As a research exercise I should perform Granger causality tests on pairs of national series (one describing output gap and the other an index of consumers' confidence).

However, I face a situation never met before. I have annual output gap data and both quarterly/monthly numbers for the consumers' index. I know that one should first test for stationarity and select the number of lags to be included but I doubt if the fact that the series are measured at different frequencies is an hindrance to my analysis.

What is an elegant way to produce results in this case? Should I annualise the quarterly data and then test for Granger causality? If so, how? I suppose case-by-case assumptions would be required.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy