I've seen the thread where historical and monte carlo results were discussed, and several people seemed to say that they got better results from Monte Carlo than from Historical. I get better results consistently from historical, and want to understand if I am setting it up correctly. I am not using CAPE, I have a start year of 1928, and and I have not checked the box for "Activate Historical Sequence Analysis" because I don't understand it, and haven't had time to see if it merely adds information or would change the result. When I set up Monte Carlo, I have checked the box for Correlated ROR's but not Custom std deviations. I seem to remember a setting somewhere for something called "deterministic" but can't remember where or what is was.
Can someone help me with the set-up parameters for these things? My wish on historical would be to have the broadest set of data possible, which is why I picked 1928. I want the most conservative results, so that hopefully I wouldn't have unpleasant surprises. Pralana is a very powerful tool, and I don't want to cause GIGO errors by not understanding how to use these parameters, even though I don't really understand what they all mean.
@debrazebra, You are misunderstanding the parameters. Easy mistake to make. 🙂
There are two features in the historical analysis: "Historical analysis" and "historic sequence analysis". The two sound identical, so please don't blame yourself for being confused. I will explain each. I will answer your specific questions about what to set in a second response.
Generic Historical Analysis: The basic idea behind Pralana's historical analysis is to take history and summarize it. Pralana has the past returns of broad asset classes from 1928-2024, which you can see Analyze>Historical Analysis>Historical Rates of Returns and Inflation.
What Pralana does is go through each year in the past a person could have retired, and then looks at the historic ups and downs of what would have happened with your own particular asset allocation if you had retired in each of those years. So for instance with a 30 year retirement, it would start in 1928 and look at what would have happened 1928 till 1958. Then it would look at 1929 through 1959. And so on.
At a certain point in this process Pralana runs out of data, since it would go beyond 2024. For instance, on a 30 year retirements, the years 1995-2025 would be one year short (2025) and 1996 on would be two years short (2025, 2026) and so on. So to provide a more complete analysis, Pralana projects that missing years will match your expected returns projections you entered in the build section (the same ones used for monte carlo).
You also need to first tell Pralana which historical asset to use for each of your own assets. You do this at Analyze > Historical Analysis > Historical Rates of Return and INflation > Map Asset Classes to Historical Classes. In most cases this is simple: you use the "Sp500: for your Stocks, "10 year treasury bonds" for your Bonds, and "Cash" or "3 month treasury bills" for your Cash.
Punchline: The percentile success rate stated in the generic Historical Analysis run is the tabulation of all of these 86 or so series of yearly return paths. If Pralana says you have a 90% chance of success (historically) it means that 90% of the runs (1928-58, 1929-59, 1930-60, etc) resulted in not running out of money. It can be thought of as this: If you had relatives exactly like you, each of whom retired in each of the years of the past, how many of them would you have not run out of money?
Historical Sequence Analysis: All this does is add in a line onto the graph showing one of those 86 or so historical sequences, starting with the year you indicate in the entry box. It shows up on the graph in a yellow line. So for instance, a common starting point is 1965, because it is one of the worst 30 year sequences in history for 60/40 portfolios.
The point of this is to go beyond averages and stress test with specific ranges of history. There is a little table on the lower right of the historical results graph that tells you the best and worst years, so you don't have to guess! It’s not always 1965. You can also test good start years.
There are other reasons you might want to look at one particular range out of all of history: say, maybe you are interested in looking what would have happened to your plan when inflation was super high (1970s) at the beginning or end of your plan. Or what would happen if we had a replication of the Great Depression right at the start of your retirement!
Like all analysis, you have to be careful not to use it as your only way of thinking. One minus of the historical results is that the analysis puts a lot of emphasis on the same years in the middle of the historical record that repeat again and again, with less repetition in the analyses of the portions of history that bookend the beginning and end 1928-2025 block. Another minus is that later runs include a lot of projected future returns (say those starting at 2000) are not actually historical but mostly your projections. But I'll leave these finer points for other discussion.
If I've missed anything or gotten anything wrong, please correct me guys.
@jkandell thank you, this is already very helpful. I’ll look forward to your next response. Thanks again!
Part 2:
I am not using CAPE, I have a start year of 1928, and and I have not checked the box for "Activate Historical Sequence Analysis" because I don't understand it, and haven't had time to see if it merely adds information or would change the result.
As I hope is clear from my longer explanation, if you don't have the historical sequence analysis checkbox checked, it is irrelevant what year you list there for the start year. (If misery loves company, I made the same mistake when I started with Pralana!)
In other words, if your "wish on historical would be to have the broadest set of data possible", you don't need to do anything for that to be true: that is exactly what happens with the generic historical analysis! It always starts at 1928 and goes forward starting 1928,29,30.... for all possible sequences, giving you the broadest range. (In fact, it goes beyond history and fills in data with projections to give you even more sequences. ;-))
I seem to remember a setting somewhere for something called "deterministic" but can't remember where or what is was.
You don't "set" deterministic results in Pralana beyond entering your expected returns and all the other information in the build section. "Deterministic" is just Pralana's term for the results using your estimated returns and entries as "fact" with no variance, as if the markets and everything else acted like a clock. These clocklike estimates are what Pralana shows in all of its tabular tables.
We all know that reality is not like a clock: so it is best to think of these as average results not as truth. My one caution is to use annualized (compounded annualized=cagr) expected returns for your build entries, since they “build in” variance. Otherwise your deterministic and Monte carlo will be too high
The "monte carlo" attempts to show what variance can look like around that average. Both deterministic and monte carlo spread are valuable in evaluating how things might unfold in the future.
This is indeed more "conservative" (your words) than not checking that box. In reality assets are somewhere between correlated and not-correlated, closer to the latter; but what you chiose is the more conservative option.
Thank you for both replies. They are very helpful. I'm going to sit with this a little later, but at least now I know what I'm looking at. I've seen quite a few posts that mention 1965 so I will use that as another reference point. My AA is currently 45/55, likely to shift to 50/50 when I get my full SS in 2 years (as opposed to survivor benefit), but that's ok, it's still a good starting point for a bad time, if that makes sense.
Thanks again.
@debrazebra i personally would look (via specific historical sequence “) at all the worse five or so years listed after running generic historical analysis, in the tiny table toward the lower right.
You might as well see the worst! And each misery had its own unique version of joy and suffering
@jkandell thank you, this is already very helpful. I’ll look forward to your next response. Thanks again!
