Re: i posted ths on anther frum, but it mst b 2 ha
Just out of curiosity, can you clarify your beliefs? Do you believe that the lack of long-term bad runs in the historical data is a random quirk, or do you believe that RTM is intrinsic to the system and that monte carlo simulation fails to account for RTM?
RTM is a form of auto-correlation, so I can't reconcile your two statements above.
Wab, I'll be glad to clarify, while concurrently apologizing to the original poster for the hijack, and also apologizing for being very grouchy. I tend to get that way when I'm trying to express ideas and opinions and someone calls me irrational and implies that my ideas are based on emotional misdirection intended to achieve some sort of "agenda".
My attempts to resolve the contradictions bought avoidance and eventually the passive name calling. I'm very guilty of being caught up in that, and assure I wont be caught up in it again.
Let me see if I can clarify some things.
- I do believe there are correlations in the market. I just do not believe they are measurable and predictable in most cases because the correlations are intertwined with highly variable social issues that inject an element of near randomness over periods of time less than 30 years. Bernsteins "guy walking the dog".
- I do not believe that in short term periods...7 to 10 years or less, that you can use any correlations - measured or implied - to predict anything meaningful about the future with regards to returns, inflation, interest rates, etc. In anything other than a very general way, eg: "interest rates are low now, they'll probably be higher at some point". "Gee, stock market returns have stunk for 10 years, they should get better sometime soon".
- I do believe that RTM happens, and it happens when the social influences weaken and investment values return to their core mathematical/financial value. I believe this can be measured in 30+ year periods.
- The whole disagreement over monte carlo sims comes from this statement. I said that monte carlo sims build their data sets from random periods, hence they can produce a random period that is worse than a historical run of comparable length. The response was: " You are simply not correct about this. Monte carlo analysis must assume something about the distribution of stock returns, bond returns and inflation. The analysis must assume something about the correlation between these distributions. And the analysis must assume somthing about the correlation of the data from year to year. What most monte carlo simulators do is fit the distributions of stock returns, bond returns and inflation to the distributions found from historical data."
I've never seen a monte carlo sim that did this and asked for one, never got it. I got evasive definition arguments over the word "correlation", rattlesnakes and weather prediction.
The root problem is that some folks think that historical data and calculators that use them are just about foolproof for future predictions of their investment returns.
Nice tool, but lets not get too carried away...