of order 1, whereas higher order autocorrelation is zero L. Magazzini (UniVR) Dynamic PD 26 / 67

27 Testing the assumption This assumption can be tested by using the standardized average residual autocovariance N T i=1 t=5 m 2 = û it û it 2 ( ) The test statistic is asymptotically N(0,1) under the null of no serial correlation in u it ( ): for details refer to Arellano and Bond (1991) L. Magazzini (UniVR) Dynamic PD 27 / 67

28 Sargan test More generally, the Sargan test of overidentifying restrictions can also be computed when A N is chosen optimally The test statistic s is asymptotically χ 2 with as many degrees of freedom as overidentifying restrictions, under the null hypothesis of the validity of the instruments # overidentifying restrictions = # moment conditions # estimated parameters L. Magazzini (UniVR) Dynamic PD 28 / 67

29 Finite sample results Monte Carlo set up by Arellano and Bond (1991) y it = ρy it 1 +x it +c i +u it with y i0 = 0 x it = 0.8x it 1 +ε it c i iidn(0,1); u it iidn(0,1); and ε it iidn(0,1) and independent of c i,u is for all t,s Sample size: N = 100; T = 7 Mean and standard deviations computed on the basis of 100 replications L. Magazzini (UniVR) Dynamic PD 29 / 67

30 Monte Carlo results GMM1 GMM2 OLS WG 1sASE R1sASE 2sASE ρ = 0.5 Mean Std.Dev β = 1 Mean Std.Dev sASE: one-step aymptotic standard errors R1sASE: robust one-step aymptotic standard errors 2sASE: two-step aymptotic standard errors L. Magazzini (UniVR) Dynamic PD 30 / 67

31 Comments - 2sASE In the Monte Carlo simulations the estimator of the asymptotic standard errors of GMM2 in the last column (2sASE) shows a sizable downward bias relative to the finite-sample standard deviation reported in the second column Why? The efficient GMM estimator uses A 2s N A 2s N ( N = 1 N A 2s as the weighting matrix ) 1 N Z Di û i û iz Di i=1 is a function of estimated fourth moments Generally, it takes a substantially larger sample size to estimate fourth moments reliably than to estimate first and second moments L. Magazzini (UniVR) Dynamic PD 31 / 67

32 Size of the test statistics Number of rejections (out of 100 cases) LoS m2 1s m2 1sr m2 2s Sargan 10% % % LoS: level of significance Experiments with ρ = 0.5 L. Magazzini (UniVR) Dynamic PD 32 / 67

33 Power of the test statistics Number of rejections (out of 100 cases) Experiments with different degree of serial correlation: corr(u it,u it 1 ) = ρ u 0 LoS m2 1s m2 1sr m2 2s Sargan ρ u = % % % ρ u = % % % L. Magazzini (UniVR) Dynamic PD 33 / 67

34 Robustness of the test statistics Number of rejections (out of 100 cases) Experiments with heteroschedasticity: var(u it ) = x 2 it LoS m2 1s m2 1sr m2 2s Sargan 10% % % L. Magazzini (UniVR) Dynamic PD 34 / 67

35 The weak instruments problem Blundell and Bond (1998) show that the instruments used in the standard first difference GMM estimator become less informative in two important cases First, as the value of ρ increases towards unity, and second as the variance of the unobserved heterogeneity component, σ 2 c increases with respect to the variance of the idiosyncratic error term, σ 2 u To see this we can focus on the case with T = 3: We are left with only one moment condition E(y i1 u i3 ) = 0 ρ is just identified L. Magazzini (UniVR) Dynamic PD 35 / 67

36 Weak instruments problem with T = 3 In this case we use one equation in first differences (i = 1,...,N) y i3 = ρ y i2 + u i3 where we use y i1 as an instrument for y i2 The GMM estimator reduces to a simple IV estimator with the following reduced form equation y i2 = πy i1 +r i For sufficiently high values of ρ or σ 2 c/σ 2 u, the OLS estimate of the reduced form coefficient π can be made arbitrarily close to zero In this case y i1 is only weakly correlated with y i2 and thus a weak instruments problem arise L. Magazzini (UniVR) Dynamic PD 36 / 67

37 Weak instruments problem with T = 3 It can be shown that plimˆπ LS = (ρ 1) σ 2 c σ 2 u (1 ρ) 2 (1 ρ 2 ) + (1 ρ)2 (1 ρ 2 ) For instance, with σ2 c = 1 and ρ = 0.8, plimˆπ σu 2 LS = 0.02 In words, the instruments available for the equations in first-differences are likely to be weak when the individual series are persistent, that is when they have near unit root properties Blundell and Bond (1998) propose additional moment conditions that allow to solve this problem See also Arellano and Bover (1995) L. Magazzini (UniVR) Dynamic PD 37 / 67

38 GMM system estimator y it = ρy it 1 +c i +u it Arellano and Bond (1991) GMM-DIF estimator uses lags of y it (lag two or older) as instruments for the equation in first differences Blundell and Bond (1998) suggest using lags of y it (lag one or older) as instruments for the equation in levels E(v it y it 1 ) = 0 t = 3,4,...,T An additional assumption is needed for consistency: the mean stationarity assumption L. Magazzini (UniVR) Dynamic PD 38 / 67

39 The mean stationarity assumption Roodman (2009) y it = ρy it 1 +c i +u it Entities in this system can evolve much like GDP per worker in the Solow growth model, converging towards mean stationarity By itself, a positive fixed effect, for instance, provides a constant, repeated boost to y in each period, like investment does for the capital stock But, assuming ρ < 1, this increment is offset by reversion towards the mean The observed entities therefore converge to steady-state levels defined by E[y it c i ] = E[y it+1 c i ] y it = ρy it +c i y it = c i 1 ρ The fixed effects and the autocorrelation coefficient interact to determine the long-run mean of the series L. Magazzini (UniVR) Dynamic PD 39 / 67

40 The mean stationarity assumption Roodman (2009) Now consider the momemt conditions proposed by Blundell and Bond (1998): E(v it y it 1 ) = 0, t = 3,4,...,T We can write (t 3) E(v it y it 1 ) = E[(c i +u it )(y it 1 y it 2 )] = E[(c i +u it )(ρy it 2 +c i +u it 1 y it 2 )] = E[(c i +u it )((ρ 1)y it 2 +c i +u it 1 )] = E[c i ((ρ 1)y it 2 +c i )] = 0 which is equivalent to E[c i ((ρ 1)y it +c i )] = 0 for t 1 L. Magazzini (UniVR) Dynamic PD 40 / 67

41 The mean stationarity assumption Roodman (2009) E[c i ((ρ 1)y it +c i )] = 0 for t 1 By dividing this condition by (1 ρ) we get: ( E [c i y it c )] i = 0 1 ρ Deviations from long-run means must not be correlated with the fixed effects It is possible to show that if this condition holds in t then it also holds in all subsequent periods Effectively, this is a condition on the initial observation L. Magazzini (UniVR) Dynamic PD 41 / 67

42 GMM system estimator The conditions imposed by Blundell and Bond (1998) imply that additional 0.5(T 1)(T 2) moment conditions for the equation in levels are available However, all but T 3 moment restrictions are redundant since they can be expressed as linear combinations of the GMM-DIF moment restrictions L. Magazzini (UniVR) Dynamic PD 42 / 67

43 For example, with T = 5 Which restrictions are redundant? Moment conditions when t = 3 Moment conditions when t = 4 Moment conditions when t = 5 E[ y i2 (y i3 ρy i2 )] = 0 E[ y i3 (y i4 ρy i3 )] = 0 E[ y i2 (y i4 ρy i3 )] = 0 E[ y i4 (y i5 ρy i4 )] = 0 E[ y i3 (y i5 ρy i4 )] = 0 E[ y i2 (y i5 ρy i4 )] = 0 L. Magazzini (UniVR) Dynamic PD 43 / 67

44 GMM system estimator Summarizing, the enlarged set of moment conditions is E[y t 2 i ( y it ρ y it 1 )] = 0 t = 3,4,...,T E[ y it 1 (y it ρy it 1 )] = 0 t = 3,4,...,T The GMM method can be applied for the estimation of ρ L. Magazzini (UniVR) Dynamic PD 44 / 67

45 Practical Suggestion The validity of the additional orthogonality conditions (spanning from the mean stationarity assumption) can be tested by using the so-called Sargan difference test of overidentified retstrictions The model is estimated twice, without and with the additional moment conditions The difference in the value of the Sargan statistic is found to have a χ 2 distribution under the null (the validity of these additional moment conditions), where the degrees of freedom are given by the number of additional restrictions L. Magazzini (UniVR) Dynamic PD 45 / 67

46 The weak instrument problem is solved To show that these additional moment restrictions remain informative when ρ increases towards unity and/or when the variance of the unobserved heterogeneity component, σ 2 c increases with respect to the variance of the idiosyncratic error term, σ 2 u, we can again focus on the case with T = 3 We are left with only one moment condition E(v i3 y i2 ) = 0, so that ρ is just identified Here, we can use one equation in level y i3 = ρy i2 +c i +u i3 where we use y i2 as an instrument for y i2 The corresponding GMM estimator is an IV estimator with the following reduced form equation y i2 = π y i2 +r i L. Magazzini (UniVR) Dynamic PD 46 / 67

47 The weak instrument problem is solved It is possible to show that, in this case plimˆπ LS = 1 1 ρ 21 ρ 2 This moment condition stays informative for high values of ρ, in contrast to the moment condition available for the first differenced method For instance, plimˆπ LS = 0.28 for ρ = 0.8 L. Magazzini (UniVR) Dynamic PD 47 / 67

48 Finite sample results Monte Carlo set up by Blundell and Bond (1998) y it = ρy it 1 +c i +u it The mean stationarity assumption is satisfied: y i0 = c i 1 ρ +ε i1 c i iidn(0,1); u it iidn(0,1) ε i1 iidn(0,σ 2 ε) and independent of c i,u is for all t,s Sample size: N = 100; T = 4,11 Mean and standard deviations computed on the basis of 1000 replications L. Magazzini (UniVR) Dynamic PD 48 / 67

49 Monte Carlo results T = 4 GMM-DIF. GMM-SYS ρ Mean Std.Dev RMSE Mean Std.Dev RMSE L. Magazzini (UniVR) Dynamic PD 49 / 67

50 Monte Carlo results T = 11 GMM-DIF GMM-SYS ρ Mean Std.Dev RMSE Mean Std.Dev RMSE L. Magazzini (UniVR) Dynamic PD 50 / 67

51 Comments In these simulations, when T = 4 we observe both the poor performance of the first-differenced GMM estimator at high values of ρ and the dramatic improvement in precision that results from exploiting the additional moment conditions The poor performance of the FD-GMM estimator improves with the number of time periods available However, even with T = 11, for high values of ρ there remain sizable gains in bias, precision and RMSE L. Magazzini (UniVR) Dynamic PD 51 / 67

52 Multivariate dynamic panel data models The multivariate dynamic panel data model is ( ρ < 1; t = 2,3,...,T) y it = ρy it 1 +x itβ +c i +u it Sequential exogeneity assumption E(u it x t i,y t 1 i,c i ) = 0 As in the first order autoregressive panel data model, note that GMM-DIF1 implies lack of serial correlation in u it L. Magazzini (UniVR) Dynamic PD 52 / 67

53 Predetermined versus endogenous variables The sequential exogeneity assumption (E(u it x t i,yt 1 i,c i ) = 0) implies that the idiosyncratic error in each time period is assumed to be uncorrelated only with past and present values of the explanatory variables (predetermined variables) This allows for feedback effects but not for simultaneity To allow for simultaneity, the idiosyncratic error term in each time period has to be assumed to be uncorrelated only with past values of the explanatory variables (endogenous variables) In this case we write: E(u it x t 1 i,y t 1 i,c i ) = 0 L. Magazzini (UniVR) Dynamic PD 53 / 67

54 Consistency The sequential exogeneity assumption also implies that the following 0.5(T 1)(T 2)+0.5(T +1)(T 2)K linear moment restrictions hold E[y t 2 i ( y it ρ y it 1 x it β)] = 0 E[x t 1 ki ( y it ρ y it 1 x it β)] = 0 The estimating equation is in first differences y it = ρ y it 1 + x it β + u it Under the sequential exogeneity assumption, values of y lagged twice or more and values of x lagged once or more are valid instruments for y it 1 and x it L. Magazzini (UniVR) Dynamic PD 54 / 67

55 Example: T = 5 1+2k moment conditions for t = 3 E[y i1 ( y i3 ρ y i2 x i3 β)] = 0 E[x i1 ( y i3 ρ y i2 x i3 β)] = 0 E[x i2 ( y i3 ρ y i2 x i3 β)] = 0 2+3K moment conditions for t = 4 E[y i1 ( y i4 ρ y i3 x i4 β)] = 0 E[y i2 ( y i4 ρ y i3 x i4 β)] = 0 E[x i1 ( y i4 ρ y i3 x i4 β)] = 0 E[x i2 ( y i4 ρ y i3 x i4 β)] = 0 E[x i3 ( y i4 ρ y i3 x i4 β)] = 0 L. Magazzini (UniVR) Dynamic PD 55 / 67

56 Finally, 3+4K moment conditions for t = 5 E[y i1 ( y i5 ρ y i4 x i5 β)] = 0 E[y i2 ( y i5 ρ y i4 x i5 β)] = 0 E[y i3 ( y i5 ρ y i4 x i5 β)] = 0 E[x i1 ( y i5 ρ y i4 x i5 β)] = 0 E[x i2 ( y i5 ρ y i4 x i5 β)] = 0 E[x i3 ( y i5 ρ y i4 x i5 β)] = 0 E[x i4 ( y i5 ρ y i4 x i5 β)] = 0 The GMM method allows you to obtain consistent estimates of the parameters L. Magazzini (UniVR) Dynamic PD 56 / 67

57 Practical suggestion In the multivariate case, the first difference GMM approach becomes even more prolific than in the simple autoregressive case For instance, with K = 2 T (T 1)(T 2)+(T +1)(T 2) In applications it is common (and wise) practice to use only less distant lags as instruments: t 2, t 3 for y and t 1 and t 2 for x (t 2, t 3 if the variables in x are endogenous) L. Magazzini (UniVR) Dynamic PD 57 / 67

58 Extensions This method extends easily to models with limited moving-average serial correlation in u it For instance, if u it is itself MA(1), then the first differenced error term u it is MA(2) As a consequence, y it 2 would not be a valid instrumental variable in these first-differenced equations but y it 3 and longer lags remain available as instruments In this case ρ is identified provided T > 4 L. Magazzini (UniVR) Dynamic PD 58 / 67

59 GMM-SYS estimator Assumptions Sequential exogeneity Mean stationarity also for x These assumptions lead to additional non-redundant moment conditions E[ w it (y it ρy it 1 x it β)] = 0 which can be exploited in the context of the GMM-SYS estimator L. Magazzini (UniVR) Dynamic PD 59 / 67

60 Variance of the 2-step GMM estimator Finite sample correction Montecarlo studies (including AB, 1991) have shown that estimated asymptotic standard errors of the efficient two-step GMM estimator are severely downward biased in small samples The weight matrix used in the calculation of the efficient two-step estimator is based on initial consistent parameter estimates Windmeijer (2005) shows that the extra variation due to the presence of these estimated parameters in the weight matrix accounts for much of the difference between the finite sample and the asymptotic variance of the two step estimator This difference can be estimated, resulting in a finite sample correction of the variance L. Magazzini (UniVR) Dynamic PD 60 / 67

61 Monte Carlo experiments Windmejier (2005) y it = βx it +c i +u it x it = 0.5x it 1 +c i +0.5u it 1 +ε it c i iidn(0,1) and ε it iidn(0,1) u it = δ i τ u ω it with ω it χ 2 1 1; δ i U(0.5,1.5); τ t = (t 1) N = replications Note that x it are correlated with c i and predetermined with respect to u it u it are skewed and heteroschedastic over time and units L. Magazzini (UniVR) Dynamic PD 61 / 67

62 Monte Carlo results β = 1 Corrected Mean Std.Dev. Std.Err. Std.Err. T = 4 ˆβ GMMd 1s ˆβ GMMd 2s T = 8 ˆβ GMMd 1s ˆβ GMMd 2s L. Magazzini (UniVR) Dynamic PD 62 / 67

63 Comments Non-corrected two-step standard errors are severely downward biased, especially when T = 8 Both one-step and corrected two-step standard errors are very close to standard deviations Additional simulations also show that Wald tests based on non-corrected two-step estimated variances are oversized This is no more the case when corrected two-step variances are used instead Finally, the size performance of the Sargan/Hansen test is not affected by the estimation of ˆβ GMMd 1s used to construct the weight matrix L. Magazzini (UniVR) Dynamic PD 63 / 67

64 Selected references Anderson, T.W. and C. Hsiao (1981), Estimation of Dynamic Models with Error Components, Journal of the American Statistical Association, 76, Arellano, M. and S. Bond (1991), Some Tests of Specification for Panel Data: Montecarlo Evidence and an Application to Employment Equations, Review of Economic Studies, 58, Blundell, R. and S. Bond (1998), Initial Conditions and Moment Restrictions in Dynamic Panel Data Models, Journal of Econometrics, 87, Nickell, S. (1981), Biases in Dynamic Models with Fixed Effects, Econometrica, 49, L. Magazzini (UniVR) Dynamic PD 64 / 67

65 Selected references Roodman, D. (2009), A Note on the Theme of Too Many Instruments, Oxford Bulletin of Economics and Statistics, 71, Windmeijer, F. (2005), A Finite Sample Correction for the Variance of Linear Two-step GMM Estimators, Journal of Econometrics, 126, L. Magazzini (UniVR) Dynamic PD 65 / 67

66 Further readings Bun, M. (2003), Bias Correction in the Dynamic Panel Data Model with a Nonscalar Disturbance Covariance Matrix, Econometric Reviews, 22, Bun, M. and M. Carree (2006), Bias Corrected Estimation in Dynamic Panel Data Models with Heteroskedasticity, Economic Letters, 92, Bun, M. and J.F Kiviet (2006), The Effects of Dynamic Feedbacks on LS and MM Estimator Accuracy in Panel Data Models, Journal of Econometrics, 127, Kiviet, J. F. (1995), On Bias, Inconsistency and Efficiency of Various Estimators in Dynamic Panel Data Models, Journal of Econometrics, 68, L. Magazzini (UniVR) Dynamic PD 66 / 67

67 Further readings Judson, R. and A. Owen (1999), Estimating Dynamic Panel Data Models: a Guide for Macroeconomists, Economic Letters, 65, 9-15 Hsiao, C., Pesaran, M.H. and Tahmiscioglu A.K.: 2002, Maximum Likelihood Estimation of Fixed Effects Dynamic Panel Data Models Covering Short Time Periods, Journal of Econometrics 109(1), L. Magazzini (UniVR) Dynamic PD 67 / 67

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

Improving GMM efficiency in dynamic models for panel data with mean stationarity

Improving GMM efficiency in dynamic models for panel data with mean stationarity Working Paper Series Department of Economics University of Verona Improving GMM efficiency in dynamic models for panel data with mean stationarity Giorgio Calzolari, Laura Magazzini WP Number: 12 July

More information

Dynamic Panels. Chapter Introduction Autoregressive Model

Dynamic Panels. Chapter Introduction Autoregressive Model Chapter 11 Dynamic Panels This chapter covers the econometrics methods to estimate dynamic panel data models, and presents examples in Stata to illustrate the use of these procedures. The topics in this

More information

Chapter 6. Panel Data. Joan Llull. Quantitative Statistical Methods II Barcelona GSE

Chapter 6. Panel Data. Joan Llull. Quantitative Statistical Methods II Barcelona GSE Chapter 6. Panel Data Joan Llull Quantitative Statistical Methods II Barcelona GSE Introduction Chapter 6. Panel Data 2 Panel data The term panel data refers to data sets with repeated observations over

More information

Bias Correction Methods for Dynamic Panel Data Models with Fixed Effects

Bias Correction Methods for Dynamic Panel Data Models with Fixed Effects MPRA Munich Personal RePEc Archive Bias Correction Methods for Dynamic Panel Data Models with Fixed Effects Mohamed R. Abonazel Department of Applied Statistics and Econometrics, Institute of Statistical

More information

1 Estimation of Persistent Dynamic Panel Data. Motivation

1 Estimation of Persistent Dynamic Panel Data. Motivation 1 Estimation of Persistent Dynamic Panel Data. Motivation Consider the following Dynamic Panel Data (DPD) model y it = y it 1 ρ + x it β + µ i + v it (1.1) with i = {1, 2,..., N} denoting the individual

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 16, 2013 Outline Univariate

More information

Dynamic panel data methods

Dynamic panel data methods Dynamic panel data methods for cross-section panels Franz Eigner University Vienna Prepared for UK Econometric Methods of Panel Data with Prof. Robert Kunst 27th May 2009 Structure 1 Preliminary considerations

More information

Panel Data Models. Chapter 5. Financial Econometrics. Michael Hauser WS17/18 1 / 63

Panel Data Models. Chapter 5. Financial Econometrics. Michael Hauser WS17/18 1 / 63 1 / 63 Panel Data Models Chapter 5 Financial Econometrics Michael Hauser WS17/18 2 / 63 Content Data structures: Times series, cross sectional, panel data, pooled data Static linear panel data models:

More information

EC327: Advanced Econometrics, Spring 2007

EC327: Advanced Econometrics, Spring 2007 EC327: Advanced Econometrics, Spring 2007 Wooldridge, Introductory Econometrics (3rd ed, 2006) Chapter 14: Advanced panel data methods Fixed effects estimators We discussed the first difference (FD) model

More information

Linear Panel Data Models

Linear Panel Data Models Linear Panel Data Models Michael R. Roberts Department of Finance The Wharton School University of Pennsylvania October 5, 2009 Michael R. Roberts Linear Panel Data Models 1/56 Example First Difference

More information

Repeated observations on the same cross-section of individual units. Important advantages relative to pure cross-section data

Repeated observations on the same cross-section of individual units. Important advantages relative to pure cross-section data Panel data Repeated observations on the same cross-section of individual units. Important advantages relative to pure cross-section data - possible to control for some unobserved heterogeneity - possible

More information

Chapter 2. Dynamic panel data models

Chapter 2. Dynamic panel data models Chapter 2. Dynamic panel data models School of Economics and Management - University of Geneva Christophe Hurlin, Université of Orléans University of Orléans April 2018 C. Hurlin (University of Orléans)

More information

Non-linear panel data modeling

Non-linear panel data modeling Non-linear panel data modeling Laura Magazzini University of Verona laura.magazzini@univr.it http://dse.univr.it/magazzini May 2010 Laura Magazzini (@univr.it) Non-linear panel data modeling May 2010 1

More information

1 Introduction The time series properties of economic series (orders of integration and cointegration) are often of considerable interest. In micro pa

1 Introduction The time series properties of economic series (orders of integration and cointegration) are often of considerable interest. In micro pa Unit Roots and Identification in Autoregressive Panel Data Models: A Comparison of Alternative Tests Stephen Bond Institute for Fiscal Studies and Nuffield College, Oxford Céline Nauges LEERNA-INRA Toulouse

More information

A Transformed System GMM Estimator for Dynamic Panel Data Models. February 26, 2014

A Transformed System GMM Estimator for Dynamic Panel Data Models. February 26, 2014 A Transformed System GMM Estimator for Dynamic Panel Data Models Xiaojin Sun Richard A. Ashley February 26, 20 Abstract The system GMM estimator developed by Blundell and Bond (998) for dynamic panel data

More information

Dynamic Panel Data Models

Dynamic Panel Data Models June 23, 2010 Contents Motivation 1 Motivation 2 Basic set-up Problem Solution 3 4 5 Literature Motivation Many economic issues are dynamic by nature and use the panel data structure to understand adjustment.

More information

10 Panel Data. Andrius Buteikis,

10 Panel Data. Andrius Buteikis, 10 Panel Data Andrius Buteikis, andrius.buteikis@mif.vu.lt http://web.vu.lt/mif/a.buteikis/ Introduction Panel data combines cross-sectional and time series data: the same individuals (persons, firms,

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

System GMM estimation of Empirical Growth Models

System GMM estimation of Empirical Growth Models System GMM estimation of Empirical Growth Models ELISABETH DORNETSHUMER June 29, 2007 1 Introduction This study based on the paper "GMM Estimation of Empirical Growth Models" by Stephan Bond, Anke Hoeffler

More information

Panel Data: Linear Models

Panel Data: Linear Models Panel Data: Linear Models Laura Magazzini University of Verona laura.magazzini@univr.it http://dse.univr.it/magazzini Laura Magazzini (@univr.it) Panel Data: Linear Models 1 / 45 Introduction Outline What

More information

A Course in Applied Econometrics Lecture 4: Linear Panel Data Models, II. Jeff Wooldridge IRP Lectures, UW Madison, August 2008

A Course in Applied Econometrics Lecture 4: Linear Panel Data Models, II. Jeff Wooldridge IRP Lectures, UW Madison, August 2008 A Course in Applied Econometrics Lecture 4: Linear Panel Data Models, II Jeff Wooldridge IRP Lectures, UW Madison, August 2008 5. Estimating Production Functions Using Proxy Variables 6. Pseudo Panels

More information

Applied Econometrics (MSc.) Lecture 3 Instrumental Variables

Applied Econometrics (MSc.) Lecture 3 Instrumental Variables Applied Econometrics (MSc.) Lecture 3 Instrumental Variables Estimation - Theory Department of Economics University of Gothenburg December 4, 2014 1/28 Why IV estimation? So far, in OLS, we assumed independence.

More information

Efficient Estimation of Dynamic Panel Data Models: Alternative Assumptions and Simplified Estimation

Efficient Estimation of Dynamic Panel Data Models: Alternative Assumptions and Simplified Estimation Efficient Estimation of Dynamic Panel Data Models: Alternative Assumptions and Simplified Estimation Seung C. Ahn Arizona State University, Tempe, AZ 85187, USA Peter Schmidt * Michigan State University,

More information

Applied Microeconometrics (L5): Panel Data-Basics

Applied Microeconometrics (L5): Panel Data-Basics Applied Microeconometrics (L5): Panel Data-Basics Nicholas Giannakopoulos University of Patras Department of Economics ngias@upatras.gr November 10, 2015 Nicholas Giannakopoulos (UPatras) MSc Applied Economics

More information

Short T Panels - Review

Short T Panels - Review Short T Panels - Review We have looked at methods for estimating parameters on time-varying explanatory variables consistently in panels with many cross-section observation units but a small number of

More information

On GMM Estimation and Inference with Bootstrap Bias-Correction in Linear Panel Data Models

On GMM Estimation and Inference with Bootstrap Bias-Correction in Linear Panel Data Models On GMM Estimation and Inference with Bootstrap Bias-Correction in Linear Panel Data Models Takashi Yamagata y Department of Economics and Related Studies, University of York, Heslington, York, UK January

More information

GMM based inference for panel data models

GMM based inference for panel data models GMM based inference for panel data models Maurice J.G. Bun and Frank Kleibergen y this version: 24 February 2010 JEL-code: C13; C23 Keywords: dynamic panel data model, Generalized Method of Moments, weak

More information

A forward demeaning transformation for a dynamic count panel data model *

A forward demeaning transformation for a dynamic count panel data model * A forward demeaning transformation for a dynamic count panel data model * Yoshitsugu Kitazawa January 21, 2010 Abstract In this note, a forward demeaning transformation is proposed for the linear feedback

More information

Econometric Methods for Panel Data

Econometric Methods for Panel Data Based on the books by Baltagi: Econometric Analysis of Panel Data and by Hsiao: Analysis of Panel Data Robert M. Kunst robert.kunst@univie.ac.at University of Vienna and Institute for Advanced Studies

More information

Økonomisk Kandidateksamen 2004 (I) Econometrics 2. Rettevejledning

Økonomisk Kandidateksamen 2004 (I) Econometrics 2. Rettevejledning Økonomisk Kandidateksamen 2004 (I) Econometrics 2 Rettevejledning This is a closed-book exam (uden hjælpemidler). Answer all questions! The group of questions 1 to 4 have equal weight. Within each group,

More information

Estimation of Dynamic Panel Data Models with Sample Selection

Estimation of Dynamic Panel Data Models with Sample Selection === Estimation of Dynamic Panel Data Models with Sample Selection Anastasia Semykina* Department of Economics Florida State University Tallahassee, FL 32306-2180 asemykina@fsu.edu Jeffrey M. Wooldridge

More information

Econometrics of Panel Data

Econometrics of Panel Data Econometrics of Panel Data Jakub Mućk Meeting # 2 Jakub Mućk Econometrics of Panel Data Meeting # 2 1 / 26 Outline 1 Fixed effects model The Least Squares Dummy Variable Estimator The Fixed Effect (Within

More information

Consistent OLS Estimation of AR(1) Dynamic Panel Data Models with Short Time Series

Consistent OLS Estimation of AR(1) Dynamic Panel Data Models with Short Time Series Consistent OLS Estimation of AR(1) Dynamic Panel Data Models with Short Time Series Kazuhiko Hayakawa Department of Economics Hitotsubashi University January 19, 006 Abstract In this paper, we examine

More information

GMM Estimation of Empirical Growth Models

GMM Estimation of Empirical Growth Models GMM Estimation of Empirical Growth Models Stephen Bond Nuffield College, University of Oxford and Institute for Fiscal Studies Anke Hoeffler St.Antony s College, University of Oxford and Centre for the

More information

Dynamic Panel Data estimators

Dynamic Panel Data estimators Dynamic Panel Data estimators Christopher F Baum EC 823: Applied Econometrics Boston College, Spring 2014 Christopher F Baum (BC / DIW) Dynamic Panel Data estimators Boston College, Spring 2014 1 / 50

More information

Dealing With Endogeneity

Dealing With Endogeneity Dealing With Endogeneity Junhui Qian December 22, 2014 Outline Introduction Instrumental Variable Instrumental Variable Estimation Two-Stage Least Square Estimation Panel Data Endogeneity in Econometrics

More information

xtdpdqml: Quasi-maximum likelihood estimation of linear dynamic short-t panel data models

xtdpdqml: Quasi-maximum likelihood estimation of linear dynamic short-t panel data models xtdpdqml: Quasi-maximum likelihood estimation of linear dynamic short-t panel data models Sebastian Kripfganz University of Exeter Business School, Department of Economics, Exeter, UK UK Stata Users Group

More information

Econometrics. Week 8. Fall Institute of Economic Studies Faculty of Social Sciences Charles University in Prague

Econometrics. Week 8. Fall Institute of Economic Studies Faculty of Social Sciences Charles University in Prague Econometrics Week 8 Institute of Economic Studies Faculty of Social Sciences Charles University in Prague Fall 2012 1 / 25 Recommended Reading For the today Instrumental Variables Estimation and Two Stage

More information

Econometrics. Week 4. Fall Institute of Economic Studies Faculty of Social Sciences Charles University in Prague

Econometrics. Week 4. Fall Institute of Economic Studies Faculty of Social Sciences Charles University in Prague Econometrics Week 4 Institute of Economic Studies Faculty of Social Sciences Charles University in Prague Fall 2012 1 / 23 Recommended Reading For the today Serial correlation and heteroskedasticity in

More information

xtseqreg: Sequential (two-stage) estimation of linear panel data models

xtseqreg: Sequential (two-stage) estimation of linear panel data models xtseqreg: Sequential (two-stage) estimation of linear panel data models and some pitfalls in the estimation of dynamic panel models Sebastian Kripfganz University of Exeter Business School, Department

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 510 B. Brown Spring 2014 Final Exam Answers

Econ 510 B. Brown Spring 2014 Final Exam Answers Econ 510 B. Brown Spring 2014 Final Exam Answers Answer five of the following questions. You must answer question 7. The question are weighted equally. You have 2.5 hours. You may use a calculator. Brevity

More information

Dynamic Panel Data Models

Dynamic Panel Data Models Models Amjad Naveed, Nora Prean, Alexander Rabas 15th June 2011 Motivation Many economic issues are dynamic by nature. These dynamic relationships are characterized by the presence of a lagged dependent

More information

Econometrics Summary Algebraic and Statistical Preliminaries

Econometrics Summary Algebraic and Statistical Preliminaries Econometrics Summary Algebraic and Statistical Preliminaries Elasticity: The point elasticity of Y with respect to L is given by α = ( Y/ L)/(Y/L). The arc elasticity is given by ( Y/ L)/(Y/L), when L

More information

Econometric Analysis of Cross Section and Panel Data

Econometric Analysis of Cross Section and Panel Data Econometric Analysis of Cross Section and Panel Data Jeffrey M. Wooldridge / The MIT Press Cambridge, Massachusetts London, England Contents Preface Acknowledgments xvii xxiii I INTRODUCTION AND BACKGROUND

More information

Common Correlated Effects Estimation of Dynamic Panels with Cross-Sectional Dependence

Common Correlated Effects Estimation of Dynamic Panels with Cross-Sectional Dependence Common Correlated Effects Estimation of Dynamic Panels with Cross-Sectional Dependence om De Groote and Gerdie Everaert SHERPPA, Ghent University Preliminary, February 00 Abstract his paper studies estimation

More information

1. Overview of the Basic Model

1. Overview of the Basic Model IRP Lectures Madison, WI, August 2008 Lectures 3 & 4, Monday, August 4, 11:15-12:30 and 1:30-2:30 Linear Panel Data Models These notes cover some recent topics in linear panel data models. They begin with

More information

Econometrics of Panel Data

Econometrics of Panel Data Econometrics of Panel Data Jakub Mućk Meeting # 1 Jakub Mućk Econometrics of Panel Data Meeting # 1 1 / 31 Outline 1 Course outline 2 Panel data Advantages of Panel Data Limitations of Panel Data 3 Pooled

More information

17/003. Alternative moment conditions and an efficient GMM estimator for dynamic panel data models. January, 2017

17/003. Alternative moment conditions and an efficient GMM estimator for dynamic panel data models. January, 2017 17/003 Alternative moment conditions and an efficient GMM estimator for dynamic panel data models Gabriel Montes-Rojas, Walter Sosa-Escudero, and Federico Zincenko January, 2017 Alternative moment conditions

More information

Bootstrap Based Bias Correction for Homogeneous Dynamic Panels*

Bootstrap Based Bias Correction for Homogeneous Dynamic Panels* FACULTEIT ECONOMIE EN BEDRIJFSKUNDE HOVENIERSBERG 24 B-9000 GENT Tel. : 32 - (0)9 264.34.61 Fax. : 32 - (0)9 264.35.92 WORKING PAPER Bootstrap Based Bias Correction for Homogeneous Dynamic Panels* Gerdie

More information

Econometrics of Panel Data

Econometrics of Panel Data Econometrics of Panel Data Jakub Mućk Meeting # 3 Jakub Mućk Econometrics of Panel Data Meeting # 3 1 / 21 Outline 1 Fixed or Random Hausman Test 2 Between Estimator 3 Coefficient of determination (R 2

More information

Applied Economics. Panel Data. Department of Economics Universidad Carlos III de Madrid

Applied Economics. Panel Data. Department of Economics Universidad Carlos III de Madrid Applied Economics Panel Data Department of Economics Universidad Carlos III de Madrid See also Wooldridge (chapter 13), and Stock and Watson (chapter 10) 1 / 38 Panel Data vs Repeated Cross-sections In

More information

Nonstationary Panels

Nonstationary Panels Nonstationary Panels Based on chapters 12.4, 12.5, and 12.6 of Baltagi, B. (2005): Econometric Analysis of Panel Data, 3rd edition. Chichester, John Wiley & Sons. June 3, 2009 Agenda 1 Spurious Regressions

More information

NEW ESTIMATION METHODS FOR PANEL DATA MODELS. Valentin Verdier

NEW ESTIMATION METHODS FOR PANEL DATA MODELS. Valentin Verdier NEW ESTIMATION METHODS FOR PANEL DATA MODELS By Valentin Verdier A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Economics - Doctor of

More information

Econometrics of Panel Data

Econometrics of Panel Data Econometrics of Panel Data Jakub Mućk Meeting # 4 Jakub Mućk Econometrics of Panel Data Meeting # 4 1 / 30 Outline 1 Two-way Error Component Model Fixed effects model Random effects model 2 Non-spherical

More information

Iterative Bias Correction Procedures Revisited: A Small Scale Monte Carlo Study

Iterative Bias Correction Procedures Revisited: A Small Scale Monte Carlo Study Discussion Paper: 2015/02 Iterative Bias Correction Procedures Revisited: A Small Scale Monte Carlo Study Artūras Juodis www.ase.uva.nl/uva-econometrics Amsterdam School of Economics Roetersstraat 11 1018

More information

Asymptotic Properties of Empirical Likelihood Estimator in Dynamic Panel Data Models

Asymptotic Properties of Empirical Likelihood Estimator in Dynamic Panel Data Models Asymptotic Properties of Empirical Likelihood Estimator in Dynamic Panel Data Models Günce Eryürük North Carolina State University February, 2009 Department of Economics, Box 80, North Carolina State University,

More information

New GMM Estimators for Dynamic Panel Data Models

New GMM Estimators for Dynamic Panel Data Models I: 2319-8753 ew GMM Estimators for Dynamic Panel Data Models Ahmed H. Youssef 1, Ahmed A. El-sheikh 2, Mohamed R. Abonazel 3 Professor, Department of tatistics, Faculty of cience, King Abdulaziz University,

More information

Economics 536 Lecture 7. Introduction to Specification Testing in Dynamic Econometric Models

Economics 536 Lecture 7. Introduction to Specification Testing in Dynamic Econometric Models University of Illinois Fall 2016 Department of Economics Roger Koenker Economics 536 Lecture 7 Introduction to Specification Testing in Dynamic Econometric Models In this lecture I want to briefly describe

More information

Econ 582 Fixed Effects Estimation of Panel Data

Econ 582 Fixed Effects Estimation of Panel Data Econ 582 Fixed Effects Estimation of Panel Data Eric Zivot May 28, 2012 Panel Data Framework = x 0 β + = 1 (individuals); =1 (time periods) y 1 = X β ( ) ( 1) + ε Main question: Is x uncorrelated with?

More information

Econometrics. Week 6. Fall Institute of Economic Studies Faculty of Social Sciences Charles University in Prague

Econometrics. Week 6. Fall Institute of Economic Studies Faculty of Social Sciences Charles University in Prague Econometrics Week 6 Institute of Economic Studies Faculty of Social Sciences Charles University in Prague Fall 2012 1 / 21 Recommended Reading For the today Advanced Panel Data Methods. Chapter 14 (pp.

More information

TOURISM DEMAND IN AUSTRIAN SKI DESTINATIONS

TOURISM DEMAND IN AUSTRIAN SKI DESTINATIONS ISSN 2074-9317 The Economics of Weather and Climate Risks Working Paper Series Working Paper No. 5/2009 TOURISM DEMAND IN AUSTRIAN SKI DESTINATIONS A DYNAMIC PANEL DATA APPROACH Franz Eigner, 1 Christoph

More information

Final Exam. Economics 835: Econometrics. Fall 2010

Final Exam. Economics 835: Econometrics. Fall 2010 Final Exam Economics 835: Econometrics Fall 2010 Please answer the question I ask - no more and no less - and remember that the correct answer is often short and simple. 1 Some short questions a) For each

More information

xtdpdml for Estimating Dynamic Panel Models

xtdpdml for Estimating Dynamic Panel Models xtdpdml for Estimating Dynamic Panel Models Enrique Moral-Benito Paul Allison Richard Williams Banco de España University of Pennsylvania University of Notre Dame Reunión Española de Usuarios de Stata

More information

Subset-Continuous-Updating GMM Estimators for Dynamic Panel Data Models

Subset-Continuous-Updating GMM Estimators for Dynamic Panel Data Models econometrics Article Subset-Continuous-Updating GMM Estimators for Dynamic Panel Data Models Richard A Ashley 1 and Xiaojin Sun 2, * 1 Department of Economics, Virginia Tech, Blacksburg, VA 24060, USA;

More information

Consistent estimation of dynamic panel data models with time-varying individual effects

Consistent estimation of dynamic panel data models with time-varying individual effects ANNALES D ÉCONOMIE ET DE STATISTIQUE. N 70 2003 Consistent estimation of dynamic panel data models with time-varying individual effects Céline NAUGES, Alban THOMAS 1 ABSTRACT. This paper proposes a new

More information

Econometrics of Panel Data

Econometrics of Panel Data Econometrics of Panel Data Jakub Mućk Meeting # 9 Jakub Mućk Econometrics of Panel Data Meeting # 9 1 / 22 Outline 1 Time series analysis Stationarity Unit Root Tests for Nonstationarity 2 Panel Unit Root

More information

Specification Test for Instrumental Variables Regression with Many Instruments

Specification Test for Instrumental Variables Regression with Many Instruments Specification Test for Instrumental Variables Regression with Many Instruments Yoonseok Lee and Ryo Okui April 009 Preliminary; comments are welcome Abstract This paper considers specification testing

More information

The regression model with one stochastic regressor (part II)

The regression model with one stochastic regressor (part II) The regression model with one stochastic regressor (part II) 3150/4150 Lecture 7 Ragnar Nymoen 6 Feb 2012 We will finish Lecture topic 4: The regression model with stochastic regressor We will first look

More information

When Should We Use Linear Fixed Effects Regression Models for Causal Inference with Longitudinal Data?

When Should We Use Linear Fixed Effects Regression Models for Causal Inference with Longitudinal Data? When Should We Use Linear Fixed Effects Regression Models for Causal Inference with Longitudinal Data? Kosuke Imai Department of Politics Center for Statistics and Machine Learning Princeton University

More information

Least Squares Estimation of a Panel Data Model with Multifactor Error Structure and Endogenous Covariates

Least Squares Estimation of a Panel Data Model with Multifactor Error Structure and Endogenous Covariates Least Squares Estimation of a Panel Data Model with Multifactor Error Structure and Endogenous Covariates Matthew Harding and Carlos Lamarche January 12, 2011 Abstract We propose a method for estimating

More information

Specification testing in panel data models estimated by fixed effects with instrumental variables

Specification testing in panel data models estimated by fixed effects with instrumental variables Specification testing in panel data models estimated by fixed effects wh instrumental variables Carrie Falls Department of Economics Michigan State Universy Abstract I show that a handful of the regressions

More information

Linear Regression with Time Series Data

Linear Regression with Time Series Data u n i v e r s i t y o f c o p e n h a g e n d e p a r t m e n t o f e c o n o m i c s Econometrics II Linear Regression with Time Series Data Morten Nyboe Tabor u n i v e r s i t y o f c o p e n h a g

More information

ECON Introductory Econometrics. Lecture 16: Instrumental variables

ECON Introductory Econometrics. Lecture 16: Instrumental variables ECON4150 - Introductory Econometrics Lecture 16: Instrumental variables Monique de Haan (moniqued@econ.uio.no) Stock and Watson Chapter 12 Lecture outline 2 OLS assumptions and when they are violated Instrumental

More information

Nearly Unbiased Estimation in Dynamic Panel Data Models

Nearly Unbiased Estimation in Dynamic Panel Data Models TI 00-008/ Tinbergen Institute Discussion Paper Nearly Unbiased Estimation in Dynamic Panel Data Models Martin A. Carree Department of General Economics, Faculty of Economics, Erasmus University Rotterdam,

More information

Missing dependent variables in panel data models

Missing dependent variables in panel data models Missing dependent variables in panel data models Jason Abrevaya Abstract This paper considers estimation of a fixed-effects model in which the dependent variable may be missing. For cross-sectional units

More information

Outline. Nature of the Problem. Nature of the Problem. Basic Econometrics in Transportation. Autocorrelation

Outline. Nature of the Problem. Nature of the Problem. Basic Econometrics in Transportation. Autocorrelation 1/30 Outline Basic Econometrics in Transportation Autocorrelation Amir Samimi What is the nature of autocorrelation? What are the theoretical and practical consequences of autocorrelation? Since the assumption

More information

Panel Data Exercises Manuel Arellano. Using panel data, a researcher considers the estimation of the following system:

Panel Data Exercises Manuel Arellano. Using panel data, a researcher considers the estimation of the following system: Panel Data Exercises Manuel Arellano Exercise 1 Using panel data, a researcher considers the estimation of the following system: y 1t = α 1 + βx 1t + v 1t. (t =1,..., T ) y Nt = α N + βx Nt + v Nt where

More information

Asymptotic distributions of the quadratic GMM estimator in linear dynamic panel data models

Asymptotic distributions of the quadratic GMM estimator in linear dynamic panel data models Asymptotic distributions of the quadratic GMM estimator in linear dynamic panel data models By Tue Gørgens Chirok Han Sen Xue ANU Working Papers in Economics and Econometrics # 635 May 2016 JEL: C230 ISBN:

More information

Fixed Effects Models for Panel Data. December 1, 2014

Fixed Effects Models for Panel Data. December 1, 2014 Fixed Effects Models for Panel Data December 1, 2014 Notation Use the same setup as before, with the linear model Y it = X it β + c i + ɛ it (1) where X it is a 1 K + 1 vector of independent variables.

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

Linear Regression with Time Series Data

Linear Regression with Time Series Data Econometrics 2 Linear Regression with Time Series Data Heino Bohn Nielsen 1of21 Outline (1) The linear regression model, identification and estimation. (2) Assumptions and results: (a) Consistency. (b)

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

Robust Standard Errors in Transformed Likelihood Estimation of Dynamic Panel Data Models

Robust Standard Errors in Transformed Likelihood Estimation of Dynamic Panel Data Models D I S C U S S I O N P A P E R S E R I E S IZA DP No. 6583 Robust Standard Errors in Transformed Likelihood Estimation of Dynamic Panel Data Models Kazuhiko Hayakawa M. Hashem Pesaran May 0 Forschungsinstitut

More information

Instrumental variables estimation using heteroskedasticity-based instruments

Instrumental variables estimation using heteroskedasticity-based instruments Instrumental variables estimation using heteroskedasticity-based instruments Christopher F Baum, Arthur Lewbel, Mark E Schaffer, Oleksandr Talavera Boston College/DIW Berlin, Boston College, Heriot Watt

More information

Robust Standard Errors in Transformed Likelihood Estimation of Dynamic Panel Data Models with Cross-Sectional Heteroskedasticity

Robust Standard Errors in Transformed Likelihood Estimation of Dynamic Panel Data Models with Cross-Sectional Heteroskedasticity Robust Standard Errors in ransformed Likelihood Estimation of Dynamic Panel Data Models with Cross-Sectional Heteroskedasticity Kazuhiko Hayakawa Hiroshima University M. Hashem Pesaran USC Dornsife IE

More information

Robust Standard Errors in Transformed Likelihood Estimation of Dynamic Panel Data Models with Cross-Sectional Heteroskedasticity

Robust Standard Errors in Transformed Likelihood Estimation of Dynamic Panel Data Models with Cross-Sectional Heteroskedasticity Robust Standard Errors in ransformed Likelihood Estimation of Dynamic Panel Data Models with Cross-Sectional Heteroskedasticity Kazuhiko Hayakawa Hiroshima University M. Hashem Pesaran University of Southern

More information

The Time Series and Cross-Section Asymptotics of Empirical Likelihood Estimator in Dynamic Panel Data Models

The Time Series and Cross-Section Asymptotics of Empirical Likelihood Estimator in Dynamic Panel Data Models The Time Series and Cross-Section Asymptotics of Empirical Likelihood Estimator in Dynamic Panel Data Models Günce Eryürük August 200 Centro de Investigación Económica, Instituto Tecnológico Autónomo de

More information

Simultaneous Equation Models Learning Objectives Introduction Introduction (2) Introduction (3) Solving the Model structural equations

Simultaneous Equation Models Learning Objectives Introduction Introduction (2) Introduction (3) Solving the Model structural equations Simultaneous Equation Models. Introduction: basic definitions 2. Consequences of ignoring simultaneity 3. The identification problem 4. Estimation of simultaneous equation models 5. Example: IS LM model

More information

A Practical Test for Strict Exogeneity in Linear Panel Data Models with Fixed Effects

A Practical Test for Strict Exogeneity in Linear Panel Data Models with Fixed Effects A Practical Test for Strict Exogeneity in Linear Panel Data Models with Fixed Effects Liangjun Su, Yonghui Zhang,JieWei School of Economics, Singapore Management University School of Economics, Renmin

More information

Bootstrapping Heteroskedasticity Consistent Covariance Matrix Estimator

Bootstrapping Heteroskedasticity Consistent Covariance Matrix Estimator Bootstrapping Heteroskedasticity Consistent Covariance Matrix Estimator by Emmanuel Flachaire Eurequa, University Paris I Panthéon-Sorbonne December 2001 Abstract Recent results of Cribari-Neto and Zarkos

More information

Econometrics Homework 4 Solutions

Econometrics Homework 4 Solutions Econometrics Homework 4 Solutions Question 1 (a) General sources of problem: measurement error in regressors, omitted variables that are correlated to the regressors, and simultaneous equation (reverse

More information

LECTURE 10: MORE ON RANDOM PROCESSES

LECTURE 10: MORE ON RANDOM PROCESSES LECTURE 10: MORE ON RANDOM PROCESSES AND SERIAL CORRELATION 2 Classification of random processes (cont d) stationary vs. non-stationary processes stationary = distribution does not change over time more

More information

A better way to bootstrap pairs

A better way to bootstrap pairs A better way to bootstrap pairs Emmanuel Flachaire GREQAM - Université de la Méditerranée CORE - Université Catholique de Louvain April 999 Abstract In this paper we are interested in heteroskedastic regression

More information

When Should We Use Linear Fixed Effects Regression Models for Causal Inference with Panel Data?

When Should We Use Linear Fixed Effects Regression Models for Causal Inference with Panel Data? When Should We Use Linear Fixed Effects Regression Models for Causal Inference with Panel Data? Kosuke Imai Department of Politics Center for Statistics and Machine Learning Princeton University Joint

More information

Empirical Economic Research, Part II

Empirical Economic Research, Part II Based on the text book by Ramanathan: Introductory Econometrics Robert M. Kunst robert.kunst@univie.ac.at University of Vienna and Institute for Advanced Studies Vienna December 7, 2011 Outline Introduction

More information

ECONOMETRICS II (ECO 2401) Victor Aguirregabiria (March 2017) TOPIC 1: LINEAR PANEL DATA MODELS

ECONOMETRICS II (ECO 2401) Victor Aguirregabiria (March 2017) TOPIC 1: LINEAR PANEL DATA MODELS ECONOMETRICS II (ECO 2401) Victor Aguirregabiria (March 2017) TOPIC 1: LINEAR PANEL DATA MODELS 1. Panel Data 2. Static Panel Data Models 2.1. Assumptions 2.2. "Fixed e ects" and "Random e ects" 2.3. Estimation

More information

Lecture 8 Panel Data

Lecture 8 Panel Data Lecture 8 Panel Data Economics 8379 George Washington University Instructor: Prof. Ben Williams Introduction This lecture will discuss some common panel data methods and problems. Random effects vs. fixed

More information

Wooldridge, Introductory Econometrics, 4th ed. Chapter 15: Instrumental variables and two stage least squares

Wooldridge, Introductory Econometrics, 4th ed. Chapter 15: Instrumental variables and two stage least squares Wooldridge, Introductory Econometrics, 4th ed. Chapter 15: Instrumental variables and two stage least squares Many economic models involve endogeneity: that is, a theoretical relationship does not fit

More information