📉What is econometrics?

Before we discuss some of the concepts of last page in more detail, it is important to define what econometrics is. There are various definitions out there, but generally econometrics is concerned with applying statistical and mathematical methods to economic and financial data. An econometrician might have multiple reasons in mind to do this. Let's discuss a few:

• Learning about relationships between variables. You might have the suspicion that different data series are related (e.g. by looking at their graphs), but that does not tell you exactly how. Econometricians build models to quantify relationships. In this way, you are able to make statements such as "If variable X goes up by 1%, then the expected increase in Y is ...%". In this way, we learn something the economy and policymakers are able to use this information to respond to changes appropriately.

• Predicting (future) values of variables. A model may not only be used for descriptive purposes as in the last bullet point, but also to learn about the future. Is inflation going to increase further? Or is it finally leveling off? Given that we have built a model, we can ask ourselves: what would be the most likely value for inflation next month, given the values of inflation (and possibly other related variables, such as the unemployment rate for example) we observed in the (recent) past? We call this forecasting.

• Testing existing economic and financial theories. According to Okun's Law (a well-known theory in macroeconomics), economic growth should be negatively related to the unemployment rate. Macroeconomics might argue that a high unemployment rate leads to less bargaining power for new workers (there are so many unemployed people, you are easily replaced!) and thus lower salaries. This leads to lower economic growth. Using econometrics, we can test whether these theories hold on real data, how strong these relationships are and whether they have remained stable over time.

We could come up with even more applications of econometrics, but it is clear that data plays an important role. There are however many different types of data:

• Data might be observed over time (e.g. monthly inflation in the Netherlands from 2000 until 2023). We call this time series data.

• Data can be observed at one point in time but from different respondents (e.g. respondents filling in a questionnaire about their customer satisfaction). We call this cross-sectional data.

• We could also have a combination of both (e.g. a supermarket asking customers to fill in a questionnaire every year). We call this panel data.

We can also have certain complicated features of data (or problems within a data set). Just to name a few:

• Data can be observed during a fixed recurrent period (e.g. every month or every quarter), but can also be measured irregularly (e.g. Google trend data).

• Data can have missing entries or breaks.

• Data can be recorded in the wrong format (e.g. text).

In short: an econometrician might face various challenges! To be able to deal with these, it is important to have a solid background in mathematics and statistics. The list of concepts mostly contains topics in statistics, as they are vital for a first course in econometrics. However, a basic background in especially linear algebra and analysis is desirable to fully understand most of the concepts.

The next pages explain informally why these topics are important to an econometrician. For a more formal treatment, please find appropriate literature here.