Characterizing a Relationship Using SAS
A question that is essential to analysts in the economic and financial world is: To what extent are certain variables related to each other? This chapter provides methods for using SAS software to determine the statistical relationship between economic and financial variables and to interpret the results. First we share useful tips for an applied time series analysis and then we examine how to estimate the statistical relationship between two (or more) variables.
USEFUL TIPS FOR AN APPLIED TIME SERIES ANALYSIS
Helpful hints are the reliable servant of any good work. An analyst needs to be familiar with a few key guidelines for applied research. The guidelines include using economic and financial theories as a benchmark and testing the theory with econometric techniques by employing time series data (or cross-section/panel data).1 Most applied econometric analysis is now done by statistical software, which produces results quickly. Such speed can be a trap, however, because the software does not care what kind of data is inputted and whether it is the best basis for the statistical results.
For example, using any statistical software, an analyst can produce correlation coefficients between the population in Gambia and the gross domestic product (GDP) of the United States.2 The correlation coefficient may be statistically significant, yet any correlation is likely to reflect a third factor: a deterministic time trend (i.e., an upward ...