I'm working on a project regarding Food inflation. I want to perform a SVAR analysis in a panel dataset. I was looking on Stata documentation and there is VAR command for panel or SVAR for time series. Any hint if there is a possibility to perform SVAR in a panel data in Stata?
Yes it can be done. Please check Holtz-Eakin, Newye and Rosen 1988 Econometrica paper for theoretical formulation. It can be conducted in STATA with the dataset declared as panel data.
I assume that your panel is a set of time series of several food prices. Therefore, it should be possible to apply SVAR. But the main question is: Why do you want to apply SVAR? I do not think that autoregressivity plays a crucial role for food inflation. Before beginning with econometric estimation, you should better analyse the problem by thinking over the possible factors influencing prices and try to find relations between your variables/data with rather simple instruments (e.g. diagrams, correlations).
If you are not really interested in the development of food prices, but only want to gain some experience in the method and its application, then, of course, applying SVAR without any economic consideration is ok.
Check the attached reference, it's clear and effective at operational level:)
Good luck with your investigation
Thank you @Benedetto Molinari for your time and suggestion. I was more into looking for a possibility to perform a Structural VAR in a panel through Stata.
Thank you @Anton Rainer! I appreciate your intervention. My dataset is a panel data (including all the countries for the last 20 years) of economical, financial, climatic indicators. The consulted literature suggested me the method of SVAR to analyze the responses to both idiosyncratic and common structural shocks, while permitting full cross member heterogeneity of the response dynamics.
I think that SVAR is often used to save oneself the trouble of a thorough specification of a model. SVAR means that all effects work with a lag of at least one year (as you have annual data). This lag is likely to be too long for some of the effects (e.g. for the effect of price changes of raw materials, exchange rate fluctuations, tax measures or wage cost increases). And I do not see any reasons, why and how the food price inflation of past years should have a significant influence on the food price inflation of the current year.
Well Fjona, if you already got hint of the theory, you can safely precede to stata either using XTVAR or PVAR. You can check documentation online (free download) for both commands.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis