For an assignment I need to perform a VAR model on the three variables real GDP, short term interest rate and inflation. While for the first two variables I have not any problem, I am struggling to understand how to obtain a reliable "inflation" time series to be used for the model.

An expression I found on "Vector Autoregressions" by James H. Stock and Mark W. Watson is: $$ \pi_t = 400 \ln \left(\frac{P_t}{P_{t-1}}\right) $$ where $P_t$ is the chain-weighted GDP index. However, I don't understand the reasoning behind this expression (why 400?) and I'd prefer to avoid using the chain-weighted GDP index since is an index I don't know very well.

I'd like to start from the GDP deflator. I tried with some reasoning to obtain an useful expression and I came to the conclusion that maybe a good variable is: $$ \pi_t = \frac{DEF_t -DEF_{t-1}}{DEF_t} $$

Is that correct? Which is the best series to use as inflation for the VAR model?


1 Answer 1


Stock and Watson use log approximation and quarterly data. Premultiplying the log approximation by 400 could be due to data being quarterly.

Your second method is fine too.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.