J
Jake
Guest
When the 4% rule is applied does your principle get used up at the end?
Hi ThirdAgeI am not a numbers person, but isn't the idea behind the 4% rule basically to live as much as possible on ones interest? Since 4% is not unreasonable given a conservative mix of investments, it would seem reasonable to figure out whether 4% of whatever you have is a reasonable amount to live on and go on from there.
Or is there something that I am missing?
If this is really a numbers-heavy concept, I probably won't get it. So KISS (keep it simple, stupid) answers will be really helpful.
thirdage
If you start withdrawal in 2003, can you use the high point of 2000 as your asset withdrawal amount?... This would give a higher dollar withdrawal. Does this make sense or is there something missing?
Valuing a portfolio in 2000, at an alltime high PE ratio didn't make sense then, or now.
A lot of REs don't feel comfortable with the real answer to your question, but if you believe in the validity of the historical analysis that the 4% rule is derived from . . . then the answer is "yes".If you start withdrawal in 2003, can you use the high point of 2000 as your asset withdrawal amount?
Based upon a 4 % withdrawal, can you use the high value (i.e. before March 2000) of your asset as your base amount to calculate the 4% withdrawal and adding that number of years ( in this case 3 years) to your end point? Instead of 30 years it is now 33 years , etc. This would give a higher dollar withdrawal.
Does this make sense or is there something missing?
Think of it this way . . . If a person had retired on March 1, 2000, they could have safely retired (assuming the validity of the 4% SWR) on 4% initial withdrawal rate of their nest egg value on that date.
This analysis shows only a few of the most glaring problems with SWR as it has been advanced. In reality, the situation is quite a bit worse.
I am afraid that Emperor SWR has no clothes.
The second reason has to do with the statistical significance of the underlying process by which SWR is calculated. Since 1871 there are 4 non-overlapping 30 year periods! 4! That means that you have 4 independent runs, not >100 as is often advanced.
Hi Telly,Did the quote above refer to a historical calculator like FIRECalc, or to some 4% hypothesis.
It's just that there is a very big ? regarding the phrase "assuming the validity of the 4% SWR". For various reasons, I believe that is an unwarranted assumption. First reason is that using several metrics, PE_10 of Shiller, Tobin's Q, etc- the market in the late 1990's and 2000 was much more overvalued than during 1929, and of course even more than during the mid 60's. So we have an out of sample condition. The existence of Year 2000 valuations proves beyond doubt that the past did not adequately encompass the year 2000. We experienced an upside outlier.
I'm not aware of any mathematical reason that the number of "non-overlapping" periods is important to validity of the analysis. Think of it this way . . . if you wanted to determine empirically the worst case of how many times a flipped coin would turn up heads out of 30 flips, you don't need to consider non-overlapping events. You can flip the coin 100 times and look at all the strings of 30 flips in a row to estimate the worst case. In fact, if you don't do that, you are throwing out a lot of valid and important data.The second reason has to do with the statistical significance of the underlying process by which SWR is calculated. Since 1871 there are 4 non-overlapping 30 year periods! 4! That means that you have 4 independent runs, not >100 as is often advanced.
Then there is the problem of stationarity. For statistical sampling to mean anything at all, there must be some reason to believe that there is not an overall drift in the data- ie. that data points from many years ago are drawn from the same universe as more recent ones. I see no reason that this would likely be true, and many reasons why it is likely not true.
This analysis shows only a few of the most glaring problems with SWR as it has been advanced. In reality, the situation is quite a bit worse.
I am afraid that Emperor SWR has no clothes.
Mikey
I'm not aware of any mathematical reason that the number of "non-overlapping" periods is important to validity of the analysis. Think of it this way . . . if you wanted to determine empirically the worst case of how many times a flipped coin would turn up heads out of 30 flips, you don't need to consider non-overlapping events. You can flip the coin 100 times and look at all the strings of 30 flips in a row to estimate the worst case. In fact, if you don't do that, you are throwing out a lot of valid and important data.
The SWR is just a tool for planning. Planning and safety have little to do with each other.
That would be true if you were trying to fit the data to some predictive curve as you have indicated. But that's not the point of the historical analysis. In this situation, you are looking for a high probability worst case of successive years. This is very different analysis than predictive curve fitting. It is true that nest egg cash flow analysis of years 1 - 30 are going to share much with nest egg cash flow from years 2 - 31 . . . or years 10 - 40. But the worst case for nest egg survival may fall in years 5 - 36 and if you only use non-overlapping intervals and don't consider all the possible 30 contiguous year intervals, you will dramatically overestimate the safe withdrawal rate.If you don't believe this, or understand it, do the exercise yourself. Try developing a safe withdrawal rate considering only non-overlapping intervals and see what answer you get. It does not make any sense mathematically to require non-overlapping intervals to do this worst case analysis.If you are trying to use standard statistical measures you need independent trials. If you consider the last 130 years of overlapping 30 year time periods to be 100 samples, you are overstating the significance of your results by a wide margin. Try it yourself if you like. Flip a coin 130 times like you suggested, and then write down the 100 overlapping 30 "year" periods. Then test the goodness of fit to the true binomial(0.5,30) distribution. You should find that it isn't even close.
I would also disagree with salaryguru's point. If you think that the 4% rule means that it is completely impossible to exhaust your nest egg using a 4% withdrawal then I guess the implication is that you can pick your highest point. However, I think the 4% rule should be interpreted as a reasonably conservative guideline supported by a reasonable amount of evidence. However, we are not looking at probabilities for the last three years -- we know they were terrible. You have to understand that there is always a chance that your portfolio will be depleted, and once it has fallen in value by 30-50%, then that chance is going to be much greater than it was when you started. In the 130 years of history the stock market has always rallied strongly off of 30%+ drops -- are you going to count on a rally like that this time?
Bernstein makes an important point, that even the best and most powerful civilizations have a less than 90% survival rate over 30 year timeframes. No portfolio is going to withstand hyperinflation, or government collapse, so any survival rate over 90% is meaningless.