My question is based on the statement in 2021 users manual regarding historical analysis........"The first projection begins with the first year of historic data (the first year for which data exists for all of your asset classes)". I have one asset class that I have created for which data is only available back to1976, so does this mean that all historic projections will be limited to 1976 or later? I see you have pre-loaded returns in the user assigned asset class columns back to 1928, would I want to delete these prior to the year (1976) of my self entered returns? Also (not my problem yet), but if hypothetically someone is using the historical analysis and has say 10-15 years remaining in their life span would the historical analysis not be somewhat less reliable as the results are only based on only 10 or 15 years of historical data?
Billy,
Even though the manual doesn't clearly state this, your historical data has to go back to at least 1928 like the tool's canned data. I suppose this is a judgment call, but in establishing the design I made the decision that we needed that much data to get a valid historical analysis. If you did have data going back to 1928 and a life expectancy of 15 years, the historical analysis would comprise 78 test cases. If your historical data went back to only 1976, the analysis would comprise only 30 test cases. In contrast, the Monte Carlo analysis consists of 500 test cases (and I'm not aware of any tool that uses fewer test cases for a Monte Carlo analysis).
I'm going to add an item to my candidate enhancement list for 2022 to consider a modification here to allow shorter historical sequences but, for now, the tool will simply refuse to perform the analysis unless all sequences go back to 1928, as a minimum.