# Difference between econometrics and time series analysis

I'd like to know the difference between time series analysis and econometrics except the fact that the observations using in TSE are in time

• Econometrics overlaps with time series analysis in so far as econometrics often deals with time series. Note that there many kinds of time series analysis that see little or no application in econometrics. If you spelled out what is puzzling you, there might be a more penetrating answer. I don't think there is very much more unity to time series analysis than the name implies, e.g. people divide sharply on whether it's a series of small variations on themes by Box and Jenkins or there are much better approaches. Commented Nov 20, 2016 at 19:09
• It's like the difference between aspirin and medicine.
– Tim
Commented Feb 7, 2022 at 21:19
• Or the difference between singing and opera. Commented Feb 7, 2022 at 21:31
• everything above is true and insightful but I believe that a key difference arises when econometrics uses a "structural" relationship and the time-series approach uses a reduced form relationship. I don't want to try get into the difference between the two here but, to me, that's a key difference. There are others also of course. such as focus on error terms, instrumental variables, simultaneous equations, the use of expectations such as RE and others that I'm forgetting.. Commented Mar 18, 2022 at 12:13

Time series analysis refers to any analysis performed on a time series dataset (as opposed to on cross-sections, panels, and pooled cross sections). These analyses can be univariate (one time series) or multivariate (many time series).

Econometrics is the statistical analysis performed by economists, who - given the nature of the discipline - are interested in both causal and predictive applications. Econometrics analysis of time series data is both econometrics and time series analysis.

Time series analysis is very different from cross-sectional analysis and more challenging mainly because the independence assumption commonly made in cross-sections does not hold in time series. Time series analysis is different and in many ways more challenging than panel data or pooled cross-section analysis because we work with only 1 observation at a point in time per time series and thus have to make the most assumptions about the process that may have generated our time series.

Econometrics make use of a wide range of statistical tools in an attempt to study or uncover relationships between key macro indicators as well as more frequently observed data, such as stock market close or open. Having said the above, it is evident that data structures here have the form of a time-series.

However, not all time-series data-sets have the same degree of difficulty. Econometics or Financial time-series often exhibit some seasonality and periodicity, however, a more rigorous statistical analysis would be required due to their random and chaotic nature.

You would certainly expect that the time-series of the temperature variation of a city during a 24h day over one year to exhibit some key structural characteristics. This is not always the case with economics or finance data sets.

So, to answer the question, time-series analysis is required to analyze econometric data-sets. I hope that this covers it. Happy to discuss further.