Simple question but one thats stumping me slightly. I am doing a study analysing historical portfolio performance related to portfolios constructed with separate factors in mind. I realise simple returns are asset additive, and that log returns are time additive, but as the portfolio is a panel-data series, consisting of asset additive returns at multiple time observations, ultimately forming the overall portfolio return, I am unsure of which to use.
At present my intention is to use the simple return of each of the portfolio's individual securities across the time-series, the take their average as the weighted return of each time observation. To annualise them, I intend to simply take their mean and multiply by 250, the number of time observations. Any help on whether this is misguided, or direction toward a more appropriate method is much appreciated.