Câmbio real e inovação tecnológica: evidências empíricas

SOUTO, KEYNIS CÂNDIDO DE; RESENDE, MARCO FLÁVIO CUNHA

doi:10.1590/0101-31572018v38n02a04

KEYNIS CÂNDIDO DE SOUTO

Universidade Federal Rural de Pernambuco (Brazil). E-mail: keynis.souto@ufrpe.br.

MARCO FLÁVIO CUNHA RESENDE

Universidade Federal de Minas Gerais (Brazil). E-mail: resende@cedeplar.ufmg.br

Tabelas (4)
Fórmulas (10)

Thumbnail

Table 1
Dynamic Model for Innovation (Equation 6)

Thumbnail

Tabela 2
Dynamic Model for Innovation with Dummies

Thumbnail

CHOW: H0: Restricted model (Pooled) H1: Unrestricted model (fixed effects) F (70, 163) = 3.48 Prob > F = 0.0000 Rejects H0: Panel data model is more appropriate. BREUSCH-PAGAN (LM TEST): H0: Pooled Model HA: Random Effect Model Chibar2(01) = 3.77 Prob > chibar2 = 0.0261 Rejects H0: Panel data model is more appropriate. HAUSMAN TEST: H0: There is no correlation between αi and the regressors Xit’s - both models are consistent, but the fixed effects model is less efficient HA: There is correlation - both are consistent, but fixed effects one is more efficient Chi2(6) = 140.74 Prob > chi2 = 0.0000 Rejects H0: Dynamic model of EF for innovation is more efficient Fonte: Own elaboration based in the regressions from stata 12.0.

Thumbnail

Explicative GMM-SYS two step GMM-SYS two step GMM-SYS two step (VPS3) (VDP5) (VDP3) β Robust SD p-value β Robust SD p-value β SD Robust p-value int -1 0,4620 0,2223 0,041** 0,4409 0,2295 0,059*** 0,4978 0,2703 0,070*** inv 0,3097 0,3264 0,346 0,2514 0,2436 0,063*** 0,2654 0,2589 0,309 educ 0,5748 0,3498 0,105 0,4818 0,3583 0,183 0,3432 0,4255 0,423 p&d 0,8820 0,4335 0,046*** 0,9413 0,4092 0,024** 0,8738 0,4828 0,075*** dcr 0,3878 0,2099 0,069*** 0,4657 0,2293 0,046** 0,2977 0,1645 0,075*** vcr -0,6227 0,2802 0,029** -3,4855 1,5210 0,025** -0,3895 0,2116 0,070*** Instrument Validity Tests (Significance Level of 5%) Arellano-Bond p-value p-value p-value frmH: Serial non-correlation AR(1) 0,002 0,019 0,010 frmI: Serial non-correlation AR(2) 0,095 0,486 0,216 Sargan 0,015 0,001 0,001 H0: non-orthogonal additional instruments Hansen 0,061 0,063 0,057 H0: Orthogonal instruments Difference-Hansen H0: All valid instruments 0,089 0,109 0,065 H0: Additional exogenous instruments 0,087 0,072 0,277 Countries 76 Observations 380 Instruments 22 Fonte: Own elaboration based in the regressions. Notes: * significant to 1%, ** significant to 5%, *** significant to 10%. The constant was not presented because it was insignificant.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(6’)

Explanatory variables	GMM-SYS two step
Explanatory variables	Coeficients	robust-SD	p-value
in _t-1	frmA= 0,6038	0,1424	0,000*
inv	frmB = 0,4486	0,2400	0,066***
educ	frmC = 0,3076	0,2244	0,175
r&d	frmD = 0,6889	0,2949	0,022**
rer	frmE = 0,2711	0,1367	0,051***
vrer(vps5)	frmF = - 0,4743	0,9880	0,000*
Constant	frmG = 1,4594	1,1143	0,195
Instrument Validation Tests (significance level of 5%)
Arellano-Bond Test		p-value
frmH: absence of serial correlation AR(1)		0,001
frmI: absence of serial correlation AR(2)		0,101
Sargan Test		0,019
H₀: non-orthogonal additional instruments
Hansen Test		0,139
H₀: orthogonal instruments
Difference-Hansen Test
H₀: all the instruments are valid		0,371
H₀: additional instruments are exogenous		0,80
Dynamic panel bias elimination test (estimates for γ)
Within	GMM-SYS		OLS
0,1663	0,6038		0,9316
(p-value = 0,016)	(p-value = 0,000)		(p-value = 0,000)
Countries		76
Observations		380
Instruments		22

Explanatory Variables	GMM-SYS two step
Explanatory Variables	Coeficients	Robust-SD	p-value
int-1	frmA= 0,5404	0,1849	0,005*
Inv	frmB = 0,5706	0,2931	0,056***
Educ	frmC = 0,1761	0,2418	0,469
r&d	frmD = 0,6864	0,3128	0,031**
rer	frmE = 0,4884	0,1443	0,736
vrer ¹	frmF = - 0,4916	0,0975	0,000*
(rer x NIS)	frmJ = 0,5040	0,2215	0,026**
(vrer x NIS)	frmJ = 0,0629	0,0335	0,0970***
Constant	frmG = 1,6758	1,0887	0,128
Instrument Validation Tests (significance level of 5%)
Arellano-Bond Test		p-value
frmH: absence of serial correlation AR(1)		0,002
frmI: absence of serial correlation AR(2)		0,099
Sargan Test		0,026
H₀: non-orthogonal additional instruments
Hansen Test		0,199
H₀: orthogonal instruments
Difference-Hansen Test
H₀: all the instruments are valid		0,418
H₀: additional instruments are exogenous		0,116
Dynamic panel bias elimination test (estimates for γ)
Within	GMM-SYS		OLS
0,1517	0,5404		0,9382
(p-value = 0,027)	(p-value = 0,005)		(p-value = 0,000)
Countries		76
Observations		380
Instruments		24

Explicative	GMM-SYS two step			GMM-SYS two step			GMM-SYS two step
	(VPS3)			(VDP5)			(VDP3)
	β	Robust SD	p-value	β	Robust SD	p-value	β	SD Robust	p-value
int _-1	0,4620	0,2223	0,041**	0,4409	0,2295	0,059***	0,4978	0,2703	0,070***
inv	0,3097	0,3264	0,346	0,2514	0,2436	0,063***	0,2654	0,2589	0,309
educ	0,5748	0,3498	0,105	0,4818	0,3583	0,183	0,3432	0,4255	0,423
p&d	0,8820	0,4335	0,046***	0,9413	0,4092	0,024**	0,8738	0,4828	0,075***
dcr	0,3878	0,2099	0,069***	0,4657	0,2293	0,046**	0,2977	0,1645	0,075***
vcr	-0,6227	0,2802	0,029**	-3,4855	1,5210	0,025**	-0,3895	0,2116	0,070***
Instrument Validity Tests (Significance Level of 5%)
Arellano-Bond	p-value			p-value			p-value
frmH: Serial non-correlation AR(1)	0,002			0,019			0,010
frmI: Serial non-correlation AR(2)	0,095			0,486			0,216
Sargan	0,015			0,001			0,001
H₀: non-orthogonal additional instruments	0,015			0,001			0,001
Hansen	0,061			0,063			0,057
H₀: Orthogonal instruments	0,061			0,063			0,057
Difference-Hansen
H₀: All valid instruments	0,089			0,109			0,065
H₀: Additional exogenous instruments	0,087			0,072			0,277
Countries						76
Observations						380
Instruments						22

[TFN1] Source: Own elaboration based in the regressions.

[TFN2] Notes: * significant at 1%, ** significant at 5%, *** significant at 10%.

CHOW:	H₀: Restricted model (Pooled)
	H₁: Unrestricted model (fixed effects)
F (70, 163) = 3.48	Prob > F = 0.0000	Rejects H₀: Panel data model is more appropriate.
BREUSCH-PAGAN (LM TEST):
H₀: Pooled Model
H_A: Random Effect Model
Chibar2(01) = 3.77	Prob > chibar2 = 0.0261	Rejects H₀: Panel data model is more appropriate.
HAUSMAN TEST:
H₀: There is no correlation between α_i and the regressors X_it’s - both models are consistent, but the fixed effects model is less efficient
H_A: There is correlation - both are consistent, but fixed effects one is more efficient
Chi2(6) = 140.74	Prob > chi2 = 0.0000	Rejects H₀: Dynamic model of EF for innovation is more efficient

Brasil

Brasil

Câmbio real e inovação tecnológica: evidências empíricas

RESUMO