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.
@smatthews51 Have you given this “enhancement” any additional consideration? It would be ideal to be able to run historical analysis that includes indices including international, emerging markets, global bond, REIT, etc. that only have return figures from 1970s/1980s to present. The cumulative of these assets classes make up a minority, yet impactful percentage of my portfolio. Thanks,
But starting in the mid 70's would miss the Great Depression and most of the 1966-1982 inflation/oil crisis/bear market. Plus, International was literally pounded into rubble in WWII and starting in the 70's would miss that too. Any conclusions derived from leaving out the bad stuff would give a giant false sense of security.
I don't understand the question about 10-15 year histories being less reliable. Nothing about stocks and bonds is reliable and the real reason to look at historical data is to notice the huge differences in outcome depending on what the unpredictable future holds - the message should be that you shouldn't fall in love with steady growth projections that show we are all going to be rich, it may or may not happen for us. The message of the past is to be diversified, have a cushion and remain flexible.
@accurate3567 Yes, and I concluded that we needed to stay with the current design so that all available historical data is included in the analysis.
Stuart
The one argument against using historical periods like the great depression is that theoretically they can't happen again, due to the checks and balances implemented since then, leveraging those experiences. For example, in 1929 the market just kept falling and falling - there were no circuit breakers to stop trading. That's why we have them now. Same with all the follow-on economic disasters.
What happened with COVID was important. It really was the ultimate black-swan event, other than alien landings or complete nuclear or climate devastation, which is certainly a possibility. Almost everyone out of work around the world almost instantly and simultaneously. Supply chains around the world ground to a halt. Massive fear and uncertainty. We should still be in a deep, deep depression due to that. But, we're not. We implemented lessons learned to keep households, businesses, bond markets afloat. It was an incredible recovery, and fingers crossed, a soft landing.
Does that mean we become overconfident? Certainly not. We're in an age of volatility and uncertainty. I worry every day about what the future holds. It's why I try to ensure my clients buy into a good safety net strategy like bucketing, so they have the assurance they can ride out a five-year recovery, and regardless can live off their pension, SS, fixed income, or other income if their investments disappear. If so, what the stock market is doing becomes irrelevant, and they can sleep at night. Its like that minimum dignity floor concept at The Retirement and IRA show. It allows you to enjoy your life after working so long.
That said, I've talked more than one client off the "bonds are stupid, I'm all in for the stock market" ledge by showing them sequence risk graphically - turning on the historical sequence analysis in Pralana Online and setting the starting year to 1965. It's an eye-opener! Just don't forget to turn it off when you're done with that exercise, uncheck the box.
That's why I run both Monte Carlo and historical, and have this exact conversation with clients.
