Sei sulla pagina 1di 8

a) Let I = Real Investment (the dependent variable)

The independent variables are GDP = Real GNP R = Interest Rate INF = Inflation Rate T = Trend u = the error term Y = 0 + 1X1 + 2X2 + 3X3 + 4X4 + u A PRIORI EXPECTATION
PARAMET ER EXPECTE D SIGN + ECONOMIC INTERPRETATION We expect a positive relationship between the constant term and real investment. This is because holding all other factors other than real GDP, interest rate, inflation rate, and trend to cause investment in the economy. There is a positive relationship between real GDP and Investment. Theoretically, GDP is partly determined by the amount of investment. As investment increases, GDP also increases, that is GDP = C + I + G. Or I = GDP C G, as the economy grows, GDP rises, investment venture becomes productive and therefore attracts more investors into the economy and existing investors also increase their production. Hence a rise in GDP leads to an increase in investment. An increase in the inflation rate causes real investment to fall depicting a negative relationship. Inflatio n is a cost to investors or firms since nominal interest rate charged by lenders (banks) includes the inflation rate. As inflation rate falls, it is expected that the real interest rate will fall which will lead to a reduction in cost of borrowing thereby increasing real investment. Again, high fluctuations in inflation discourage investment since investors find it difficult to plan and try to cut down investment. Interest rate is the cost of investment thus the price investors pay to get the needed funds to produce as well as adding to capital and inventory. As the cost of borrowing goes up, real investment will fall because investors find it costly to borrow to undertake new investment projects. The relationship between investment and trend in indeterminate, thus it can be positive or negative. There exists a relationship between real investment and trend, it must be noted that correlation does not necessarily imply causality. If there exist in the economy favourable conditions like stable political environment, stability in the general price levels, technological progress, low business taxes, low interest rates etc then investors or firms would be encouraged to increase their investment spending over time which might not be due to time trend but favourable conditions. Firms may

+ or

reduce investment not because of time but because of unfavourable conditions.

After running the regression; I = -0.509 + 0.670GDP - 0.00233R 0.0000940INF - 0.0166T SE (0.0551) (0.0550) t* (-9.234) (12.19) (0.00122) (-1.908) (0.00135) (-0.06977) (0.00197) (-8.409)

Adjusted R2 = 0.961406 R2 = 0.972 DW *= 1.963642 t 0.025,10 = 2.228 F0.5, V1,V2 = 3.47805 F* = 88.18825

INTERPRETATION OF RESULTS The R2 of 0.972433 means approximately 98 percent of the variation in Real investment is explained by Real GNP, Interest rate, Inflation rate, and Trend. The rest are presented in the table below; PARAMETER 0 REGRESSION VALUE 0.509 INTERPRETATION The -0.509071 (0) is the intercept of the regression model. That is real investment will change by 0.509071 unit in opposite direction when T, G, R, and INF are held constant. This shows a negative relationship, but this is opposite to the a priori sign This value measures the percentage change in Real Investment with respect to real GDP when T, R, and INF are held constant. Positive relationship exists between Real Investment and GNP confirming the a priori sign. This implies that a unit change in real GNP will lead to 0.670 unit change in the same direction in investment holding T, R, and INF constant. There is a negative relationship between real investment and interest rate confirming the a priori sign. The 0.00233 unit measures the percentage change in Real Investment with respect to Interest rate when T, GNP, and INF are held constant. Thus a unit change in

+0.670

- 0.00233

interest rate will lead to 0.00233 unit change in real investment in opposite direction holding T, GNP, and INF constant. 3
0.0000940

As expected, there is a negative relationship exists between real investment. The 0.0000940 measures the percentage change in Real Investment with respect to Inflation rate when T, GNP, and R are held constant. Thus a unit change in inflation rate will lead to 0.0000940 unit change in investment in opposite direction holding T, GNP, and R constant. There is a negative relationship between real investment and trend confirming the a priori sign. The 0.0166 measures the percentage change in Real Investment with respect to Trend when GNP, R, and INF are held constant. Thus a unit change in Trend will lead to 0.0165804 unit change in opposite direction in real investment holding GNP, R, and INF constant.

- 0.0166

TESTING FOR OVERALL SIGNIFICANCE H0: 0 = 1 = 2 = 3 = 4 = 0 H1: Not all the s (betas) are equal to zero. F0.5,V1,V2 = 3.47805 F* = 88.188 Since F* > F0.5,V1,V2 (88.188 > 3.478) H0 is rejected and H1 is accepted. Therefore not all the betas are equal to zero. Therefore the coefficients (betas) are jointly significant. In other words the independent variables (T, GNP, R, and INF) are able to explain the dependent variable (I).

TESTING FOR SIGNIFICANCE OF INDIVIDUAL PARAMETERS (1) H0: 0 = 0 H1: 0 0 t0.025,10 = 2.228 Since t > t* thus -2.228 > -9.234, Ho is rejected and H1 is accepted. Therefore -0.509071 is statistically significant. Thus holding T, GDP, R, and INF constant, other factors which are not captured in this model/regression will make Real Investment to fall. This shows an inverse relationship.

(2) H0: 1 = 0 H1: 1 0 Since t > t* thus -2.228 > -8.4089, Ho is rejected and H1 is accepted. Therefore -0.01658 is statistically significant. Thus holding GDP, R, and INF constant, a unit change in Trend will cause real investment to change in opposite direction by 0.0165804. This shows an inverse relationship between Real Investment and Trend. (3) H0: 2 = 0 H1: 2 0 Since t < t* thus 2.228 < 12.189, Ho is rejected and H1 is accepted. Therefore 0.67038 is statistically significant. Thus a unit change in GDP, holding T, R, and INF constant, will cause real investment to change in the same direction by 0.6708. This shows a positive relationship between Real Investment and Real GDP. (4) H0: 3 = 0 H1: 3 0 Since t < t* thus -2.228 < -1.9083, Ho is accepted and H1 is rejected. Therefore 0.00233 is statistically insignificant. Thus there is insufficient evidence to claim that a unit change in R will cause real investment to change by 0.00233 in opposite direction. (5) H0: 4 = 0 H1: 4 0 Since t < t* thus -2.228 < -0.0698, Ho is accepted and H1 is rejected. Therefore 0.00004010 is statistically insignificant. Thus there is insufficient evidence to claim that a unit change in inflation rate will cause real investment to change by 0.00004010 in opposite direction.

CONDUCTING DW TEST

H0: p = 0 H1: p 0

5% critical values for Durbin-Watson statistic, n = 15, k = 4

dL = 0.6852 dU = 1.9774

4 dL = 3.3148 4 dU = 2.0226

(2) (i) RESULTS OF COCHRAN E ORCUTT (C O) INTERACTIVE PROCESS I = -0.516 + 0.682GDP - 0.00174R - 0.000615INF - 0.0175TREND SE (0.0466) (0.0467) T = 14, rho = -0.046 R2 = 0.979 DW*C-O = 2.477564 (0.00105) (0.00116) (0.00173)

(ii) RESULTS OF HILDRETH LU (H L) INTERACTIVE PROCESS

I = -0.516 + 0.682GDP - 0.00174R - 0.000615INF - 0.0175TREND Se (0.0466) (0.0467) (0.00105) (0.00116) (0.00173)

T = 14, rho = -0.047

R2 = 0.979 DW*H-L = 2.477314

(iii) PRAIS - WINSTEN TRANSFORMATION PROCEDURE

I = -0.502 + 0.665GDP - 0.00237R - 0.000218INF - 0.0163TREND SE (0.0520) (0.0519) (0.00111) (0.00129) (0.00188)

T = 15, rho = -0.130 R2 = 0.973 DW*P-W = 1.798094

(iv)

There is no autocorrelation.

APPENDIX
Model 1: OLS, using observations 1968-1982 (T = 15)

Dependent variable: INVT Const TREND GNP IR INFL Mean dependent var Sum squared resid R-squared F(4, 10) Log-likelihood Schwarz criterion Rho Coefficient Std. Error -0.509071 0.0551277 -0.0165804 0.00197176 0.670383 0.0549972 -0.00232593 0.00121887 -9.40107e-05 0.00134748 0.203333 0.000451 0.972433 88.18825 56.80975 -100.0793 -0.098137 t-ratio -9.2344 -8.4089 12.1894 -1.9083 -0.0698 p-value <0.00001 <0.00001 <0.00001 0.08545 0.94575 *** *** *** *

S.D. dependent var S.E. of regression Adjusted R-squared P-value(F) Akaike criterion Hannan-Quinn Durbin-Watson

0.034177 0.006714 0.961406 9.33e-08 -103.6195 -103.6572 1.963642

(i) COCHRANE ORCUTT (C O) INTERATIVE RESULTS

Model 1: Cochrane-Orcutt, using observations 1951-1964 (T = 14) Dependent variable: I const GDP R INF TREND Mean dependent var Sum squared resid R-squared F(4, 9) rho Coefficient Std. Error -0.516271 0.0465828 0.681586 0.0466684 -0.00174084 0.00104887 -0.000615123 0.00116138 -0.017478 0.00172705 t-ratio -11.0829 14.6049 -1.6597 -0.5296 -10.1201 p-value <0.00001 <0.00001 0.13134 0.60917 <0.00001 *** *** ***

Statistics based on the rho-differenced data: 0.206357 S.D. dependent var 0.000297 S.E. of regression 0.979436 Adjusted R-squared 115.1726 P-value(F) -0.292461 Durbin-Watson

0.033320 0.005743 0.970296 1.01e-07 2.477564

(ii) HILDRETH LU (H U) INTERACTIVE PROCESS

Model 2: Hildreth-Lu, using observations 1951-1964 (T = 14) Dependent variable: I

const GDP R INF TREND Mean dependent var Sum squared resid R-squared F(4, 9) rho

Coefficient Std. Error -0.516262 0.0465797 0.681577 0.0466653 -0.00174112 0.00104873 -0.000615151 0.00116132 -0.0174774 0.00172694

t-ratio -11.0834 14.6057 -1.6602 -0.5297 -10.1204

p-value <0.00001 <0.00001 0.13124 0.60914 <0.00001

*** *** ***

Statistics based on the rho-differenced data: 0.206357 S.D. dependent var 0.000297 S.E. of regression 0.979436 Adjusted R-squared 115.2054 P-value(F) -0.292363 Durbin-Watson

0.033320 0.005743 0.970296 1.01e-07 2.477314

(iii) PRAIS WINSTEN TRANSFORMATION PROCESURE

Model 3: Prais-Winsten, using observations 1950-1964 (T = 15) Dependent variable: I const GDP R INF TREND Coefficient Std. Error -0.502261 0.0520398 0.664649 0.0518993 -0.00237339 0.00110636 -0.000218259 0.00129206 -0.01635 0.00187975 t-ratio -9.6515 12.8065 -2.1452 -0.1689 -8.6980 p-value <0.00001 <0.00001 0.05753 0.86922 <0.00001 *** *** * ***

Statistics based on the rho-differenced data: Mean dependent var 0.203333 S.D. dependent var Sum squared resid 0.000443 S.E. of regression R-squared 0.972920 Adjusted R-squared F(4, 10) 124.6837 P-value(F) rho -0.018806 Durbin-Watson

0.034177 0.006657 0.962088 1.73e-08 1.798094

Potrebbero piacerti anche