behavor of a cross-sectonal data over tme the ncreasng use of anel data, one new research area s examnng the roertes of non-statonary tme-seres data n

Size: px
Start display at page:

Download "behavor of a cross-sectonal data over tme the ncreasng use of anel data, one new research area s examnng the roertes of non-statonary tme-seres data n"

Transcription

1 On the Estmaton Inference of a Contegrated Regresson n Panel Data Chhwa Kao Mn-Hsen Chang Syracuse Unversty Frst Draft: May, 995 Frst Revsed Draft: January 997 Ths Draft: February 7, 997 Abstract In ths aer, we study the asymtotc dstrbutons for least-squares (OLS), fully moded (FM), dynamc OLS (DOLS) estmators n contegrated regresson models n anel data. We show that the OLS, FM, DOLS estmators are all asymtotcally normally dstrbuted. However, the asymtotc dstrbuton of the OLS estmator s shown to have a non-zero mean. Monte Carlo results examne the samlng behavor of the roosed estmators show that () the OLS estmator has a non-neglgble bas n nte samles, () the FM estmator does not mrove over the OLS estmator n general, () the DOLS out-erforms both the OLS FM estmators. Key Words: Panel Data, OLS Estmator; FM Estmator, DOLS Estmator, Heterogeneous Panels. Introducton Evaluatng the statstcal roertes of data along the tme dmenson has roven to be very derent from analyss of the cross-secton dmenson. As economsts have ganed access to better data wth more observatons across tme, understng these roertes has grown ncreasngly mortant. An area of artcular concern n tme seres econometrcs has been the use of non-statonary data. Wth the desre to study the We are thankful to three anonymous referees Peter Phlls for ontng out several techncal errors n an earler verson rovdng comments that led to mrovements of ths aer. We also thank Peter Pedron for hs helful comments Bangtan Chen for hs research assstance on an earler draft of ths aer. An electronc verson of the aer n ostscrt format can be retreved from htt://web.syr.edu/~cdkao. Address corresondence to: Chhwa Kao, Center for Polcy Research, 46 Eggers, Syracuse Unversty, Syracuse, Y 44; e-mal: cdkao@maxwell.syr.edu.

2 behavor of a cross-sectonal data over tme the ncreasng use of anel data, one new research area s examnng the roertes of non-statonary tme-seres data n anel form. It s an ntrgung queston to ask: how exactly does ths hybrd style of data combne the statstcal elements of tradtonal cross-sectonal analyss tme-seres analyss? In artcular, what s the correct way to analyze non-statonarty, the surous regresson roblem, contegraton n anel data? Gven the mmense nterest n testng for unt roots contegraton n tme-seres data, not much attenton has been ad to testng the unt roots n anel data. The only theoretcal studes as far as we know n ths area are Bretung Meyer (994), Quah (994), Levn Ln (99), Im, Pesaran, Shn (995), Maddala Wu (996). Bretung Meyer (994) derved the asymtotc normalty of the Dckey- Fuller test statstc for anel data wth a large cross-secton dmenson a small tme-seres dmenson. Quah (994) studed a unt root test for anel data that have smultaneously extensve cross-secton tme-seres varaton. He showed that the asymtotc dstrbuton for the roosed test s a mxture of the stard normal Dckey-Fuller-Phlls asymtotcs. Levn Ln (99) derved the asymtotc dstrbutons for unt roots on anel data showed that the ower of these tests ncreases dramatcally as the cross-secton dmenson ncreases. Im et al. (995) crtqued the Levn Ln anel unt root statstcs roosed alternatves. Maddala Wu (996) rovded a comarson of the tests of Im et al. (995) Levn Ln (99). They suggested a new test based on the Fsher test. However, to ths date, lttle s known about contegraton tests estmaton wth regresson models n anel data. Excetons are Kao (996), McCoskey Kao (996), Pedron (995, 996). In the rst half of Kao (996), he studed a surous regresson n anel data. Asymtotc roertes of the least-squares (OLS) estmator other conventonal statstcs were examned. Kao (996) showed that the OLS estmator s consstent for ts true value; but the t-statstc dverges so that nferences about the regresson coecent, ; are wrong wth robablty that goes to one as! T!:Furthermore, Kao (996) examned the Dckey-Fuller (DF) the augmented Dckey-Fuller (ADF) tests to test the null hyothess of no contegraton n anel data. McCoskey Kao (996) roosed further tests for the null hyothess of contegraton n anel data. Pedron (995) derved asymtotc dstrbutons for resdual based tests of contegraton for both homogenous heterogenous anels. Pedron (996) roosed a fully moded estmator for heterogenous anels. On the other h, Park Ogak (99) derved asymtotc dstrbutons for contegraton coecent estmators related t-statstcs for anel data usng CCR transformatons. Although they used an SUR aroach rather than dmenson asymtotcs, many of the ssues they dealt wth are smlar. In ths aer, we study the lmtng dstrbutons for the ordnary least squares (OLS), fully moded (FM), Dynamc OLS (DOLS) estmators n anel contegrated regresson models.

3 Secton ntroduces the model assumtons. Secton develos the asymtotc theory for OLS, FM DOLS estmators. Secton 4 gves the lmtng dstrbutons of FM DOLS estmators for heterogeneous anels. Secton 5 develos the lmtng dstrbutons of the Wald statstcs. Secton 6 resents some Monte Carlo results to evaluate the nte samle roertes of the OLS, FM, dynamc OLS estmators. Secton 7 summarzes the ndngs. All roofs are n the Aendx. R R Aword on notaton. We wrte the ntegral W (s)ds as W when there s no ambguty over lmts. 0 We dene to be any matrx such that n 0 : We use kak to denote tr A Ao 0 ; jaj to denote the determnant of A, d! to denote convergence n dstrbuton,! to denote convergence n robablty, [x] to denote the largest nteger x, I(0) I() to sgnfy a tme seres that s ntegrated of order zero one, resectvely, BM () to denote Brownan moton wth covarance matrx. The Model Assumtons Consder the followng xed eect anel regresson: y t + x 0 t + u t; ; :::; ; t ; :::T; () where fy t g are ; s a k vector of the sloe arameters, f g are the ntercets, fu t g are the statonary dsturbance terms. We assume that fx t g are k ntegrated rocesses of order one for all ; where x t x t + t : Under these seccatons, () descrbes a system of contegrated regressons,.e., y t s contegrated wth x t : The ntalzaton of ths system s y 0 x 0 0for all. Assumton fy t; x t g are ndeendent across. Assumton The cross-secton dmenson s a monotonc functon of the tme-seres dmenson,.e., (T ), so that the law of large numbers (Theorem 6., Bllngsley, 986,. 8) the central lmt theorem (Theorem 7., Bllngsley, 986,. 69) for trangular arrays can be aled. ext, we characterze the nnovaton vector w t u t ; t 0 0. We assume that w t s a lnear rocess that satses the followng assumton. Assumton (e.g., Phlls, 995) (a) w t (L) t P j0 j t j ; P j0 ja k j k < ; j()j 60for some a>:

4 (b) t s..d. wth zero mean, varance matrx ; nte fourth order cumulants. Assumton mles that (e.g., Phlls Solo, 99) the artal sum rocess T P [Tr] t w t satses the followng multvarate nvarance rncle: where X [Tr] d w t! B (r) BM () as T!; () T t B The long-run covarance matrx of fw t g s gven by 4 B u B X j 5 : E w j w 0 0 () () 0 where X j u u u 5 ; E w j w u u u E w 0 w u u u 5 () 5 (4) are arttoned conformably wth w t : Assumton 4 s non-sngular,.e., fx t g are not contegrated. Dene u: u u u: (5) Then, B can be rewrtten as B 4 B u B 5 4 u: u V W 5 ; 4

5 where 4 V W 5 BM (I) s a stardzed Brownan moton. Dene the one-sded long-run covarance wth + X j0 E w j w u u u 5 : Remark Here we assume that anels are homogeneous,.e., the varances are constant across the crosssecton unts. We wll relax ths assumton n Secton 4 to allow for derent varances for derent. OLS, Fully Moded, Dynamc OLS Estmators Let us rst study the lmtng dstrbuton of the OLS estmator for equaton (). The OLS estmator of s It follows that b OLS X t TX X (x t x )(x t x ) 0 TX t T bols h P P T h P T (x t t x )(x t x ) 0 h P h P T T [ T ] T ; (x t x )(y t y ) P T T (x t t : (6) x ) u t where x T P T t x t; y T P T t y t; T T P T t (x t x ) u t ; T T P T t (x t x )(x t x ) 0, P T P T, T T : Before gong nto the next theorem, we need to consder some relmnary results. All lmts n (a) (d) n Lemma are taken as T!:Also, all lmts n (e) (f ) n Lemma Theorems 4 are taken as! T!: Lemma If Assumtons (a) T d! (b) T d! (c) E [ ] u + u ; 4 hold, then R fw dv u: + R fw f W 0 ; R fw dw 0 u + u ; 5

6 (d) E [ ] 6 ; (e) T! u + u ; (f) T! 6 ; where f W W R W : Remark u s due to the endogenety of the regressor x t ; u s due to the seral correlaton. Thus, we have establshed the followng theorem: Theorem If Assumtons 4 hold, then (a) T bols! u +6 u; (b) T bols T d! 0;6 u: ; where T X T t TX (x t x t )(x t x ) 0 X fw dw 0 u + u : Remark We notce that T! u +6 u: Remark 4 The normalty of the OLS estmator comes naturally. When summng across, the nonstard asymtotc dstrbuton due to unt root n the tme dmenson s smoothed out. However, t s mortant to note that the OLS estmator s asymtotcally based. The asymtotc bas s b OLS T T u +6 u T whch decreases as T ncreases. Remark 5 u: can be seen as the long-run sgnal-to-nose rato. 0 Remark 6 If w t u t ; 0 t are..d., then whch was examned by Kao Chen (995). T! u 6

7 Chen, McCoskey, Kao (996) nvestgated the nte samle roretes of the OLS estmator n (6), the t-statstc, the bas-corrected OLS estmator, the bas-corrected t-statstc. They found that the bas-corrected OLS estmator does not mrove over the OLS estmator n general. The results of Chen, McCoskey, Kao (996) suggest that alternatves, such as the FM estmator or DOLS estmator (e.g., Sakkonen, 99; Stock Watson, 99) may be more romsng n contegrated anel regressons. Thus, we begn our study by examnng the lmtng dstrbuton of FM estmator, b FM : Followng Pedron (996), we begn our study by examnng the lmtng dstrbuton of the FM estmator, b FM. In contrast to Pedron (996), we ntally consder the case where s common across members of the anel n order to focus on the role that the sgnal to nose rato, u:, can lay n the asymtotc dstrbuton of an FM estmator. The FM estmator s constructed by makng correctons for endogenety seral correlaton to the OLS estmator b OLS n (6). Let b u b are consstent estmates of u : Dene ote that 4 u+ t t whch has the long-run covarance matrx u + t u t u t ; bu + t u t b u b t; y + t y t u t; by + t y t b u b t : 5 4 u 0 Ik 4 u: u t t where Ik s a k k dentty matrx. The endogenety correcton s acheved by modfyng the varable y t n () wth the transformaton by + t y t b u b t The seral correlaton correcton term has the form 5 ; 5 ; + x 0 t + u t b u b t: b + u bu b b u b b b u ; 7 b b u A

8 where u b b are kernel estmates of u : Therefore, the FM estmator s X TX X! TX b FM (x t x )(x t x ) 0 (x t x ) by t + T b + u t t ow, we state the lmtng dstrbuton of b FM : : (7) Theorem If Assumtons 4 hold, then T bfm d! 0;6 u: : Remark 7 ote that Pedron (996) allowed the drfts for the ntegrated regressors n hs contegrated system. Ths aer only consders the regresson n whch ntegrated regressors do not have drfts. Also we roose the FM estmators for multle regresson. ext, we roose a DOLS estmator, b D ; whch uses the ast future values of 4x t as addtonal regressors. We then show that the lmtng dstrbuton of b D s the same as the FM estmator, b FM : But rst, we need the followng addtonal assumton: Assumton 5 The rocess fu t g can be rojected ontof t g to get where u t X j X j c j t+j + v t ; (8) kc j k < ; fv t g s statonary wth zero mean, fv t g f t g are uncorrelated not only contemoraneously but also n all lags leads. Remark 8 Assumton 5 can be guaranteed by followng the condtons n Sakkonen (99,. ). Remark 9 In ractce, the leads lags may be truncated whle retanng Assumton 5 aroxmately, so that u t q X j q c j t+j + v t : Ths s because fc j g are assumed tobe absolutely summable,.e., P j kc jk < : We also need to requre that q q tend to nnty wth T at a sutable rate,.e., Assumton 6 q T! 0; q T! 0; T X jjj>q or q kc j k!0: (9) 8

9 We then substtute (8) nto () to have y t + x 0 t + q X j q c j t+j + v t : Therefore, we obtan the DOLS of ; b D ; by runnng the followng regresson: y t + x 0 t + q X j q c j 4x t+j + v t : (0) ext, we show that b D has the same lmtng dstrbuton b FM as n Theorem. Theorem If Assumtons 6 hold, then T bd d! 0;6 u: : 4 Heterogeneous Panels The aer so far assumes that the anel data are homogeneous. The substantal heterogenety exhbted by actual data n the cross-sectonal dmenson severely restrcts the ractcal alcablty of such estmators. Also, the estmators n Sectons are not easly extended to cases of broader cross-sectonal heterogenety snce the varances bases are seced n terms of the asymtotc covarance arameters that are assumed to be shared cross sectonally. Recently, Pedron (996) roosed an FM estmator for heterogeneous anels. Pedron (996) roosed the followng anel FM estmator (usng hs notatons):! X TX b T bl b TX L (x t x ) u t Tb ; () bl t (x t x )! X where b L s the lower trangular decomoston of a consstent estmator of the asymtotc covarance matrx t ; where u t s gven by u t u t bl bl 4 x t the seral correlaton adjustment arameter b s gven by b b + b 0 bl b + 0 : bl Pedron (996) then derved the followng result (hs Prooston.): T b T! (0;v); 9

10 where v 8 < : x y 0 6 else In ths secton, we roose an alternatve reresentaton of the anel FM estmator for heterogeneous anels. Agan, n contract to Pedron (996), ths secton only consders the regresson that ntegrated regressors do not have drfts. Also we roose an FM estmator for multle regresson. Before we dscuss the FM estmator we need the followng assumtons: Assumton 7 We assume the anels are heterogeneous,.e., ; : are vared for derent : We also assume the nvarance rncle n (), (8) n Assumton 5, (9) n Assumton 6 stll hold. Let x t b x t ; () u t b u: bu + t ; () y t b u: by + t ; (4) where b b u: are consstent estmators of u: ; resectvely. Assumton 8 b s not sngular for all. where Then, we dene the FM estmator for heterogeneous anels as b FM X b FM can be wrtten as T b h h FM P TX X (x t x )(x t x )0 t b + u bu b b u b b b u :! TX (x t x ) y t T b + u t P T T t (x t x )(x t x h P )0 P 4T h P T [ 4T ] T ; b b u A ; (5) P T T t (x t x ) u t T b + u 0

11 where x T P T t x t ; T T P T t (x t x ) u t T b + u ; 4T T P T t (x t x )(x t x )0, T P P T, 4T 4T : It s clear that from Lemma that T d! fw dv ; 4T d! fw f W 0; It follows that 4T! 6 I k; T! 6 I k: T b FM d! (0; 6I k) : Hence, we have establshed the followng theorem: Theorem 4 If Assumtons 7 8 hold, then T b FM d! (0; 6I k) : The DOLS estmator for heterogeneous anels, b D; can be obtaned by runnng the followng regresson: yt + x 0 t + q X j q c j 4x t+j + v t : It s straghtforward to show that b D also has the same lmtng dstrbuton as b FM: Theorem 5 If Assumtons 7 8 hold, then T b D d! (0; 6I k) : Remark 0 Theorems 4 5 show that the lmtng dstrbutons of b FM b D are free of nusance arameters. 5 Hyothess Testng We now consder a lnear hyothess that nvolves the elements of the coecent vector : We show that hyothess tests constructed usng the FM DOLS estmators have asymtotc ch-squared dstrbutons. The null hyothess has the form: H 0 : R r; (6)

12 where r s a m known vector R s a known m k matrx descrbng the restrctons. A natural test statstc of the Wald test usng FM b or D b for homogeneous anels s W Rb 6 T FM r 0h Rb b u: R Rb 0 FM r : (7) For the heterogeneous anels, a natural statstc usng bfm or bd to test the null hyothess s W Rb 0h 6 T FM r RR Rb 0 FM r : (8) It s clear that W W converge n dstrbuton to a ch-squared rom varable wth k degrees of freedom, k ; as! T!under the null hyothess. Hence, we establsh the followng theorem: Theorem 6 If Assumtons 8 hold, then under the null hyothess (6), (a) W d! k ; (b) W d! k : Remark Because the FM the DOLS estmators have the same asymtotc dstrbuton, t s easy to verfy that the Wald statstcs based on the FM estmator share the same lmtng dstrbutons wth those based on the DOLS estmator. 6 Monte Carlo Smulatons To comare the erformance of OLS, FM, DOLS estmators, we conducted Monte Carlo exerments based on the desgn smlar to Phlls Hansen (990) Phlls Loretan (99). The data generatng rocess (DGP) was y t + x t + u t x t x t + t for ; :::; ; t ; :::T; where u t t wth u t 0 u t t t A d 0 A : 0:4 0:6 0 A 5 ; u t t 5 A : A

13 We generated from a unform dstrbuton, U [0; 0]; set. We allowed to vary consdered values of f0:8; 0:4; 0:0; 0:8g for f 0:8; 0:4; 0:4g for : Rom numbers for (u t ; t ) were generated by the GAUSS rocedure RDS. At each relcaton, we generated (T + 000) length of rom numbers then slt t nto seres so that each seres had the same mean varance. The rst ; 000 observatons were dscarded for each seres. fu t g f t g were constructed wth u : Once the estmates of w t ; bw t were estmated, we used to estmate : was estmated by ( X b T TX t b T bw t bw 0 t + T X TX t lx $ l T X bw t bw 0 t (9) t+ bwt bw 0 t + bw t bw 0 t ) ; (0) where $ l saweght functon or a kernel: Usng Phlls Durlauf (986) the law of large numbers for trangular arrays, b b can be shown to be consstent for : The estmate of the long-run covarance matrx n (0) was obtaned by usng the rocedure KEREL n COIT :0 wth a Bartlett wndow of lag length ve. Results wth other kernels, such as Parzen QS kernels, are not reorted, because no essental derences were found for most cases. ext, we recorded the results from our Monte Carlo exerments that examned the nte-samle roertes of the OLS estmator, b OLS ; the FM estmator, b FM ; the DOLS estmator, b D : The smulatons were erformed by a Sun SarcServer ; 000. GAUSS :: COIT :0 were used to erform the smulatons. The results we reort are based on 0; 000 relcatons are summarzed n Tables. The FM estmator was obtaned by usng a Bartlett wndow of lag length ve as n (0). Four lags two leads were used for the DOLS estmator. Table reorts the Monte Carlo means stard devatons (n arentheses) of bfm ; bd bols ; for samle szes T ( )(0;40; 60) : The bases of the OLS estmator, b OLS ; decrease at a rate of T. For examle, wth 0:8 0:8;the bas at T 0s 0:0 at T 40s 0:04. Also, the bases ncrease n (f > 0) decrease n : In general, the FM estmator, b FM ; resents the same degree of dculty wth bas as does the OLS estmator, b OLS : For examle, whle the FM estmator, b FM ; reduces the bas substantally outerforms b OLS when > 0 < 0; the ooste s true when > 0 > 0; the ooste s true. Lkewse, when 0:8; b FM s less based than b OLS for values of 0:8: Yet, for values of 0:4; the bas n b OLS s less than the bas n b FM : There seems to be lttle to choose between b OLS b FM when

14 < 0: Ths s robably due to the falure of the semarametrc correcton rocedure n the resence of a negatve seral correlaton of the errors,.e., a negatve MAvalue, < 0: Fnally, for the cases where 0:0,b FM outerforms b OLS when < 0: On the other h, b FM s more based than b OLS when > 0: In contrast, the results n Table show that the DOLS, D b ; s dstnctly sueror to the OLS FM estmators for all cases n terms of the mean bases. Clearly, the DOLS out-erformed both the OLS FM estmators. Whle the lmtng theory deends on the assumton that the cross-secton tme-seres dmensons are comarable n magntude, the actual anel data have a wde varety of cross-secton tme-seres dmensons. It s mortant to know the eects of the varatons n anel dmensons. Table consders 0 derent settngs for T, each rangng from 0 to 0. Frst, we notce that the cross-secton dmenson has sgncant eect on the bases of OLS b ; FM b ; D b when s ncreased from to 0. However, when s ncreased from 0 to 40 on, there s lttle eect on the bases of OLS b ; FM b ; D b : The results n Table also conrm the suerorty of the DOLS. 7 Concluson Ths aer derves lmtng dstrbutons for the OLS, FM, DOLS estmators n a contegrated regresson shows they are asymtotcally normal. We also nvestgated the nte samle roretes of the OLS, FM, DOLS estmators. Our ndngs are summarzed as follows:. The OLS estmator has a non-neglgble bas n nte samles.. The FM estmator does not mrove over the OLS estmator n general.. The DOLS estmator may be more romsng than OLS or FM estmators n estmatng the contegrated anel regressons. Aendx A Proof of Lemma Proof. (a) (b) are taken from the lterature,.e., 4

15 where P T T T (x t t x )(x t x ) 0 d R! fww f 0 ; fw W T T P T t (x t x ) u t d! ; R fw dv u: + W ; R fw dw 0 u + u () where 4 u u + u : The row, column dagonal element of R f W f W 0 element s W W j s R W W R W, the row, column j o-dagonal W j : Usng E W W 6 Var W W 45 ; the corresondng generalzaton s E fww f 0 6 I k; where IK s a k k dentty matrx. It then follows that E [ ] 6 Ik () 6 () establshng (d): To rove (c); we use the fact that E 0 hr fw dv E hr fw dw 0 I k 5

16 resultng n E [ ] u + u : Fnally, T X T T! X T! u + u as requred for (e) (f ) by the law of large numbers for trangular arrays. 6 (4) B Proof of Var h R gw dv u: 6 u: Proof. ote that Var fw dv u: 0 E fw dv u: fw dv u: 0 fw dv fw dv u: E R usng E fw R 0 dv fw dv 6 I k: u: 6 I k 6 u: C Proof of Theorem Proof. (a) s mmedately evdent from Lemma. Recall that h T bols P h P h P T [ T ] T : ote that the sequence n P T T (x hp t t x )(x t x ) 0 n T T ( T ) P R fw dw 0 R fw dw 0 n T u + u o o u + u n R o fw dw 0 u + u R fw dw 0 u + u 6

17 s a trangular array sequence; thus, a central lmt theorem of a trangular array s needed. Recall that the varances are assumed to be the same across ;whch mles that the Lndeberg condton wll hold. We can readly see that the T wll converge to a normal varable wth an arorate normalzaton that T wll converge to 6 n robablty byalaw of large number for a trangular array. Frst, we note from Lemma that It follows that X T T fw dw 0 d u + u! fw dv u: : fw dw 0 u + u Var fw dv u: 6 u: from Aendx B. Usng the Slutsky theorem, we obtan! E fw dv u: 0 [ T ] T d! 0;6 u: : Hence, rovng (b), where T X TX T t Therefore, we establsh Theorem. T bols : d! 0;6 u: T (x t x )(x t x ) 0 X ; (5) fw dw 0 u + u : (6) D Proof of Theorem The FM estmator of can be rewrtten as follows b FM hp P T hp (x t t x )(x t x ) 0 P T (x t t x ) by t + T b + u hp P T hp + (x t t x )(x t x ) 0 P T (x t t x ) bu + t T b + u : (7) 7

18 Frst, we note that T bfm P where T T T t Let w t + where u + t h h P P T [ T ] h(x t x ) bu + t b + u 0 t 0 P T P T (x t t x )(x t x ) 0 h P T P T T t h(x t x ) bu b + t + u T ; (8) we have X [Tr] w t + T t + 4 B+ u Let B d! P ; T T. 4 B+ u B 4 u: I u + T T From Lemma the consstency of b u b + T 5 BM + as T!; (9) 0 I d! TX t 5 (x t x ) bu + t : 4 B u B we note that eb db u + ++ u ; 5 : where + u u u u: u A It follows n terms of stard Wener rocesses that + d T! fw dv u: ++ u : 8

19 ow let Clearly, from Lemma we know that T + T b + u : E [ T ]0 Var[ T ] 6 u: : A smlar argument to the roof of Theorem yelds as requred. T bfm d! 0;6 u: E Proof of Theorem Frst we wrte (0) n vector from: y e + x + q C + v x + D + v (say), where y s a T vector of y t ; e astunt vector, q s the T (q + q ) matrx of observatons on the q + q regressors 4x t q ; ;4x t+q ; x savector of T k of x t ; C s a (q + q ) vector of c j ; v s a T vector of v t ; s a T (q + q +)matrx, (e; q ); D s a (q + q +)vector of arameters. Let Q I ( 0 ) 0 : It follows that We then wrte where 5T P bd X T bd P [ 6T ] T (x 0 Q x ) X x 0 Q x (x 0 Q v ) P P 6T h 5T ; 5T ; 5T T (x0 Q v ) ; 6T 9 P P 5T : T (x0 Q v ) 6T ; 6T T (x 0 Q x ) :

20 Observe that from Sakkonen (99) 6T X X X X 6T T (x0 Q x ) x 0 T W T x + o () T TX q 0 (x t x )(x t x ) + o (); tq + 5T X X X X 5T T (x0 Q v ) x 0 T W T v + o () T TX q tq + (x t x ) v t + o (); T TX q tq + (x t x ) v t d! eb db u + ; where W T I T T ee 0 : Usng arguments smlar to those n Theorem, we establsh the followng: as requred. T bd d! 0;6 u: References [] Bllngsley, P. (986), Probablty Measure, John Wley, ew York. [] Bretung, J., Meyer, W. (994), Testng for Unt Roots n Panel Data: Are Wages on Derent Barganng Levels Contegrated? Aled Economcs, 6,

21 [] Chen, B., McCoskey, S., Kao, C. (996), Estmaton Inference of a Contegrated Regresson n Panel Data: A Monte Carlo Study, Manuscrt, Center for Polcy Research, Syracuse Unversty. [4] Im, K., Pesaran, H., Shn, Y. (995), Testng for Unt Roots n Heterogeneous Panels, Manuscrt, Unversty of Cambrdge. [5] Kao, C. (996), Surous Regresson Resdual-Based Tests for Contegraton n Panel Data when the Cross-Secton Tme-Seres Dmensons are Comarable, Manuscrt, Center for Polcy Research, Syracuse Unversty. [6] Kao, C., Chen, B. (995), On the Estmaton Inference for Contegraton n Panel Data when the Cross-Secton Tme-Seres Dmensons are Comarable, Manuscrt, Center for Polcy Research, Syracuse Unversty. [7] Levn, A., Ln, C.-F. (99), Unt Root Tests n Panel Data: ew Results, Dscusson Paer, Deartment of Economcs, UC-San Dego. [8] Maddala, G. S., Wu, S. (996), A Comaratve Study of Unt Root Tests wth Panel Data a ew Smle Test: Evdence From Smulatons the Bootstra, Manuscrt, Deartment of Economcs, Oho State Unversty.. [9] McCoskey, S., Kao, C. (996), Resduals-Based Tests of the ull of Contegraton n Panel Data, Manuscrt, Center for Polcy Research, Syracuse Unversty. [0] Park, J., Ogak, M. (99), Seemngly Unrelated Canoncal Contegratng Regressons, The Rochester Center for Economc Research, Workng Paer o. 80. [] Pedron, P. (995), Panel Contegraton: Asymtotcs Fnte Samle Proertes of Pooled Tme Seres Tests wth an Alcaton to the PPP Hyothess, Indana Unversty Workng Paers n Economcs, o [] Pedron, P. (996), Fully Moded OLS for Heterogeneous Contegrated Panels the Case of Purchasng Power Party, Indana Unversty Workng Paers n Economcs, o [] Phlls, P. C. B. (995), Fully Moded Least Squares Vector Autoregresson, Econometrca, 6, [4] Phlls, P. C. B., Durlauf, S.. (986), Multle Tme Seres Regresson wth Integrated Processes, Revew of Economc Studes, 5,

22 [5] Phlls, P. C. B., Hansen, B. E. (990), Statstcal Inference n Instrumental Varables Regresson wth I() Processes, Revew of Economc Studes, 57, [6] Phlls, P. C. B., Loretan, M. (99), Estmatng Long-Run Economc Equlbra, Revew of Economc Studes, 58, [7] Phlls, P. C. B., Solo, V. (99), Asymtotcs for lnear rocesses, Annals of Statstcs, 0, [8] Quah, D. (994), Exlotng Cross Secton Varaton for Unt Root Inference n Dynamc Data, Economcs Letters, 44, 9-9. [9] Sakkonen, P. (99), Asymtotcally Ecent Estmaton of Contegratng Regressons, Econometrc Theory, 58, -. [0] Stock, J., Watson, M. (99), A Smle Estmator of Contegratng Vectors n Hgher Order Integrated Systems, Econometrca, 6,

23 Table : Means Bases Stard Devatons of OLS, FM, DOLS Estmators 0:8 0:4 0:8 b OLS b FM b D b OLS b FM b D b OLS b FM b D 0:8 T (.049) (.047) (.040) (.0) (.05) (.0) (.0) (.06) (.009) T (.09) (.07) (.0) (.0) (.0) (.0) (.004) (.006) (.00) T (.00) (.009) (.007) (.007) (.007) (.006) (.00) (.00) (.00) 0:4 T (.08) (.06) (.07) (.00) (.09) (.0) (.0) (.08) (.0) T (.04) (.009) (.07) (.0) (.0) (.009) (.005) (.006) (.004) T (.007) (.005) (.005) (.006) (.006) (.005) (.00) (.00) (.00) 0:0 T (.07) (.05) (.07) (.06) (.0) (.06) (.06) (.09) (.07) T (.009) (.005) (.005) (.06) (.0) (.06) (.00) (.007) (.005) T (.005) (.00) (.00) (.005) (.008) (.008) (.00) (.004) (.00) 0:8.000 T (.06) (.0) (.008) (.07) (.05) (.04) (.04) (.07) (.0) T (.006) (.00) (.00) (.006) (.005) (.004) (.0) (.009) (.009) T (.00) (.00) (.00) (.00) (.00) (.00) (.007) (.005) (.005) ote: (a) T. (b) The lag length 5 of the Bartlett wndows s used for the FM estmator. (c) The 4 lags leads are used for the DOLS estmator.

24 Table : Means Bases Stard Devatons of OLS, FM, DOLS Estmators for Derent T (,T) OLS b FM b D b (,0) (.84) (.89) (.97) (,40) (.09) (.09) (.06) (,60) (.06) (.06) (.064) (,0) (.0) (.0) (.09) (0,0) (.00) (.09) (.0) (0,40) (.06) (.05) (.04) (0,60) (.00) (.009) (.009) (0,0) (.005) (.005) (.005) (40,0) (.0) (.0) (.0) (40,40) (.0) (.0) (.009) (40,60) (.007) (.007) (.007) (40,0) (.004) (.00) (.00) (60,0) (.07) (.07) (.08) (60,40) (.009) (.009) (.008) (60,60) (.006) (.006) (.005) (60,0) (.00) (.00) (.00) (0,0) (.0) (.0) (.0) (0,40) (.006) (.006) (.006) (0,60) (.004) (.004) (.004) (0,0) (.00) (.00) (.00) ote: (a) The lag length 5 of the Bartlett wndows s used for the FM estmator. (b) The 4 lags leads are used for the DOLS estmator. 4

Testing for seasonal unit roots in heterogeneous panels

Testing for seasonal unit roots in heterogeneous panels Testng for seasonal unt roots n heterogeneous panels Jesus Otero * Facultad de Economía Unversdad del Rosaro, Colomba Jeremy Smth Department of Economcs Unversty of arwck Monca Gulett Aston Busness School

More information

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4) I. Classcal Assumptons Econ7 Appled Econometrcs Topc 3: Classcal Model (Studenmund, Chapter 4) We have defned OLS and studed some algebrac propertes of OLS. In ths topc we wll study statstcal propertes

More information

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur Analyss of Varance and Desgn of Exerments-I MODULE III LECTURE - 2 EXPERIMENTAL DESIGN MODELS Dr. Shalabh Deartment of Mathematcs and Statstcs Indan Insttute of Technology Kanur 2 We consder the models

More information

ANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U)

ANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U) Econ 413 Exam 13 H ANSWERS Settet er nndelt 9 deloppgaver, A,B,C, som alle anbefales å telle lkt for å gøre det ltt lettere å stå. Svar er gtt . Unfortunately, there s a prntng error n the hnt of

More information

A Comparative Study for Estimation Parameters in Panel Data Model

A Comparative Study for Estimation Parameters in Panel Data Model A Comparatve Study for Estmaton Parameters n Panel Data Model Ahmed H. Youssef and Mohamed R. Abonazel hs paper examnes the panel data models when the regresson coeffcents are fxed random and mxed and

More information

Not-for-Publication Appendix to Optimal Asymptotic Least Aquares Estimation in a Singular Set-up

Not-for-Publication Appendix to Optimal Asymptotic Least Aquares Estimation in a Singular Set-up Not-for-Publcaton Aendx to Otmal Asymtotc Least Aquares Estmaton n a Sngular Set-u Antono Dez de los Ros Bank of Canada dezbankofcanada.ca December 214 A Proof of Proostons A.1 Proof of Prooston 1 Ts roof

More information

Generalized fixed-t Panel Unit Root Tests Allowing for Structural Breaks

Generalized fixed-t Panel Unit Root Tests Allowing for Structural Breaks ATHES UIVERSITY OF ECOOMICS AD BUSIESS DEPARTMET OF ECOOMICS WORKIG PAPER SERIES 08-0 Generalzed fxed-t Panel Unt Root Tests Allowng for Structural Breaks Yanns Karavas and Elas Tzavals 76 Patsson Str.,

More information

ECONOMICS 351*-A Mid-Term Exam -- Fall Term 2000 Page 1 of 13 pages. QUEEN'S UNIVERSITY AT KINGSTON Department of Economics

ECONOMICS 351*-A Mid-Term Exam -- Fall Term 2000 Page 1 of 13 pages. QUEEN'S UNIVERSITY AT KINGSTON Department of Economics ECOOMICS 35*-A Md-Term Exam -- Fall Term 000 Page of 3 pages QUEE'S UIVERSITY AT KIGSTO Department of Economcs ECOOMICS 35* - Secton A Introductory Econometrcs Fall Term 000 MID-TERM EAM ASWERS MG Abbott

More information

PARTIAL QUOTIENTS AND DISTRIBUTION OF SEQUENCES. Department of Mathematics University of California Riverside, CA

PARTIAL QUOTIENTS AND DISTRIBUTION OF SEQUENCES. Department of Mathematics University of California Riverside, CA PARTIAL QUOTIETS AD DISTRIBUTIO OF SEQUECES 1 Me-Chu Chang Deartment of Mathematcs Unversty of Calforna Rversde, CA 92521 mcc@math.ucr.edu Abstract. In ths aer we establsh average bounds on the artal quotents

More information

A Monte Carlo Study for Swamy s Estimate of Random Coefficient Panel Data Model

A Monte Carlo Study for Swamy s Estimate of Random Coefficient Panel Data Model A Monte Carlo Study for Swamy s Estmate of Random Coeffcent Panel Data Model Aman Mousa, Ahmed H. Youssef and Mohamed R. Abonazel Department of Appled Statstcs and Econometrcs, Instute of Statstcal Studes

More information

Econometrics of Panel Data

Econometrics of Panel Data Econometrcs of Panel Data Jakub Mućk Meetng # 8 Jakub Mućk Econometrcs of Panel Data Meetng # 8 1 / 17 Outlne 1 Heterogenety n the slope coeffcents 2 Seemngly Unrelated Regresson (SUR) 3 Swamy s random

More information

Testing Weak Cross-Sectional Dependence in Large Panels

Testing Weak Cross-Sectional Dependence in Large Panels estng Weak Cross-Sectonal Deendence n Large Panels M. Hashem Pesaran January CWPE 8 estng Weak Cross-Sectonal Deendence n Large Panels M. Hashem Pesaran Unversty of Cambrdge & Unversty of Southern Calforna

More information

PPP TESTS IN COINTEGRATED PANELS: EVIDENCE FROM ASIAN DEVELOPING COUNTRIES

PPP TESTS IN COINTEGRATED PANELS: EVIDENCE FROM ASIAN DEVELOPING COUNTRIES PPP TESTS IN COINTEGRATED PANELS: EVIDENCE FROM ASIAN DEVELOPING COUNTRIES Syed Abul Basher Department of Economcs York Unversty Toronto, ON M3J 1P3 basher@yorku.ca and Mohammed Mohsn * Department of Economcs

More information

Statistics for Economics & Business

Statistics for Economics & Business Statstcs for Economcs & Busness Smple Lnear Regresson Learnng Objectves In ths chapter, you learn: How to use regresson analyss to predct the value of a dependent varable based on an ndependent varable

More information

Digital PI Controller Equations

Digital PI Controller Equations Ver. 4, 9 th March 7 Dgtal PI Controller Equatons Probably the most common tye of controller n ndustral ower electroncs s the PI (Proortonal - Integral) controller. In feld orented motor control, PI controllers

More information

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur Analyss of Varance and Desgn of Exerments-I MODULE II LECTURE - GENERAL LINEAR HYPOTHESIS AND ANALYSIS OF VARIANCE Dr. Shalabh Deartment of Mathematcs and Statstcs Indan Insttute of Technology Kanur 3.

More information

Comparing two Quantiles: the Burr Type X and Weibull Cases

Comparing two Quantiles: the Burr Type X and Weibull Cases IOSR Journal of Mathematcs (IOSR-JM) e-issn: 78-578, -ISSN: 39-765X. Volume, Issue 5 Ver. VII (Se. - Oct.06), PP 8-40 www.osrjournals.org Comarng two Quantles: the Burr Tye X and Webull Cases Mohammed

More information

Nonlinear IV unit root tests in panels with cross-sectional dependency

Nonlinear IV unit root tests in panels with cross-sectional dependency Journal of Econometrcs 11 (22) 261 292 www.elsever.com/locate/econbase Nonlnear IV unt root tests n panels wth cross-sectonal dependency Yoosoon Chang Department of Economcs-MS22, Rce Unversty, 61 Man

More information

Solutions (mostly for odd-numbered exercises)

Solutions (mostly for odd-numbered exercises) Solutons (mostly for odd-numbered exercses) c 005 A. Coln Cameron and Pravn K. Trved "Mcroeconometrcs: Methods and Alcatons" 1. Chater 1: Introducton o exercses.. Chater : Causal and oncausal Models o

More information

Logistic regression with one predictor. STK4900/ Lecture 7. Program

Logistic regression with one predictor. STK4900/ Lecture 7. Program Logstc regresson wth one redctor STK49/99 - Lecture 7 Program. Logstc regresson wth one redctor 2. Maxmum lkelhood estmaton 3. Logstc regresson wth several redctors 4. Devance and lkelhood rato tests 5.

More information

6. Hamilton s Equations

6. Hamilton s Equations 6. Hamlton s Equatons Mchael Fowler A Dynamcal System s Path n Confguraton Sace and n State Sace The story so far: For a mechancal system wth n degrees of freedom, the satal confguraton at some nstant

More information

A General Class of Selection Procedures and Modified Murthy Estimator

A General Class of Selection Procedures and Modified Murthy Estimator ISS 684-8403 Journal of Statstcs Volume 4, 007,. 3-9 A General Class of Selecton Procedures and Modfed Murthy Estmator Abdul Bast and Muhammad Qasar Shahbaz Abstract A new selecton rocedure for unequal

More information

Computing MLE Bias Empirically

Computing MLE Bias Empirically Computng MLE Bas Emprcally Kar Wa Lm Australan atonal Unversty January 3, 27 Abstract Ths note studes the bas arses from the MLE estmate of the rate parameter and the mean parameter of an exponental dstrbuton.

More information

The Small Noise Arbitrage Pricing Theory

The Small Noise Arbitrage Pricing Theory The Small Nose Arbtrage Prcng Theory S. Satchell Trnty College Unversty of Cambrdge and School of Fnance and Economcs Unversty of Technology, Sydney December 998 Ths aer was wrtten when the Author was

More information

Panel cointegration rank test with cross-section dependence

Panel cointegration rank test with cross-section dependence Panel contegraton rank test wth cross-secton dependence Josep Lluís Carron--Slvestre Laura Surdeanu y AQR-IREA Research Group Department of Econometrcs, Statstcs and Spansh Economy Unversty of Barcelona

More information

Algorithms for factoring

Algorithms for factoring CSA E0 235: Crytograhy Arl 9,2015 Instructor: Arta Patra Algorthms for factorng Submtted by: Jay Oza, Nranjan Sngh Introducton Factorsaton of large ntegers has been a wdely studed toc manly because of

More information

Bias Corrections in Testing and Estimating Semiparametric, Single Index Models

Bias Corrections in Testing and Estimating Semiparametric, Single Index Models Bas Correctons n Testng and Estmatng Semarametrc, Sngle Index Models Roger Klen and Chan Shen July, 2009 Abstract Semarametrc methods are wdely emloyed n aled work where the ablty to conduct nferences

More information

Economics 130. Lecture 4 Simple Linear Regression Continued

Economics 130. Lecture 4 Simple Linear Regression Continued Economcs 130 Lecture 4 Contnued Readngs for Week 4 Text, Chapter and 3. We contnue wth addressng our second ssue + add n how we evaluate these relatonshps: Where do we get data to do ths analyss? How do

More information

Econ Statistical Properties of the OLS estimator. Sanjaya DeSilva

Econ Statistical Properties of the OLS estimator. Sanjaya DeSilva Econ 39 - Statstcal Propertes of the OLS estmator Sanjaya DeSlva September, 008 1 Overvew Recall that the true regresson model s Y = β 0 + β 1 X + u (1) Applyng the OLS method to a sample of data, we estmate

More information

Managing Capacity Through Reward Programs. on-line companion page. Byung-Do Kim Seoul National University College of Business Administration

Managing Capacity Through Reward Programs. on-line companion page. Byung-Do Kim Seoul National University College of Business Administration Managng Caacty Through eward Programs on-lne comanon age Byung-Do Km Seoul Natonal Unversty College of Busness Admnstraton Mengze Sh Unversty of Toronto otman School of Management Toronto ON M5S E6 Canada

More information

On New Selection Procedures for Unequal Probability Sampling

On New Selection Procedures for Unequal Probability Sampling Int. J. Oen Problems Comt. Math., Vol. 4, o. 1, March 011 ISS 1998-66; Coyrght ICSRS Publcaton, 011 www.-csrs.org On ew Selecton Procedures for Unequal Probablty Samlng Muhammad Qaser Shahbaz, Saman Shahbaz

More information

Chapter 2 - The Simple Linear Regression Model S =0. e i is a random error. S β2 β. This is a minimization problem. Solution is a calculus exercise.

Chapter 2 - The Simple Linear Regression Model S =0. e i is a random error. S β2 β. This is a minimization problem. Solution is a calculus exercise. Chapter - The Smple Lnear Regresson Model The lnear regresson equaton s: where y + = β + β e for =,..., y and are observable varables e s a random error How can an estmaton rule be constructed for the

More information

x i1 =1 for all i (the constant ).

x i1 =1 for all i (the constant ). Chapter 5 The Multple Regresson Model Consder an economc model where the dependent varable s a functon of K explanatory varables. The economc model has the form: y = f ( x,x,..., ) xk Approxmate ths by

More information

A Quadratic Cumulative Production Model for the Material Balance of Abnormally-Pressured Gas Reservoirs F.E. Gonzalez M.S.

A Quadratic Cumulative Production Model for the Material Balance of Abnormally-Pressured Gas Reservoirs F.E. Gonzalez M.S. Natural as Engneerng A Quadratc Cumulatve Producton Model for the Materal Balance of Abnormally-Pressured as Reservors F.E. onale M.S. Thess (2003) T.A. Blasngame, Texas A&M U. Deartment of Petroleum Engneerng

More information

Chapter 5 Multilevel Models

Chapter 5 Multilevel Models Chapter 5 Multlevel Models 5.1 Cross-sectonal multlevel models 5.1.1 Two-level models 5.1.2 Multple level models 5.1.3 Multple level modelng n other felds 5.2 Longtudnal multlevel models 5.2.1 Two-level

More information

A Quadratic Cumulative Production Model for the Material Balance of Abnormally-Pressured Gas Reservoirs F.E. Gonzalez M.S.

A Quadratic Cumulative Production Model for the Material Balance of Abnormally-Pressured Gas Reservoirs F.E. Gonzalez M.S. Formaton Evaluaton and the Analyss of Reservor Performance A Quadratc Cumulatve Producton Model for the Materal Balance of Abnormally-Pressured as Reservors F.E. onale M.S. Thess (2003) T.A. Blasngame,

More information

Testing for Granger Non-causality in Heterogeneous Panels

Testing for Granger Non-causality in Heterogeneous Panels Testng for Granger on-causalty n Heterogeneous Panels Chrstophe Hurln y June 27 Abstract Ths paper proposes a very smple test of Granger (1969) non-causalty for heterogeneous panel data models. Our test

More information

1. Inference on Regression Parameters a. Finding Mean, s.d and covariance amongst estimates. 2. Confidence Intervals and Working Hotelling Bands

1. Inference on Regression Parameters a. Finding Mean, s.d and covariance amongst estimates. 2. Confidence Intervals and Working Hotelling Bands Content. Inference on Regresson Parameters a. Fndng Mean, s.d and covarance amongst estmates.. Confdence Intervals and Workng Hotellng Bands 3. Cochran s Theorem 4. General Lnear Testng 5. Measures of

More information

Chat eld, C. and A.J.Collins, Introduction to multivariate analysis. Chapman & Hall, 1980

Chat eld, C. and A.J.Collins, Introduction to multivariate analysis. Chapman & Hall, 1980 MT07: Multvarate Statstcal Methods Mke Tso: emal mke.tso@manchester.ac.uk Webpage for notes: http://www.maths.manchester.ac.uk/~mkt/new_teachng.htm. Introducton to multvarate data. Books Chat eld, C. and

More information

Asymptotic Properties of the Jarque-Bera Test for Normality in General Autoregressions with a Deterministic Term

Asymptotic Properties of the Jarque-Bera Test for Normality in General Autoregressions with a Deterministic Term Asymptotc Propertes of the Jarque-Bera est for Normalty n General Autoregressons wth a Determnstc erm Carlos Caceres Nuffeld College, Unversty of Oxford May 2006 Abstract he am of ths paper s to analyse

More information

Machine Learning. Classification. Theory of Classification and Nonparametric Classifier. Representing data: Hypothesis (classifier) Eric Xing

Machine Learning. Classification. Theory of Classification and Nonparametric Classifier. Representing data: Hypothesis (classifier) Eric Xing Machne Learnng 0-70/5 70/5-78, 78, Fall 008 Theory of Classfcaton and Nonarametrc Classfer Erc ng Lecture, Setember 0, 008 Readng: Cha.,5 CB and handouts Classfcaton Reresentng data: M K Hyothess classfer

More information

e i is a random error

e i is a random error Chapter - The Smple Lnear Regresson Model The lnear regresson equaton s: where + β + β e for,..., and are observable varables e s a random error How can an estmaton rule be constructed for the unknown

More information

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 30 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 2 Remedes for multcollnearty Varous technques have

More information

Professor Chris Murray. Midterm Exam

Professor Chris Murray. Midterm Exam Econ 7 Econometrcs Sprng 4 Professor Chrs Murray McElhnney D cjmurray@uh.edu Mdterm Exam Wrte your answers on one sde of the blank whte paper that I have gven you.. Do not wrte your answers on ths exam.

More information

Statistics for Managers Using Microsoft Excel/SPSS Chapter 13 The Simple Linear Regression Model and Correlation

Statistics for Managers Using Microsoft Excel/SPSS Chapter 13 The Simple Linear Regression Model and Correlation Statstcs for Managers Usng Mcrosoft Excel/SPSS Chapter 13 The Smple Lnear Regresson Model and Correlaton 1999 Prentce-Hall, Inc. Chap. 13-1 Chapter Topcs Types of Regresson Models Determnng the Smple Lnear

More information

Comparison of Outlier Detection Methods in Crossover Design Bioequivalence Studies

Comparison of Outlier Detection Methods in Crossover Design Bioequivalence Studies Journal of Pharmacy and Nutrton Scences, 01,, 16-170 16 Comarson of Outler Detecton Methods n Crossover Desgn Boequvalence Studes A. Rasheed 1,*, T. Ahmad,# and J.S. Sddq,# 1 Deartment of Research, Dow

More information

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction ECONOMICS 5* -- NOTE (Summary) ECON 5* -- NOTE The Multple Classcal Lnear Regresson Model (CLRM): Specfcaton and Assumptons. Introducton CLRM stands for the Classcal Lnear Regresson Model. The CLRM s also

More information

Chapter 11: Simple Linear Regression and Correlation

Chapter 11: Simple Linear Regression and Correlation Chapter 11: Smple Lnear Regresson and Correlaton 11-1 Emprcal Models 11-2 Smple Lnear Regresson 11-3 Propertes of the Least Squares Estmators 11-4 Hypothess Test n Smple Lnear Regresson 11-4.1 Use of t-tests

More information

Interval Estimation in the Classical Normal Linear Regression Model. 1. Introduction

Interval Estimation in the Classical Normal Linear Regression Model. 1. Introduction ECONOMICS 35* -- NOTE 7 ECON 35* -- NOTE 7 Interval Estmaton n the Classcal Normal Lnear Regresson Model Ths note outlnes the basc elements of nterval estmaton n the Classcal Normal Lnear Regresson Model

More information

Exam. Econometrics - Exam 1

Exam. Econometrics - Exam 1 Econometrcs - Exam 1 Exam Problem 1: (15 ponts) Suppose that the classcal regresson model apples but that the true value of the constant s zero. In order to answer the followng questons assume just one

More information

Factor models with many assets: strong factors, weak factors, and the two-pass procedure

Factor models with many assets: strong factors, weak factors, and the two-pass procedure Factor models wth many assets: strong factors, weak factors, and the two-pass procedure Stanslav Anatolyev 1 Anna Mkusheva 2 1 CERGE-EI and NES 2 MIT December 2017 Stanslav Anatolyev and Anna Mkusheva

More information

Lecture 4 Hypothesis Testing

Lecture 4 Hypothesis Testing Lecture 4 Hypothess Testng We may wsh to test pror hypotheses about the coeffcents we estmate. We can use the estmates to test whether the data rejects our hypothess. An example mght be that we wsh to

More information

Supplementary Material for Spectral Clustering based on the graph p-laplacian

Supplementary Material for Spectral Clustering based on the graph p-laplacian Sulementary Materal for Sectral Clusterng based on the grah -Lalacan Thomas Bühler and Matthas Hen Saarland Unversty, Saarbrücken, Germany {tb,hen}@csun-sbde May 009 Corrected verson, June 00 Abstract

More information

Lecture 3 Stat102, Spring 2007

Lecture 3 Stat102, Spring 2007 Lecture 3 Stat0, Sprng 007 Chapter 3. 3.: Introducton to regresson analyss Lnear regresson as a descrptve technque The least-squares equatons Chapter 3.3 Samplng dstrbuton of b 0, b. Contnued n net lecture

More information

Global Sensitivity. Tuesday 20 th February, 2018

Global Sensitivity. Tuesday 20 th February, 2018 Global Senstvty Tuesday 2 th February, 28 ) Local Senstvty Most senstvty analyses [] are based on local estmates of senstvty, typcally by expandng the response n a Taylor seres about some specfc values

More information

Computation of Higher Order Moments from Two Multinomial Overdispersion Likelihood Models

Computation of Higher Order Moments from Two Multinomial Overdispersion Likelihood Models Computaton of Hgher Order Moments from Two Multnomal Overdsperson Lkelhood Models BY J. T. NEWCOMER, N. K. NEERCHAL Department of Mathematcs and Statstcs, Unversty of Maryland, Baltmore County, Baltmore,

More information

Department of Statistics University of Toronto STA305H1S / 1004 HS Design and Analysis of Experiments Term Test - Winter Solution

Department of Statistics University of Toronto STA305H1S / 1004 HS Design and Analysis of Experiments Term Test - Winter Solution Department of Statstcs Unversty of Toronto STA35HS / HS Desgn and Analyss of Experments Term Test - Wnter - Soluton February, Last Name: Frst Name: Student Number: Instructons: Tme: hours. Ads: a non-programmable

More information

Multiple Regression Analysis

Multiple Regression Analysis Multle Regresson Analss Roland Szlág Ph.D. Assocate rofessor Correlaton descres the strength of a relatonsh, the degree to whch one varale s lnearl related to another Regresson shows us how to determne

More information

PANEL UNIT ROOT TESTS UNDER CROSS-SECTIONAL DEPENDENCE: AN OVERVIEW

PANEL UNIT ROOT TESTS UNDER CROSS-SECTIONAL DEPENDENCE: AN OVERVIEW Journal of Statstcs: Advances n Theory and Applcatons Volume, Number 2, 2009, Pages 7-58 PANEL UNIT ROOT TESTS UNDER CROSS-SECTIONAL DEPENDENCE: AN OVERVIEW LAURA BARBIERI Dpartmento d Scenze Economche

More information

Composite Hypotheses testing

Composite Hypotheses testing Composte ypotheses testng In many hypothess testng problems there are many possble dstrbutons that can occur under each of the hypotheses. The output of the source s a set of parameters (ponts n a parameter

More information

See Book Chapter 11 2 nd Edition (Chapter 10 1 st Edition)

See Book Chapter 11 2 nd Edition (Chapter 10 1 st Edition) Count Data Models See Book Chapter 11 2 nd Edton (Chapter 10 1 st Edton) Count data consst of non-negatve nteger values Examples: number of drver route changes per week, the number of trp departure changes

More information

Testing for Unit Roots in Panel Data: An Exploration Using Real and Simulated Data

Testing for Unit Roots in Panel Data: An Exploration Using Real and Simulated Data Testng for Unt Roots n Panel Data: An Exploraton Usng Real and Smulated Data Bronwyn H. HALL UC Berkeley, Oxford Unversty, and NBER Jacques MAIRESSE INSEE-CREST, EHESS, and NBER Introducton! Our Research

More information

Endogenous timing in a mixed oligopoly consisting of a single public firm and foreign competitors. Abstract

Endogenous timing in a mixed oligopoly consisting of a single public firm and foreign competitors. Abstract Endogenous tmng n a mxed olgopoly consstng o a sngle publc rm and oregn compettors Yuanzhu Lu Chna Economcs and Management Academy, Central Unversty o Fnance and Economcs Abstract We nvestgate endogenous

More information

Transfer Functions. Convenient representation of a linear, dynamic model. A transfer function (TF) relates one input and one output: ( ) system

Transfer Functions. Convenient representation of a linear, dynamic model. A transfer function (TF) relates one input and one output: ( ) system Transfer Functons Convenent representaton of a lnear, dynamc model. A transfer functon (TF) relates one nput and one output: x t X s y t system Y s The followng termnology s used: x y nput output forcng

More information

Ryan (2009)- regulating a concentrated industry (cement) Firms play Cournot in the stage. Make lumpy investment decisions

Ryan (2009)- regulating a concentrated industry (cement) Firms play Cournot in the stage. Make lumpy investment decisions 1 Motvaton Next we consder dynamc games where the choce varables are contnuous and/or dscrete. Example 1: Ryan (2009)- regulatng a concentrated ndustry (cement) Frms play Cournot n the stage Make lumpy

More information

Are Health Expenditure and GDP Cointegrated: A Panel Analysis

Are Health Expenditure and GDP Cointegrated: A Panel Analysis Journal of Busness and Economcs ISSN 2155-7950 USA December 2010 Volume 1 No. 1 Academc Star Publshng Company 2010 http://www.academcstar.us Are Health Expendture and GDP Contegrated: A Panel Analyss Engn

More information

Simulated Power of the Discrete Cramér-von Mises Goodness-of-Fit Tests

Simulated Power of the Discrete Cramér-von Mises Goodness-of-Fit Tests Smulated of the Cramér-von Mses Goodness-of-Ft Tests Steele, M., Chaselng, J. and 3 Hurst, C. School of Mathematcal and Physcal Scences, James Cook Unversty, Australan School of Envronmental Studes, Grffth

More information

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number

More information

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography CSc 6974 and ECSE 6966 Math. Tech. for Vson, Graphcs and Robotcs Lecture 21, Aprl 17, 2006 Estmatng A Plane Homography Overvew We contnue wth a dscusson of the major ssues, usng estmaton of plane projectve

More information

Probability, Statistics, and Reliability for Engineers and Scientists SIMULATION

Probability, Statistics, and Reliability for Engineers and Scientists SIMULATION CHATER robablty, Statstcs, and Relablty or Engneers and Scentsts Second Edton SIULATIO A. J. Clark School o Engneerng Department o Cvl and Envronmental Engneerng 7b robablty and Statstcs or Cvl Engneers

More information

a. (All your answers should be in the letter!

a. (All your answers should be in the letter! Econ 301 Blkent Unversty Taskn Econometrcs Department of Economcs Md Term Exam I November 8, 015 Name For each hypothess testng n the exam complete the followng steps: Indcate the test statstc, ts crtcal

More information

Complete Variance Decomposition Methods. Cédric J. Sallaberry

Complete Variance Decomposition Methods. Cédric J. Sallaberry Comlete Varance Decomoston Methods Cédrc J. allaberry enstvty Analyss y y [,,, ] [ y, y,, ] y ny s a vector o uncertan nuts s a vector o results s a comle uncton successon o derent codes, systems o de,

More information

On Comparison of Some Ridge Parameters in Ridge Regression

On Comparison of Some Ridge Parameters in Ridge Regression Sr Lankan Journal of Aled Statstcs, Vol (15-1) On Comarson of Some Rdge Parameters n Rdge Regresson Ashok V. Dorugade Y C Mahavdyalaya Halkarn, Tal-Chandgad, Kolhaur, Maharashtra, Inda Corresondng Author:

More information

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal Inner Product Defnton 1 () A Eucldean space s a fnte-dmensonal vector space over the reals R, wth an nner product,. Defnton 2 (Inner Product) An nner product, on a real vector space X s a symmetrc, blnear,

More information

Basic Business Statistics, 10/e

Basic Business Statistics, 10/e Chapter 13 13-1 Basc Busness Statstcs 11 th Edton Chapter 13 Smple Lnear Regresson Basc Busness Statstcs, 11e 009 Prentce-Hall, Inc. Chap 13-1 Learnng Objectves In ths chapter, you learn: How to use regresson

More information

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers Psychology 282 Lecture #24 Outlne Regresson Dagnostcs: Outlers In an earler lecture we studed the statstcal assumptons underlyng the regresson model, ncludng the followng ponts: Formal statement of assumptons.

More information

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6 Department of Quanttatve Methods & Informaton Systems Tme Seres and Ther Components QMIS 30 Chapter 6 Fall 00 Dr. Mohammad Zanal These sldes were modfed from ther orgnal source for educatonal purpose only.

More information

Hidden Markov Models & The Multivariate Gaussian (10/26/04)

Hidden Markov Models & The Multivariate Gaussian (10/26/04) CS281A/Stat241A: Statstcal Learnng Theory Hdden Markov Models & The Multvarate Gaussan (10/26/04) Lecturer: Mchael I. Jordan Scrbes: Jonathan W. Hu 1 Hdden Markov Models As a bref revew, hdden Markov models

More information

be a second-order and mean-value-zero vector-valued process, i.e., for t E

be a second-order and mean-value-zero vector-valued process, i.e., for t E CONFERENCE REPORT 617 DSCUSSON OF TWO PROCEDURES FOR EXPANDNG A VECTOR-VALUED STOCHASTC PROCESS N AN ORTHONORMAL WAY by R. GUTkRREZ and M. J. VALDERRAMA 1. ntroducton Snce K. Karhunen [l] and M. Lo&e [2]

More information

LECTURE 9 CANONICAL CORRELATION ANALYSIS

LECTURE 9 CANONICAL CORRELATION ANALYSIS LECURE 9 CANONICAL CORRELAION ANALYSIS Introducton he concept of canoncal correlaton arses when we want to quantfy the assocatons between two sets of varables. For example, suppose that the frst set of

More information

Systems of Equations (SUR, GMM, and 3SLS)

Systems of Equations (SUR, GMM, and 3SLS) Lecture otes on Advanced Econometrcs Takash Yamano Fall Semester 4 Lecture 4: Sstems of Equatons (SUR, MM, and 3SLS) Seemngl Unrelated Regresson (SUR) Model Consder a set of lnear equatons: $ + ɛ $ + ɛ

More information

290 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I: FUNDAMENTAL THEORY AND APPLICATIONS, VOL. 45, NO. 3, MARCH H d (e j! ;e j!

290 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I: FUNDAMENTAL THEORY AND APPLICATIONS, VOL. 45, NO. 3, MARCH H d (e j! ;e j! 9 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I: FUNDAMENTAL THEORY AND APPLICATIONS, VOL. 45, NO. 3, MARCH 998 Transactons Brefs Two-Dmensonal FIR Notch Flter Desgn Usng Sngular Value Decomoston S.-C. Pe,

More information

Journal of Multivariate Analysis

Journal of Multivariate Analysis Journal of Multvarate Analyss 07 (202) 232 243 Contents lsts avalable at ScVerse ScenceDrect Journal of Multvarate Analyss journal homeage: www.elsever.com/locate/jmva James Sten tye estmators of varances

More information

Lecture 12: Discrete Laplacian

Lecture 12: Discrete Laplacian Lecture 12: Dscrete Laplacan Scrbe: Tanye Lu Our goal s to come up wth a dscrete verson of Laplacan operator for trangulated surfaces, so that we can use t n practce to solve related problems We are mostly

More information

Lecture 10 Support Vector Machines II

Lecture 10 Support Vector Machines II Lecture 10 Support Vector Machnes II 22 February 2016 Taylor B. Arnold Yale Statstcs STAT 365/665 1/28 Notes: Problem 3 s posted and due ths upcomng Frday There was an early bug n the fake-test data; fxed

More information

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS Avalable onlne at http://sck.org J. Math. Comput. Sc. 3 (3), No., 6-3 ISSN: 97-537 COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

More information

STAT 511 FINAL EXAM NAME Spring 2001

STAT 511 FINAL EXAM NAME Spring 2001 STAT 5 FINAL EXAM NAME Sprng Instructons: Ths s a closed book exam. No notes or books are allowed. ou may use a calculator but you are not allowed to store notes or formulas n the calculator. Please wrte

More information

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours UNIVERSITY OF TORONTO Faculty of Arts and Scence December 005 Examnatons STA47HF/STA005HF Duraton - hours AIDS ALLOWED: (to be suppled by the student) Non-programmable calculator One handwrtten 8.5'' x

More information

/ n ) are compared. The logic is: if the two

/ n ) are compared. The logic is: if the two STAT C141, Sprng 2005 Lecture 13 Two sample tests One sample tests: examples of goodness of ft tests, where we are testng whether our data supports predctons. Two sample tests: called as tests of ndependence

More information

Negative Binomial Regression

Negative Binomial Regression STATGRAPHICS Rev. 9/16/2013 Negatve Bnomal Regresson Summary... 1 Data Input... 3 Statstcal Model... 3 Analyss Summary... 4 Analyss Optons... 7 Plot of Ftted Model... 8 Observed Versus Predcted... 10 Predctons...

More information

Maximum Likelihood Estimation of Binary Dependent Variables Models: Probit and Logit. 1. General Formulation of Binary Dependent Variables Models

Maximum Likelihood Estimation of Binary Dependent Variables Models: Probit and Logit. 1. General Formulation of Binary Dependent Variables Models ECO 452 -- OE 4: Probt and Logt Models ECO 452 -- OE 4 Maxmum Lkelhood Estmaton of Bnary Dependent Varables Models: Probt and Logt hs note demonstrates how to formulate bnary dependent varables models

More information

5 The Rational Canonical Form

5 The Rational Canonical Form 5 The Ratonal Canoncal Form Here p s a monc rreducble factor of the mnmum polynomal m T and s not necessarly of degree one Let F p denote the feld constructed earler n the course, consstng of all matrces

More information

Here is the rationale: If X and y have a strong positive relationship to one another, then ( x x) will tend to be positive when ( y y)

Here is the rationale: If X and y have a strong positive relationship to one another, then ( x x) will tend to be positive when ( y y) Secton 1.5 Correlaton In the prevous sectons, we looked at regresson and the value r was a measurement of how much of the varaton n y can be attrbuted to the lnear relatonshp between y and x. In ths secton,

More information

Simulation and Probability Distribution

Simulation and Probability Distribution CHAPTER Probablty, Statstcs, and Relablty for Engneers and Scentsts Second Edton PROBABILIT DISTRIBUTION FOR CONTINUOUS RANDOM VARIABLES A. J. Clark School of Engneerng Department of Cvl and Envronmental

More information

An (almost) unbiased estimator for the S-Gini index

An (almost) unbiased estimator for the S-Gini index An (almost unbased estmator for the S-Gn ndex Thomas Demuynck February 25, 2009 Abstract Ths note provdes an unbased estmator for the absolute S-Gn and an almost unbased estmator for the relatve S-Gn for

More information

1 Convex Optimization

1 Convex Optimization Convex Optmzaton We wll consder convex optmzaton problems. Namely, mnmzaton problems where the objectve s convex (we assume no constrants for now). Such problems often arse n machne learnng. For example,

More information

Comparison of the Population Variance Estimators. of 2-Parameter Exponential Distribution Based on. Multiple Criteria Decision Making Method

Comparison of the Population Variance Estimators. of 2-Parameter Exponential Distribution Based on. Multiple Criteria Decision Making Method Appled Mathematcal Scences, Vol. 7, 0, no. 47, 07-0 HIARI Ltd, www.m-hkar.com Comparson of the Populaton Varance Estmators of -Parameter Exponental Dstrbuton Based on Multple Crtera Decson Makng Method

More information

Comparison of Regression Lines

Comparison of Regression Lines STATGRAPHICS Rev. 9/13/2013 Comparson of Regresson Lnes Summary... 1 Data Input... 3 Analyss Summary... 4 Plot of Ftted Model... 6 Condtonal Sums of Squares... 6 Analyss Optons... 7 Forecasts... 8 Confdence

More information

x = , so that calculated

x = , so that calculated Stat 4, secton Sngle Factor ANOVA notes by Tm Plachowsk n chapter 8 we conducted hypothess tests n whch we compared a sngle sample s mean or proporton to some hypotheszed value Chapter 9 expanded ths to

More information

2016 Wiley. Study Session 2: Ethical and Professional Standards Application

2016 Wiley. Study Session 2: Ethical and Professional Standards Application 6 Wley Study Sesson : Ethcal and Professonal Standards Applcaton LESSON : CORRECTION ANALYSIS Readng 9: Correlaton and Regresson LOS 9a: Calculate and nterpret a sample covarance and a sample correlaton

More information