NGHIÊN cứu mối QUAN hệ GIỮA lạm PHÁT và THÂM hụt NGÂN SÁCH ở VIỆT NAM

B GIÁO D O I H C KINH T TP.HCM -oOo THANH XUÂN NGHIÊN C U M I QUAN H GI A L M PHÁT VÀ THÂM H T NGÂN SÁCH Chuyên ngành : Tài VI T NAM Ngân hàng Mã s : 60340201 LU I NG D N KHOA H C: PGS TS NGUY N TH LIÊN HOA Thành ph H Chí Minh, L NGHIÊN C U M I QUAN H PHÁT VÀ THÂM H T NGÂN SÁCH c a tác gi , n VI T NAM GI A L M công trình nghiên c u t t trình h c t p k t qu nghiên c u th c ti n th i gian qua, s li u s d ng trung th c có ngu n g c trích d n rõ ràng Lu c th c hi is ng d n khoa h c c a PGS TS NGUY N TH LIÊN HOA Tác gi lu THANH XUÂN L IC Tôi chân thành c ih h c Kinh T Thành ph H i u ki n thu n l i cho h c t p nghiên c u su t th i gian qua Tôi chân thành c i h c Kinh T Thành ph H Chí t tình gi ng d y cho su t trình tham gia h c t p t c bi t c Cô PGS TS NGUY N TH LIÊN HOA ng, n tình ch b o, góp ý ng viên su t trình th c hi n lu Trong trình hoàn t tài, m nhi u ý ki g ng tham kh o tài li u, tham kh o u không th tránh kh i R t mong nh n Quý Th ng nghi p b n Xin trân tr ng c Tác gi lu THANH XUÂN M CL C TÓM T T 1 I THI U 1.1 L im 1.2 M c tiêu nghiên c u 1.3 Câu h i nghiên c u 1.4 ng ph m vi nghiên c u u NG QUAN CÁC NGHIÊN C 2.1 T ng quan h c thuy t kinh t Keynes m i 2.2 T 2.3 T ng quan nghiên c ng cong Laffer LI U 17 3.1 17 3.2 u c a lu 18 3.3 Mô hình nghiên c u 19 3.4 Mô t d li u 19 T QU NGHIÊN C U 20 4.1 Ki m ch ng y u t tài chính: Chi tiêu ph , cung ti n l m phát tác n thâm h t ngân sách t i Vi t Nam 20 4.2 K t qu nghiên c u y u t ng lên thâm h t ngân sách Vi t Nam 21 4.2.1 Phân tích quan h nhân qu 21 4.2.1.1 tr t 21 4.2.1.2 Phân tích mô hình VAR 22 4.2.1.3 Phân tích m i quan h nhân qu Granger theo ti p c n Dolado Lütkepohl 23 4.2.1.4 4.2.2 Ki Phân tích s nh c a mô hình VAR 24 nh quan h nhân qu Granger truy n th ng 25 4.2.2.1 Ki nh nghi 25 4.2.2.2 Ki ng liên k t 27 4.2.2.3 Ki nh VECM 32 4.2.2.4 nh m i quan h ng n h n gi a bi n nghiên c u 35 4.2.2.5 Ki m tra tính b n v ng mô hình VECM 36 4.2.2.6 Ki m nh quan h nhân qu Granger 36 4.2.2.7 So sánh k t qu ki Granger truy n th ng 37 NG K T 39 5.1 K t qu nghiên c u 39 5.2 Ki n ngh gi i pháp 40 5.3 Nh ng h n ch c a lu 5.3.1 H n ch c a lu 5.3.2 ng nghiên c u ti p theo 41 41 ng nghiên c u ti p theo 42 TÀI LI U THAM KH O 43 DANH M C T VI T T T ADB Ngân hàng phát tri n Châu Á ADF nh nghi theo Augmented Dickey- Fuller AIC Akaike Information Criterion DL Dolado Lütkepohl FD Thâm h t ngân sách GE Chi tiêu ph IMF Qu ti n t qu c t INF L m phát MS Cung ti n M2 NHNN c OLS PP SIC VAR VECM i thi u m nh nghi Schwarz Information Criterion h i quy VAR Mô hình hi u ch nh sai s VECM theo Philips-Perron DANH M C B NG B ng 2.1: Tóm t t t ng quan nghiên c B ng 4.1: Tóm t t ch tiêu th ng kê mô t c a bi n tài B ng 4.2 tr t a mô hình VAR B ng 4.3: K t qu phân tích mô hình VAR B ng 4.4: Phân tích quan h nhân qu Granger theo cách ti p c n DL B ng 4.5: Phân tích s nh c a mô hình VAR B ng 4.6: Ki nh nghi B ng 4.7: Ki nh tính d ng ph B ng 4.8: Ki m tra l a ch n mô hình B ng 4.9: Ki ng liên k t B ng 4.10: Ki nh VECM B ng 4.11: Ki nh tính d ng ph B ng 4.12: K t qu ng m i quan h ng n h n B ng 4.13: Ki m tra tính b n v ng mô hình VECM B ng 4.14: Ki nh nhân qu Granger Nghiên c c th c hi ki m tra m i quan h nhân qu gi a thâm h t ngân sách, chi tiêu ph , cung ti n l m phát Tác gi ti n hành ki thuy t cho Vi t Nam v i d li u t gi d nh gi 2012 Trong nghiên c u này, tác a Dolado Lütkepohl ti p c n quan h nhân qu Granger truy n th ng ki m tra quan h nhân qu tr c ti p gi a bi n nghiên c u V i k t qu th c nghi d ng Unit root test, ki c qua ki nh nhân qu Granger, ki nh tính ng liên k t Engle- Granger Johansen Juselius (1990) mô hình VECM (Vector Error Correction Models) cho th y r ng: Phân tích quan h nhân qu d thâm h t ngân sách, l m phát cung ti p c n theo DL cho r ng ng gây chi tiêu ph không t n t i m i quan h gây thâm h t ngân sách Ki nh quan h Granger truy n th ng cho r ng có m i quan h hai chi u gi a thâm h t ngân sách chi tiêu ph , m i quan h m t chi u gi a l m phát cung ti có m i quan h ngu n cung ti n chi tiêu ph Ngoài ra, k t qu cho th y ng m t chi u gây l m phát I THI U 1.1 L i m u Kinh t Vi i m t v i nhi u b t n kinh t k t cu c kh ng ho ng kinh t tài toàn c u x y vào g a kinh t th gi Nh ng bi c l nh ng m khuy v im n c a n n kinh t c m t mà coi nh s ng kinh t nh lâu dài c suy gi m, t m c 8, 2007, xu ng x p x n 2008 n 2004 2012 Trong kh m c hai s , bình quân kho ng g l l m phát n 2007-2012 bi t thâm h t ngân sách cao n p t c nh ng làm x u thêm ch s kinh t as c n GDP vào cu i x p x 5c l c 57% 42% px Các nghiên c u th c nghi m nhi tài khóa y u nguyên nhân d c th gi r ng, s qu n lý n hàng lo t v c a n n kinh t S ng qui t c kinh t L m phát m t không ph i xa l m Vì th , vi quan tr ng nh t ng qua l i c a thâm h t ngân sách l m phát h t s c ph c t p không ph kinh t v i m ng ng âm Do v y, m i quan h gi a thâm h t ngân sách, cung ti n, chi tiêu ngân sách l m phát m t nh ng v m c a n n kinh t hàng hóa T i m i th i kì ng kinh t khác s có nh ng m c l m phát phù h p nh m i quan h gi a thâm h t ngân sách l c quan tâm nghiên c u c a nhi u nhà kinh t , tác gi u m i quan h gi a l m phát thâm h t ngân sách lu kinh t nghiêm m phát cao dai d ng, thâm h t cán cân tài kho n vãng lai l th p, ho c th v m n nh c a n n kinh t lai Thâm h t ngân sách nh n công c u qu c a sách kích thích kinh t tr ng tiêu c c ch n Vi tài: cho a V i m c tiêu nghiên c u l m phát có ph i t t c nguyên nhân gây nên thâm h t ngân sách hay không y u t chi tiêu ph , cung ti n có tác n thâm h t ngân sách Vi t Nam 1.2 M c tiêu nghiên c u nh l m phát, chi tiêu ph , cung ti n, y u t nhân t tác n thâm h t ngân sách Vi t Nam Tìm hi u m i quan h nhân qu gi a y u t tài chính: Chi tiêu ph , l m phát, cung ti n thâm h t ngân sách Vi t Nam 1.3 Câu h i nghiên c u gi i thích m c tiêu nghiên c u trên, tác gi t s câu h i nghiên c u sau: Chi tiêu ph , l m phát, cung ti n có quan h ng n thâm h t ngân sách Vi t Nam? Có t n t i m i quan h nhân qu hai chi ng gi a bi n nghiên c u: Thâm h t ngân sách, chi tiêu ph , l m phát, cung ti n hay không? 1.4 ng ph m vi nghiên c u ng nghiên c u: M i quan h gi a chi tiêu ph , l m phát, cung ti n thâm h t ngân sách Vi t Nam Ph m vi nghiên c u: N n kinh t Vi n 1990 2012 73 Ki nh nghi n, xu th ): Bi n LFD: Null Hypothesis: LFD has a unit root Exogenous: Constant, Linear Trend Bandwidth: (Newey-West using Bartlett kernel) Adj t-Stat Phillips-Perron test statistic Test critical values: -5.119618 -4.440739 -3.632896 -3.254671 1% level 5% level 10% level Prob.* 0.0024 *MacKinnon (1996) one-sided p-values Residual variance (no correction) HAC corrected variance (Bartlett kernel) 0.155810 0.155810 Phillips-Perron Test Equation Dependent Variable: D(LFD) Method: Least Squares Date: 08/07/13 Time: 14:28 Sample (adjusted): 1991 2012 Included observations: 22 after adjustments Variable Coefficient Std Error t-Statistic Prob LFD(-1) C @TREND(1990) -1.116642 8.245076 0.240044 0.218110 1.598655 0.047974 -5.119618 5.157509 5.003660 0.0001 0.0001 0.0001 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.581149 0.537060 0.424749 3.427815 -10.76634 13.18111 0.000257 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat 0.180988 0.624266 1.251485 1.400264 1.286533 1.484495 74 Bi n D(LFD): Null Hypothesis: D(LFD) has a unit root Exogenous: Constant, Linear Trend Bandwidth: (Newey-West using Bartlett kernel) Adj t-Stat Phillips-Perron test statistic Test critical values: -13.66545 -4.467895 -3.644963 -3.261452 1% level 5% level 10% level Prob.* 0.0000 *MacKinnon (1996) one-sided p-values Residual variance (no correction) HAC corrected variance (Bartlett kernel) 0.204347 0.051676 Phillips-Perron Test Equation Dependent Variable: D(LFD,2) Method: Least Squares Date: 08/07/13 Time: 14:30 Sample (adjusted): 1992 2012 Included observations: 21 after adjustments Variable Coefficient Std Error t-Statistic Prob D(LFD(-1)) C @TREND(1990) -1.387473 0.487064 -0.013725 0.170999 0.237369 0.017615 -8.113943 2.051931 -0.779154 0.0000 0.0550 0.4460 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.788686 0.765207 0.488267 4.291280 -13.12437 33.59067 0.000001 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat 0.078543 1.007661 1.535654 1.684872 1.568038 2.434523 75 Bi n LGE: Null Hypothesis: LGE has a unit root Exogenous: Constant, Linear Trend Bandwidth: (Newey-West using Bartlett kernel) Adj t-Stat Phillips-Perron test statistic Test critical values: -2.886529 -4.440739 -3.632896 -3.254671 1% level 5% level 10% level Prob.* 0.1851 *MacKinnon (1996) one-sided p-values Residual variance (no correction) HAC corrected variance (Bartlett kernel) 0.010688 0.008914 Phillips-Perron Test Equation Dependent Variable: D(LGE) Method: Least Squares Date: 08/07/13 Time: 14:31 Sample (adjusted): 1991 2012 Included observations: 22 after adjustments Variable Coefficient Std Error t-Statistic Prob LGE(-1) C @TREND(1990) -0.316339 3.306085 0.051070 0.111937 1.059081 0.021842 -2.826049 3.121654 2.338219 0.0108 0.0056 0.0305 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.437538 0.378331 0.111247 0.235142 18.70803 7.390027 0.004226 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat 0.204231 0.141094 -1.428003 -1.279224 -1.392955 1.708739 76 Bi n D(LGE): Null Hypothesis: D(LGE) has a unit root Exogenous: Constant, Linear Trend Bandwidth: (Newey-West using Bartlett kernel) Adj t-Stat Phillips-Perron test statistic Test critical values: -3.666638 -4.467895 -3.644963 -3.261452 1% level 5% level 10% level Prob.* 0.0480 *MacKinnon (1996) one-sided p-values Residual variance (no correction) HAC corrected variance (Bartlett kernel) 0.014173 0.017483 Phillips-Perron Test Equation Dependent Variable: D(LGE,2) Method: Least Squares Date: 08/07/13 Time: 14:31 Sample (adjusted): 1992 2012 Included observations: 21 after adjustments Variable Coefficient Std Error t-Statistic Prob D(LGE(-1)) C @TREND(1990) -0.799936 0.283231 -0.009981 0.224540 0.094630 0.005032 -3.562558 2.993045 -1.983498 0.0022 0.0078 0.0628 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.421397 0.357107 0.128588 0.297626 14.89492 6.554695 0.007268 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat -0.006501 0.160372 -1.132850 -0.983632 -1.100466 0.967104 77 Bi n LINF: Null Hypothesis: LINF has a unit root Exogenous: Constant, Linear Trend Bandwidth: (Newey-West using Bartlett kernel) Adj t-Stat Phillips-Perron test statistic Test critical values: -6.817153 -4.440739 -3.632896 -3.254671 1% level 5% level 10% level Prob.* 0.0001 *MacKinnon (1996) one-sided p-values Residual variance (no correction) HAC corrected variance (Bartlett kernel) 0.008258 0.006324 Phillips-Perron Test Equation Dependent Variable: D(LINF) Method: Least Squares Date: 08/08/13 Time: 15:17 Sample (adjusted): 1991 2012 Included observations: 22 after adjustments Variable Coefficient Std Error t-Statistic Prob LINF(-1) C @TREND(1990) -0.645425 2.667151 0.016477 0.103862 0.400732 0.005474 -6.214279 6.655692 3.010224 0.0000 0.0000 0.0072 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.721704 0.692409 0.097786 0.181680 21.54543 24.63626 0.000005 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat 0.068017 0.176315 -1.685948 -1.537169 -1.650900 1.910980 78 Bi n D(LINF): Null Hypothesis: D(LINF) has a unit root Exogenous: Constant, Linear Trend Bandwidth: 11 (Newey-West using Bartlett kernel) Adj t-Stat Phillips-Perron test statistic Test critical values: -7.090700 -4.467895 -3.644963 -3.261452 1% level 5% level 10% level Prob.* 0.0001 *MacKinnon (1996) one-sided p-values Residual variance (no correction) HAC corrected variance (Bartlett kernel) 0.015550 0.002725 Phillips-Perron Test Equation Dependent Variable: D(LINF,2) Method: Least Squares Date: 08/08/13 Time: 15:17 Sample (adjusted): 1992 2012 Included observations: 21 after adjustments Variable Coefficient Std Error t-Statistic Prob D(LINF(-1)) C @TREND(1990) -0.831999 0.069264 -0.003337 0.186421 0.073616 0.005107 -4.463020 0.940885 -0.653432 0.0003 0.3592 0.5217 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.532639 0.480710 0.134690 0.326547 13.92117 10.25708 0.001064 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat -0.037970 0.186910 -1.040111 -0.890894 -1.007727 2.182511 79 Bi n LMS: Null Hypothesis: LMS has a unit root Exogenous: Constant, Linear Trend Bandwidth: (Newey-West using Bartlett kernel) Adj t-Stat Phillips-Perron test statistic Test critical values: -3.595030 -4.440739 -3.632896 -3.254671 1% level 5% level 10% level Prob.* 0.0537 *MacKinnon (1996) one-sided p-values Residual variance (no correction) HAC corrected variance (Bartlett kernel) 0.005377 0.005377 Phillips-Perron Test Equation Dependent Variable: D(LMS) Method: Least Squares Date: 08/07/13 Time: 14:33 Sample (adjusted): 1991 2012 Included observations: 22 after adjustments Variable Coefficient Std Error t-Statistic Prob LMS(-1) C @TREND(1990) -0.700949 6.836457 0.178655 0.194977 1.812171 0.051210 -3.595030 3.772523 3.488652 0.0019 0.0013 0.0025 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.468849 0.412939 0.078902 0.118285 26.26608 8.385700 0.002452 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat 0.263043 0.102978 -2.115099 -1.966320 -2.080051 1.035577 80 Bi n D(LMS): Null Hypothesis: D(LMS) has a unit root Exogenous: Constant, Linear Trend Bandwidth: (Newey-West using Bartlett kernel) Adj t-Stat Phillips-Perron test statistic Test critical values: -5.115947 -4.467895 -3.644963 -3.261452 1% level 5% level 10% level Prob.* 0.0027 *MacKinnon (1996) one-sided p-values Residual variance (no correction) HAC corrected variance (Bartlett kernel) 0.005381 0.004629 Phillips-Perron Test Equation Dependent Variable: D(LMS,2) Method: Least Squares Date: 08/07/13 Time: 14:33 Sample (adjusted): 1992 2012 Included observations: 21 after adjustments Variable Coefficient Std Error t-Statistic Prob D(LMS(-1)) C @TREND(1990) -0.890538 0.229841 -0.000934 0.178549 0.069306 0.002973 -4.987642 3.316337 -0.314148 0.0001 0.0038 0.7570 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.592142 0.546824 0.079229 0.112991 25.06443 13.06648 0.000312 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat -0.019590 0.117694 -2.101374 -1.952156 -2.068990 1.867097 81 Ph l c 6: L a ch n mô hình ki nh Johansen: Date: 08/08/13 Time: 16:36 Sample: 1990 2012 Included observations: 21 Series: LFD LGE LINF LMS Lags interval: to Selected (0.05 level*) Number of Cointegrating Relations by Model Data Trend: Test Type Trace Max-Eig None No Intercept No Trend 1 None Intercept No Trend 1 Linear Intercept No Trend 1 Linear Intercept Trend Quadratic Intercept Trend *Critical values based on MacKinnon-Haug-Michelis (1999) Information Criteria by Rank and Model Data Trend: Rank or No of CEs None No Intercept No Trend None Intercept No Trend Linear Intercept No Trend Linear Intercept Trend Quadratic Intercept Trend Log Likelihood by Rank (rows) and Model (columns) 59.56567 74.96379 82.21798 85.19444 85.45423 59.56567 80.62334 89.23455 95.39712 97.95091 72.20836 86.73377 94.80141 97.39254 97.95091 72.20836 87.14326 101.1252 107.0779 109.3931 73.20718 88.14166 101.3792 107.1293 109.3931 Akaike Information Criteria by Rank (rows) and Model (columns) -4.149111 -4.853694 -4.782664 -4.304232 -3.567070 -4.149111 -5.297461 -5.260433 -4.990202 -4.376277 -4.972225 -5.593692 -5.600134 -5.085003 -4.376277 -4.972225 -5.537454 -6.011925* -5.721705 -5.085059 -4.686398 -5.346825 -5.845641 -5.631362 -5.085059 Schwarz 82 Criteria by Rank (rows) and Model (columns) -3.353284 -3.659954 -3.191011 -2.314665 -1.179590 Ph l c 7: Ki -3.353284 -4.053982 -3.569302 -2.851418 -1.789840 -3.977442 -4.200996* -3.809524 -2.896480 -1.789840 -3.977442 -4.095018 -4.121837 -3.383964 -2.299666 -3.492658 -3.755171 -3.856075 -3.243882 -2.299666 nh Johansen: Date: 08/08/13 Time: 16:34 Sample (adjusted): 1992 2012 Included observations: 21 after adjustments Trend assumption: Linear deterministic trend Series: LFD LGE LINF LMS Lags interval (in first differences): to Unrestricted Cointegration Rank Test (Trace) Hypothesized No of CE(s) Eigenvalue Trace Statistic 0.05 Critical Value Prob.** None * At most At most At most 0.749268 0.536221 0.218683 0.051789 51.48509 22.43428 6.299003 1.116744 47.85613 29.79707 15.49471 3.841466 0.0219 0.2750 0.6602 0.2906 Trace test indicates cointegrating eqn(s) at the 0.05 level * denotes rejection of the hypothesis at the 0.05 level **MacKinnon-Haug-Michelis (1999) p-values Unrestricted Cointegration Rank Test (Maximum Eigenvalue) Hypothesized No of CE(s) Eigenvalue Max-Eigen Statistic 0.05 Critical Value Prob.** None * At most At most At most 0.749268 0.536221 0.218683 0.051789 29.05081 16.13527 5.182259 1.116744 27.58434 21.13162 14.26460 3.841466 0.0322 0.2171 0.7187 0.2906 Max-eigenvalue test indicates cointegrating eqn(s) at the 0.05 level * denotes rejection of the hypothesis at the 0.05 level **MacKinnon-Haug-Michelis (1999) p-values Unrestricted Cointegrating Coefficients (normalized by b'*S11*b=I): LFD -0.976037 5.038900 1.603634 0.754699 LGE -4.833094 2.589505 -9.799628 -4.420177 LINF -7.454596 -6.703864 7.270949 0.677792 LMS 4.556475 -4.870822 4.249464 2.950625 83 Unrestricted Adjustment Coefficients (alpha): D(LFD) D(LGE) D(LINF) D(LMS) 0.068406 0.041860 0.054205 0.016486 Cointegrating Equation(s): -0.187639 -0.005162 0.055598 -0.013026 -0.125377 0.002096 -0.027512 0.008118 Log likelihood 86.73377 Normalized cointegrating coefficients (standard error in parentheses) LFD LGE LINF LMS 1.000000 -4.951751 -7.637612 4.668340 (1.84639) (1.85457) (1.16294) Adjustment coefficients (standard error in parentheses) D(LFD) -0.066767 (0.09679) D(LGE) -0.040857 (0.00677) D(LINF) -0.052906 (0.02548) D(LMS) -0.016091 (0.01699) Cointegrating Equation(s): Log likelihood 94.80141 Normalized cointegrating coefficients (standard error in parentheses) LFD LGE LINF LMS 1.000000 0.000000 -2.368920 -0.537987 (0.62707) (0.07679) 0.000000 1.000000 2.020807 -0.834120 (0.40407) (0.04948) Adjustment coefficients (standard error in parentheses) D(LFD) -1.012258 -0.816503 (0.44410) (0.47443) D(LGE) -0.066865 -0.215679 (0.03495) (0.03734) D(LINF) 0.227247 -0.118006 (0.11189) (0.11954) D(LMS) -0.081727 -0.113409 (0.08768) (0.09367) Cointegrating Equation(s): Log likelihood 97.39254 Normalized cointegrating coefficients (standard error in parentheses) LFD LGE LINF LMS 1.000000 0.000000 0.000000 -0.772928 (0.04026) 0.000000 1.000000 0.000000 -0.633704 (0.02477) 0.000000 0.000000 1.000000 -0.099176 (0.01523) -0.020859 -0.001852 -0.000762 -0.014100 84 Adjustment coefficients (standard error in parentheses) D(LFD) -1.213317 0.412141 (0.43148) (0.90106) D(LGE) -0.063504 -0.236219 (0.03651) (0.07623) D(LINF) 0.183128 0.151599 (0.11083) (0.23145) D(LMS) -0.068709 -0.192962 (0.09117) (0.19039) Ph l c 8: Ki -0.163642 (0.99377) -0.262206 (0.08408) -0.976835 (0.25526) 0.023451 (0.20998) nh VECM: Vector Error Correction Estimates Date: 08/08/13 Time: 16:37 Sample (adjusted): 1992 2012 Included observations: 21 after adjustments Standard errors in ( ) & t-statistics in [ ] Cointegrating Eq: CointEq1 LFD(-1) 1.000000 LGE(-1) -4.951751 (1.84639) [ 2.68186] LINF(-1) -7.637612 (1.85457) [ 4.11827] LMS(-1) 4.668340 (1.16294) [-4.01424] C 43.41242 Error Correction: D(LFD) D(LGE) D(LINF) D(LMS) CointEq1 -0.066767 (0.09679) [-0.68981] -0.040857 (0.00677) [-6.03203] -0.052906 (0.02548) [-2.07660] -0.016091 (0.01699) [-0.94684] D(LFD(-1)) -0.572002 (0.27996) [-2.04317] -0.056694 (0.01959) [-2.89382] 0.072986 (0.07369) [ 0.99043] 0.036534 (0.04916) [ 0.74324] D(LGE(-1)) 1.832213 (1.26183) [ 1.45203] 0.331529 (0.08830) [ 3.75449] -0.309360 (0.33214) [-0.93141] -0.229683 (0.22155) [-1.03669] D(LINF(-1)) -0.680892 (0.75250) [-0.90484] 0.104337 (0.05266) [ 1.98136] 0.033107 (0.19807) [ 0.16714] -0.108624 (0.13212) [-0.82213] 85 D(LMS(-1)) -0.650715 (1.69716) [-0.38341] 0.117722 (0.11877) [ 0.99120] -0.168322 (0.44673) [-0.37679] -0.009458 (0.29799) [-0.03174] C 0.194555 (0.59318) [ 0.32799] 0.105576 (0.04151) [ 2.54337] 0.138047 (0.15614) [ 0.88413] 0.301597 (0.10415) [ 2.89576] 0.456790 0.275720 3.097683 0.454436 2.522722 -9.702094 1.495438 1.793873 0.254274 0.533973 0.963592 0.951456 0.015170 0.031801 79.39931 46.14851 -3.823668 -3.525233 0.205972 0.144336 0.401789 0.202386 0.214626 0.119618 2.014955 18.32779 -1.174075 -0.875640 0.042789 0.133937 0.183448 -0.088737 0.095498 0.079790 0.673984 26.83064 -1.983870 -1.685435 0.247914 0.076470 R-squared Adj R-squared Sum sq resids S.E equation F-statistic Log likelihood Akaike AIC Schwarz SC Mean dependent S.D dependent Determinant resid covariance (dof adj.) Determinant resid covariance Log likelihood Akaike information criterion Schwarz criterion Ph l 1.17E-08 3.04E-09 86.73377 -5.593692 -4.200996 ng mô hình ng n h n: Dependent Variable: DLFD Method: Least Squares Date: 09/06/13 Time: 15:18 Sample (adjusted): 1992 2011 Included observations: 20 after adjustments Variable Coefficient Std Error t-Statistic Prob C DLFD(1) DLGE(1) DLINF(1) DLMS(1) ECT(-1) 0.064349 -0.545045 -0.024880 0.445420 1.006840 -0.082326 0.501812 0.286980 2.259459 0.992648 1.494894 0.054348 0.128233 -1.899247 -0.011012 0.448719 0.673519 -1.514775 0.8998 0.0783 0.9914 0.6605 0.5116 0.0862 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.398715 0.183970 0.494830 3.427990 -10.74119 1.856690 0.166226 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat 0.252419 0.547775 1.674119 1.972838 1.732432 2.205436 86 Ph l c 10: Ki a ph : Breusch-Godfrey Serial Correlation LM Test: F-statistic Obs*R-squared 0.178524 0.264414 Prob F(1,14) Prob Chi-Square(1) 0.6791 0.6071 Test Equation: Dependent Variable: RESID Method: Least Squares Date: 09/10/13 Time: 03:15 Sample: 1992 2012 Included observations: 21 Presample missing value lagged residuals set to zero Variable Coefficient Std Error t-Statistic Prob C(1) C(2) C(3) C(4) C(5) C(6) RESID(-1) 0.001429 0.142161 -0.305202 0.164913 0.299611 -0.083336 -0.210921 0.049928 0.405090 1.273612 0.877118 1.394288 1.007095 0.499196 0.028624 0.350936 -0.239635 0.188017 0.214885 -0.082749 -0.422521 0.9776 0.7309 0.8141 0.8536 0.8330 0.9352 0.6791 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.012591 -0.410584 0.474463 3.151605 -9.883299 0.029754 0.999842 Ph l c 11: Ki Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat 1.45E-16 0.399487 1.607933 1.956107 1.683496 2.027540 ng n c a ph : Heteroskedasticity Test: Breusch-Pagan-Godfrey F-statistic Obs*R-squared Scaled explained SS 2.456176 13.03777 8.374814 Prob F(8,12) Prob Chi-Square(8) Prob Chi-Square(8) 0.0779 0.1106 0.3977 Test Equation: Dependent Variable: RESID^2 Method: Least Squares Date: 09/10/13 Time: 03:32 Sample: 1992 2012 Included observations: 21 Variable Coefficient Std Error t-Statistic Prob C LFD(-1) -2.531917 -0.413664 2.199196 0.141464 -1.151292 -2.924162 0.2720 0.0127 87 LGE(-1) LINF(-1) LMS(-1) LFD(-2) LGE(-2) LINF(-2) LMS(-2) 1.007037 -0.009368 -0.433935 -0.039771 0.303097 -0.281806 -0.132167 R-squared Adjusted R-squared S.E of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) Ph l c 12: Ki 0.620846 0.368077 0.196459 0.463153 10.25161 2.456176 0.077900 0.982658 0.520507 0.617204 0.189161 0.679265 0.400590 0.766864 1.024808 -0.017998 -0.703066 -0.210248 0.446213 -0.703476 -0.172348 Mean dependent var S.D dependent var Akaike info criterion Schwarz criterion Hannan-Quinn criter Durbin-Watson stat 0.3257 0.9859 0.4954 0.8370 0.6634 0.4952 0.8660 0.151990 0.247138 -0.119201 0.328451 -0.022049 2.407222 nh quan h nhân qu Granger: Pairwise Granger Causality Tests Date: 08/11/13 Time: 17:59 Sample: 1990 2012 Lags: Null Hypothesis: Obs F-Statistic Prob D(LGE) does not Granger Cause D(LFD) D(LFD) does not Granger Cause D(LGE) 21 5.68131 25.1655 0.0284 9.E-05 D(LINF) does not Granger Cause D(LFD) D(LFD) does not Granger Cause D(LINF) 21 0.10782 0.28245 0.7464 0.6016 D(LMS) does not Granger Cause D(LFD) D(LFD) does not Granger Cause D(LMS) 21 0.06438 0.00017 0.8026 0.9898 D(LINF) does not Granger Cause D(LGE) D(LGE) does not Granger Cause D(LINF) 21 9.88225 0.00167 0.0056 0.9678 D(LMS) does not Granger Cause D(LGE) D(LGE) does not Granger Cause D(LMS) 21 21.1919 0.96061 0.0002 0.3400 D(LMS) does not Granger Cause D(LINF) D(LINF) does not Granger Cause D(LMS) 21 3.48638 1.74312 0.0782 0.2033