Sie sind auf Seite 1von 4

tsset time ** Stationarity of fdi flows scatter fdi time gen dfdi = fdi - fdi[_n-1] gen fdi_lag1 = fdi[_n-1]

reg dfdi fdi_lag1 predict e, residuals ac e drop e ** Since there is no autocorrelation in the residuals, no lags are used in the A ugmented Dickey-Fuller test dfuller fdi, trend lags(0) ** Since p-value is 0.9962 > 10%, we fail to reject the possibility of nonstatio narity of this series scatter dfdi time gen ddfdi = dfdi - dfdi[_n-1] gen dfdi_lag1 = dfdi[_n-1] reg ddfdi dfdi_lag1 time predict e, residuals ac e drop e ** Since there is no autocorrelation in the residuals, no lags are used in the A ugmented Dickey-Fuller test dfuller dfdi, trend lags(0) ** Since p-value is 0.3030 > 10%, we fail to reject the possibility of nonstatio narity of this series scatter ddfdi time gen dddfdi = ddfdi - ddfdi[_n-1] gen ddfdi_lag1 = ddfdi[_n-1] reg dddfdi ddfdi_lag1, noconstant predict e, residuals ac e drop e ** Since there is no autocorrelation in the residuals, no lags are used in the A ugmented Dickey-Fuller test dfuller ddfdi, noconstant lags(0) ** Since test stat is -3.276 < -1.600, we reject the possibility of nonstationar ity of this series ** Stationarity of exchange rate volitility scatter ervol time gen dervol = ervol - ervol[_n-1] gen ervol_lag1 = ervol[_n-1] reg dervol ervol_lag1 predict e, residuals ac e drop e ** Since there is no autocorrelation in the residuals, no lags are used in the A ugmented Dickey-Fuller test dfuller ervol, drift lags(0) ** Since p-value is 0.0011 < 10%, we reject the possibility of nonstationarity o f this series

** Stationarity of inflation scatter inflation time gen dinflation = inflation - inflation[_n-1] gen inflation_lag1 = inflation[_n-1] reg dinflation inflation_lag1 predict e, residuals ac e drop e ** Since there is no autocorrelation in the residuals, no lags are used in the A ugmented Dickey-Fuller test dfuller inflation, drift lags(0) ** Since p-value is 0.1478 > 10%, we fail to reject the possibility of nonstatio narity of this series scatter dinflation time gen ddinflation = dinflation - dinflation[_n-1] gen dinflation_lag1 = dinflation[_n-1] reg ddinflation dinflation_lag1, noconstant predict e, residuals ac e drop e ** Since there is no autocorrelation in the residuals, no lags are used in the A ugmented Dickey-Fuller test dfuller dinflation, noconstant lags(0) ** Since test stat is -3.020 < -1.600, we reject the possibility of nonstationar ity of this series ** Stationarity of exchange rate scatter exchangerate time gen dex = exchangerate - exchangerate[_n-1] gen ex_lag1 = exchangerate[_n-1] reg dex ex_lag1 time predict e, residuals ac e drop e ** There is autocorrelation, so we increase lags of dependent variable on right hand side gen dex_lag1 = dex[_n-1] reg dex ex_lag1 dex_lag* time predict e, residuals ac e drop e ** There is autocorrelation, so we increase lags of dependent variable on right hand side gen dex_lag2 = dex[_n-2] reg dex ex_lag1 dex_lag* time predict e, residuals ac e drop e ** There is autocorrelation, so we increase lags of dependent variable on right hand side gen dex_lag3 = dex[_n-3] reg dex ex_lag1 dex_lag* time predict e, residuals ac e drop e

** There is autocorrelation, so we increase lags of dependent variable on right hand side gen dex_lag4 = dex[_n-4] reg dex ex_lag1 dex_lag* time predict e, residuals ac e drop e ** There is autocorrelation, so we increase lags of dependent variable on right hand side gen dex_lag5 = dex[_n-5] reg dex ex_lag1 dex_lag* time predict e, residuals ac e drop e ** There is autocorrelation, so we increase lags of dependent variable on right hand side gen dex_lag6 = dex[_n-6] reg dex ex_lag1 dex_lag* time predict e, residuals ac e drop e ** Since there is no autocorrelation in the residuals now, 6 lags are used in th e Augmented Dickey-Fuller test dfuller exchangerate, lags(6) trend ** Since p-value is 0.7027 > 10%, we fail to reject the possibility of nonstatio narity of this series scatter dex time gen ddex = dex - dex_lag1 reg ddex dex_lag1 predict e, residuals ac e drop e ** There is autocorrelation, hand side gen ddex_lag1 = ddex[_n-1] reg ddex dex_lag1 ddex_lag* predict e, residuals ac e drop e ** There is autocorrelation, hand side gen ddex_lag2 = ddex[_n-2] reg ddex dex_lag1 ddex_lag* predict e, residuals ac e drop e ** There is autocorrelation, hand side gen ddex_lag3 = ddex[_n-3] reg ddex dex_lag1 ddex_lag* predict e, residuals ac e drop e ** There is autocorrelation, hand side gen ddex_lag4 = ddex[_n-4] reg ddex dex_lag1 ddex_lag* predict e, residuals ac e

so we increase lags of dependent variable on right

so we increase lags of dependent variable on right

so we increase lags of dependent variable on right

so we increase lags of dependent variable on right

drop e ** There is autocorrelation, so we increase lags of dependent variable on right hand side gen ddex_lag5 = ddex[_n-5] reg ddex dex_lag1 ddex_lag* predict e, residuals ac e drop e ** Since there is no autocorrelation now in the residuals, 5 lags are used in th e Augmented Dickey-Fuller test dfuller dex, drift lags(5) ** Since p-value is 0.0010 < 10%, we reject the possibility of nonstationarity o f this series ** Stationarity of GDP Growth scatter gdpgrowth time gen dgdp = gdpgrowth - gdpgrowth[_n-1] gen gdp_lag1 = gdpgrowth[_n-1] reg dgdp gdp_lag1 predict e, residuals ac e drop e ** Since there is no autocorrelation in the residuals, 0 lags are used in the Au gmented Dickey-Fuller test dfuller gdpgrowth, drift lags(0) ** Since p-value is 0.0032 < 10%, we reject the possibility of nonstationarity o f this series ** Model reg ddfdi ervol dex dex_lag1 dex_lag2 dex_lag3 dex_lag4 dex_lag5 dex_lag6 dex_la g7 estat hettest ** Hence, there is no heteroskedasticity estat ovtest ** Hence, there is no functional form misspecification estat durbinalt, robust ** Hence, there is no autocorrelation estat archlm ** Hence, there are no arch effects test dex + dex_lag1 + dex_lag2 + dex_lag3 + dex_lag4 + dex_lag5 + dex_lag6 + dex _lag7 == 0 ** This suggests that the long run multiplier is not statistically different fro m 0 test dex_lag2 + dex_lag3 + dex_lag5 + dex_lag7 == 0 ** Only using the significant coefficients shows the robustness of the above res ult ** Hence, while there are short term fluctuations in the flow of fdi due to chan ges in exchange rate, there is no long term effect.

Das könnte Ihnen auch gefallen