I think you should search for the topic called "whitening". This is in Box and Jenkins,  Time Series Analysis: Forecasting and Control (Wiley Series in Probability and Statistics). The aim is to treat all time series the same manner, i.e. bring one to stationarity and apply the same parameters to the other series. Then they should be comparable. Hope this helps.

Dear Swapnil,

Stationarity is not the assumption of ANOVA. Get ANOVA assumption fulfill and go ahead. Your objective is to compare average growth rate of three sectors over the time period. If you forecast future growth, then you are required to address stationarity issue.

regards.

Have you plotted the raw data for each of the sectors across the time period?  I'd suggest doing that before deciding on a modelling approach. 

Thanks for kind suggestions.

Kindly suggest now.

I'd fit a time series model with a trend.  Then compare the trends across the sectors.