Abstract
Concurrent with Mamingi findings using Monte Carlo experiment (Economics
Letters, 1996, 52, 7-14), this paper analyses the existence of Granger causality distortion
in error correction models using the relationship between the theoretical cross
covariances of the basic and aggregated process. The analytical result confirms that the
causal distortion depends on the degree of cointegaration, the data span, the sample size
and the type of aggregation. The more prominent result is that the systematic sampling
preserves the causal direction among the variables while most of the causal distortion
occurs due to temporal aggregation. For the bivariate cointegarated system with no short
run dynamics, this paper develops the testing procedure to test for the long run Granger
causality between the variables in the basic disaggregated form using the temporally
aggregated data.
Letters, 1996, 52, 7-14), this paper analyses the existence of Granger causality distortion
in error correction models using the relationship between the theoretical cross
covariances of the basic and aggregated process. The analytical result confirms that the
causal distortion depends on the degree of cointegaration, the data span, the sample size
and the type of aggregation. The more prominent result is that the systematic sampling
preserves the causal direction among the variables while most of the causal distortion
occurs due to temporal aggregation. For the bivariate cointegarated system with no short
run dynamics, this paper develops the testing procedure to test for the long run Granger
causality between the variables in the basic disaggregated form using the temporally
aggregated data.
| Original language | English |
|---|---|
| Publication status | Published - 2001 |
| Externally published | Yes |
| Event | 12th (EC)2 Conference - Causality and Exogeneity in Econometrics: Causality and Exogeneity in Econometrics - Centre for Operation Research and Econometrics (CORE), Louvain-la-Neuve, Belgium Duration: 12 Dec 2001 → 15 Dec 2001 https://sites.google.com/site/ecpower2/conferences |
Conference
| Conference | 12th (EC)2 Conference - Causality and Exogeneity in Econometrics |
|---|---|
| Country/Territory | Belgium |
| City | Louvain-la-Neuve |
| Period | 12/12/01 → 15/12/01 |
| Internet address |