Ch08: Characterizing a Relationship Using SAS

Chapter Summary:

This chapter cover the applied side of the econometrics, with the application of SAS. Along with the application of econometrics techniques, the author also suggested some important tips to determine a statistical relationship between variables of interest.

If the assumption of stationary full fill, then an analyst can start with a simple econometric analysis by estimating correlation and regression analysis for the underlying variables. On the other hand, if the assumption of stationary not full fill, then the next step is, co-integration and ECM tests should be utilized to determine a statistical relationship between variables of interest. If the data in hand is nonstationary then the regression analysis will gives us spurious results. Most of the series in the economics are time trended, this time trend in the variables will leads to spurious regression (relationship between variables is not actual rather it is due to time trend). Co-integration and ECM provide reliable results, when the series are non-stationary. For decision makers, it is important to determine the direction of the relationship, particularly which variable is leading and which one is lagging, i.e. ARIMA models. Financial data series mostly suffer with the volatility cluster. To obtain meaningful results, we suggest utilizing the ARCH approach.

The final point we stress here is that the decision of what econometrics techniques to utilize is dependent on the objective of the study. Therefore, familiarity with all of these techniques would be helpful in determining which technique is more suitable to fulfill the requirement.