In short run, the foreign Inflation and Interest rate changes are affecting future change in Exchange Rate significantly. Domestic and Foreign Inflation is effecting the future Domestic Inflation. The foreign Inflation is only caused by past foreign Inflation and Interest Rate change. The domestic interest rate change was only the reason to the future domestic interest rate change and the past foreign Inflation and Interest Rate change is causing the change in foreign Interest Rate.

The VECM stability test in showed that the equilibrium is not stable enough that is why it takes more time than usual to return back to equilibrium after a shock. The LM residual test showed no sign of autocorrelation among the residuals from 1st to 4th order hence the results are considered valid. The normality test showed for some of the variables to be different from the skew-ness of zero and kurtosis to 3 hinting presence of few extreme values present in the data.

5.9 Testing Parities

The restrictions will be applied on the Long Run Vector [S, P, P*, i, i*] to see some specific parities with its specifications given in this vector form. For K variables and l co-integrating equations, there must me K-I eigenvalues real and near to unity, if any other is near to one, it will suggest that either the equilibrium pattern is non-stationary or the equilibriums are lower than the actual. That is the reason why the convergence speed is slow. Stata Documentation. Four restrictions are only used to purify the relationship and testing the significance of the desired outcome. P-Value higher than 0.05 suggests that the restrictions applied is statistically acceptable at 5%. This is the specification of the restriction applied; the signs of the variables other than Exchange Rate are actually opposite to what it is written here in the restriction. Here it is corrected to show the desired direction. This restriction is theoretically fine until a > 0. Both hypothesis 6 and 7 are technically same in the later one restriction is applied on the unity of the Exchange Rate instead of the Domestic Price Level, the results are same the Parameter of the Domestic Price Level is still statistically near to unity but this improved the results.

(MacDonald & Taylor, 1991) and (Juselius & MacDonald, 2004) used series of test restrictions to see the validity of the restrictions and to see it the parameters are near to the mathematical model. This paper uses this idea and tests some restrictions for the parities on the long run equilibrium to see if they hold or not.

**Table 12. Testing parities**

Hypothesis | Pt-i | P*t-i | it-i | i*t-1 | Chi-Square(v) | P-Value |

1 | Purchasing Power Parity [S,P,P*,i, i*] = ~ I | |||||

1 | -1 | 0 | 0 | 10.07 | 0.019 | |

2 | Uncovered Interest Parity [S,P,P*,i, i*] = ~ I | |||||

0 | 0 | 1 | -1 | 10.08 | 0.018 | |

3 | External Economy Effect [S,P,P*,i, i*] = ~ I | |||||

0 | -1 | 0 | -1 | 14.35 | 0.002 | |

4 | Internal | Economy Effect [S,P,P*,i, i*] = ~ I | ||||

1 | 0 | 1 | 0 | 9.305 | 0.025 | |

5 | Strict PPP and Loose UIP [S,P,P*,i, i*] = [1,1,-1,n,m] ~ I | |||||

1 | -1 | 0.15 | -0.66 | 0.7954 | 0.672 | |

6 | Strict PPP and UIP (I) | [S, P, P*, i, i*] = [a,1,-1,1,-1] ~ I | ||||

1 | -1 | 1 | -1 | 7.466 | 0.058 | |

7 | Strict PPP and UIP (II) [S, P, P*, i, i*] = [1,b,-1,1,-1] | ~ I | ||||

1.37 | – 1 | 1 | -1 | 6.339 | 0.096 | |

Bold parameters are statistically significant at 5% |

In short run, the foreign Inflation and Interest rate changes are affecting future change in Exchange Rate significantly. Domestic and Foreign Inflation is effecting the future Domestic Inflation. The foreign Inflation is only caused by past foreign Inflation and Interest Rate change. The domestic interest rate change was only the reason to the future domestic interest rate change and the past foreign Inflation and Interest Rate change is causing the change in foreign Interest Rate. The VECM stability test in showed that the equilibrium is not stable enough that is why it takes more time than usual to return back to equilibrium after a shock. The LM residual test showed no sign of autocorrelation among the residuals from 1st to