ERROR CORRECTION MODEL Flashcards
To look at the long-run we used? But now to look at the short-run we use?
Upon establishment of cointegration, we assume an LR equilibrium relationship but to look at the short-run we use the error correction model.
What does the ECM do?
The ECM uses lagged error term in the short-run to account for any disequilibrium between Eyt* and Yt.
How do we capture deviations from E?
We use the error term
What is the dependent we observe and do not? What does the ECM do about it?
Yt* is the E value that is not observed
Yt is the observed value.
ECM captures the deviations between the two
What do Ut and Ut-1 do?
Capture the disequilibrium in time period t
Capture the disequilibrium in time period t-1
How is the LR equilibrium achieved with a short-run disequilibrium?
Treat error term Ut = ln(Ct) - B0 - B1lnYt as the equilibrium error term that corrects for deviations of ln(Ct) from its E value, this gives the cointegrating equation of Ln(Ct) = Bo + B1ln(Yt) + ut
In the short-run what can the changes in the dependent variable be due to?
- Changes in the X variables.
2. Disequilibrium between two periods which is captured by ut-1 (et-1).
What is the SR Granger Causality test? What is the regression that is run? How do u test it? What would a negative ECM coefficient in the LC mean?
The test helps us to find if LY is or is not affected by the disequilibrium in LC.
reg dlc emc1 dlc1 dlc2 dlc3 dlc4 dly dly1 dly2 dly3 dly4
Based on the p-values it is insignificant. However, the EMC in LC is -0.06. This means if there were any departures from the E in the previous period, this departure is reduced by 6% in the current period.
Why do we only interpret EMC in regression?
It gives the error correction interpretation from the last year’s error term to this current error term. Therefore the amount of disequilibrium that was causing changes in the dependent.
What is the short-run ECM equation?
Change in Ln(Ct) = A1 + A2Chnage in Ln(Yt) + A3ut-1 + vt.
Error correction model must include Ut-1 or et-1.
What does the short-run ECM equation show?
Changes in ln(Ct) depend on changes in ln(yt) and the lagged error term
What if the error term = 0 in the SR ECM equation?
then Ln(Ct) = B1 + B2ln(Yt) + Ut this the long-run E equsation which is the same as the original regression (the cointegrating regression).
Two things must include in the short-run ECM model?
Want to include triangle (change)
Want to include Ut-1.
