I have time series, for example, gdp and unemployment(unemp), freq= 4.
What if I interpret it as cross-sectional data and do cross-sectional analysis instead of time series?
My task is to test how unemployment affects gdp.
Is it allowed to do that kind of analysis?
Do the coefficients in the model lm(gdp~unemp) have an economic explanation?
Best Answer
The major difference between time series data and cross-section data is that the former focuses on results gained over an extended period of time, often within a small area, whilst the latter focuses on the information received from surveys and opinions at a particular time, in various locations, depending on the information sought. Moreover, gdp in one time lag is likely to be dependent on the next time lag and so on. In a cross sectional point of view, you ignore this correlation. For your problem, I guess you will be trying to see how gdp is being affected by employment over time so that you can also estimate the future scenario.