What Are the Different Methods of Econometrics?

Econometrics is the application of statistical analysis to economic data. The various methods of econometrics can be divided into two types: theoretical and applied. Roughly speaking, the former is based on testing whether theories work in a mathematical sense, while the latter tests whether the theories are borne out in the real world, as well as for forecasting.

Most methods of econometrics are simply variants of more general data analysis. Such analysis involves looking at collections of data and trying to both identify patterns and to identify how strong those patterns are and whether they could be caused by freak results. Some analysts will attempt to simply find patterns and then consider possible explanations, while others may start with a hypothesis and then look for data to bear it out.

Some methods of econometrics are purely theoretical. They generally involve looking at the techniques of gathering and analyzing data, rather than the data itself. For example, a theoretical econometrics project could involve looking at ways to improve the accuracy with which a survey sample group represents the entire population.

Other methods of econometrics are practical, known as applied methods, and work with real-life data. One use of such methods is to take an economic theory, such as that decreasing tax rates increases the total tax revenue, and seeing whether it works with real data. Another type of applied econometrics is to look at patterns and relationships that are shown by past data and then predict what would happen if those patterns continued in the future.

Such techniques are often extremely complex because every economic decision and action is often influenced by multiple factors. As a result, one of the most common econometrics techniques is regression analysis, which is a technique designed to isolate the effects of individual factors. For example, if an economist wasn't sure if it was income levels, local tax levels or mortgage rates that were causing a decrease in consumer spending, she would cross-reference the data to see what effect varying mortgage rates had on people who were on identical or very similar salaries and lived in areas with the same level of local taxes.

Economists are usually forced to use regression analysis because they cannot carry out controlled experiments as can be done in science. This means the quality of the analysis is often restricted by the availability of data. For example, a study of 3,000 people can be enough for the results to be considered statistically significant in representing the entire population. However, in the example above there may only be a couple of hundred people within the study who have similar income and local tax levels. This means that any conclusions about how mortgage rates affect their spending may have to be treated with more caution.