Hi Dear

If you are not acquainted with the theoretical framework of conditional volatility models like ARCH/GARCH, the simplest approach is to regress revenue/expenditure ratio on its own past series and then extract the corresponding residuals and square them. That is the volatility series. I did not say which type of regression you should use. This depends on the results of your simple regression diagnostic checking. Volatility is the known scientific term for instability. Good luck.

A moving standard deviation has been most common measure for volatility in literature when the series under concern is low-frequency data.

thank you all the suggestions. can ARCH/GARCH can be applied for yearly data?

Paul Louangrath please elaborate your answer.i cont understand. any reference please suggest. thanks for the response.