VECM
Vector Error Correction Model is a specialized type of VAR used when multiple non-stationary time series are cointegrated.
What is VECM?
When time series are non-stationary (have unit roots) but share a long-term equilibrium relationship (they are cointegrated), a VECM is used. It models the short-term dynamic changes in the variables while simultaneously ensuring they correct deviations from their shared long-term equilibrium trend.
Why VECM Matters
It allows researchers to analyze both the short-term fluctuations and the long-term relationships of macroeconomic variables, avoiding the loss of long-run information that happens when simply differencing data for a standard VAR.
Example
An energy economist models the relationship between crude oil prices and gasoline prices. Because they move together in the long run but fluctuate separately in the short run, a VECM captures how quickly gasoline prices 'correct' after a shock to crude oil.
Common Mistakes
- Applying a VECM when the variables are not actually cointegrated.
- Misinterpreting the error correction term; it must be negative to imply the system is returning to equilibrium.