Asset allocation puzzle

[simpletron]

I must apologize for my absent from this discussion. I went to my parent’s house for my dad’s 50th birthday last weekend. And now, I have to head up a new project at work and burn time off of the old contact, which means I’ll be working 80 hours a week until the end of September. But it isn’t all bad, I renegotiated the increase in pay rate and got it on both contacts, so I will be paid 140% more(versus 115% if I didn’t) more for only 100% more work. Needless to say I won’t have as much time as I would like to continue this discussion outside of the weekends. (working ~13.5hrs/day, six days a week is the plan)

I going to try an answer all your questions tonight, post by post like you did.

[simpletron]

Is your math solid here?

Assume a $1 M portfolio which generates a 40k pre tax income stream.

60-40 suggests that $400k is in bonds and this 10 years income-before inflation. Anything more aggressive than 60-40 would not have 10 years expenses in bonds.

Not necessary, you can go more aggressive and still have 10 years expenses in bonds if your bonds have more growth than inflation. But I will agree that it is a good estimate for 10 years expense.

In another sentence you said you had 5 years expenses in bonds- which is it- 10 or 5? A 5 year buffer in cash when market could stay down for 10 years does not make sense (to me). If market went down in 2000, would the stocks have caught up in 2005.

I am going to do 80% stock, 16% bonds, 4% cash, which is 5 years expenses in cash and bonds.

But I don’t see the point in going more conservative than 60% stock, 40% bonds, which is 10 years expense. Unless you know when you will die and going more conservative will guarantee that you will make it.

Are the 10 year periods rolling periods? If someone invested in 2000, do you think they will be ahead in 2010? I don't know if market has fully recovered from 2000-2002 bear, but thought I would ask.

Yes and no, the ten year periods are rolling but are based on only the first day of the years(aka take the first of any year and the first of the year ten years later, the second one is always higher). The ‘no’ comes from the fact that I know of days that this is not true(most are from 1929 right before the crash and a couple of random days in the early 70’s).

But just because you are positive, doesn’t mean you beat inflation. You would have lost purchasing power for a lot 10 year periods starting in the 30’s, late 60’s, early 70’s, and probably the 2000’s.

The 5 year periods that failed were 1928-1933(great depression), 1929-1934(great depression), 1937-1942(???), 1973-1978(OPEC was formed in 73), 1999-2004(tech bubble), 2000-2005(tech bubble). On a side note: every one of these except the OPEC one had extreme growth right before:
1927-1929 until crash: 150+%, best three run, 3 consecutive 20+% years
1933-1936: 200+%, best four year run includes best year(1933) and best two year(1935-36), 3 of 4 years 30+%
1995-1999: 250+%, best five year run, 5 consecutive 20+% years(I switch up my investments in 1999 after 3 20+% years because I thought great depression 2 was going to happen. It is the reason why I have any bonds(I-bonds))

The market did fully recover in 2007, but now we aren’t fully recovered. The market needs to go up by about 10% in the next 17 months and avoid a 5% lost in the next 5 months for the trend to continue. It is doable, so I am going to say yes because the trend is my friend.


In general, design withdraw portfolios for a worst case scenario.

What is the worst case? Are you sure that is the worst case and can’t be avoid? I ask about the can be avoided part because I have never suffer a negative return year for the 14 years I have been investing. Am I lucky or am I good? (a little of both is my answer)

Also you mention the 7 asset strategy (never negative for 3 years according to you). Imagine doing the 7 asset strategy in stocks + cash only (utilities/steady dividend stocks for bonds, property holding groups for REIT, mining/oil stocks for commodities). Just something to think about.

added bold parts

[simpletron]

HUH? I was taking the data in the original post and showing it needed more detail.

I knew (and I thought I stated) that the aggressive portfolio had more when retirement started.

This was my whole point- the aggressive portfolio gave a higher starting amount- so using a moderate portfolio (to stablize returns) in retirement would increase the probability of success.

What I was saying is that it isn’t a fair comparison.

This is how the calculator works:
A case is one run though from start to finish. It deposit money then applied a gain/lost for each year until the beginning of retirement then whatever that amount is, it starts withdrawing and then applying the gain/lost for each year. That is one case and it is given a pass/fail. Repeat 9,999 times for the 10,000 cases. The percentage is the number of passes divided by 10,000. the gain/lost is from a bell curve based on growth rate and standard deviation inputs.

You will agree than stocks are more volatile than bonds. Than means that stocks have greater upside and greater downside. For example the best year stock return is over 50% and the worst is more than -30% whereas for bonds the best year is about 30% and the worst about -20%.

So during the accumulation stage, if you take 10,000 cases like the calculator, that some will have lost more than 30% after a year which is worst that the worst possible for the bonds. Do this for 40 years and the range of possible outcomes for the higher risk portfolio is greater that the lower risk portfolio in both direction up and down. Therefore there will be cases of the 10,000 aggressive cases that start the withdraw phase with less than the lowest case of the 10,000 conservative cases. This gives a natural advantage to being conservative in your analysis because if the lowest of the conservative case failed, how can you expect cases which start even lower in the aggressive case succeed?

It is like this for two cases. After accumulation, the aggressive cases are $4 and $0, for the conservative cases are $1.5 and $.5. the average value of the aggressive cases are $2, which is greater than the $1 average of the conservative case, but the aggressive case will have at most 50% success whereas the conservative case can have up to 100% success.


But I could not find a way to make the calculator give me that data or let me run that scenario (except to hand type the returns and deviations).

I did not conclude it was better to be aggressive all the time- how did you come to that conclusion?

I conclude that is better to be ‘aggressive all the time’ over being ‘conservative all the time’ from the fact that you ended up being able to withdraw more money at the same percentage chance of success.

There is no way to compare ‘aggressive all the time’ to ‘aggressive until retirement then conservative’ with the data you had.

I agree the longer retirement needs more stocks.

One issue when discussing withdraw rates is how the withdraws occur.

There are many camps-
always leave 95% of porfolio there (max withdraw of 5%)
always take 4% of value (money will never run out- 4% of $1 leaves $.96).
4% adjusted upwards for inflation- do not adjust for inflation in down markets.
3 year rule (avoid a negative return over 3 year periods)- there is a 7 asset allocation which shows how to do this- add a commodities index, REIT index and various other securities to prevent portfolio from losing money in any 3 year period.
always take withdraws from dividends

aren’t the first two the same except, you are withdrawing 5% in the first and 4% in the second?

I tend to think that 3-4% starting withdraw and adjust upwards for inflation(no matter what) is really the only way to go from maintaining a standard of living point of view.

Where that money comes from is the interesting part, because there are so many ways.

I am in the dividend camp, or the 7 asset class camp to keep a stable portfolio.

The 7 asset class portfolio would be the bucket approach already mentioned.

The dividend approach is preferred because it has 100% success. This is where the calculator was flawed, because it had a 2.4% withdraw rate in some of the cases mentioned and kicked out an 85-90% success rate in those 2.4% SWR cases. If that 100% equity (or 93% equity) portfolio was 100% dividend paying stocks, every person I have spoke to has more money now than they did when they retired. 2-3.5% withdraw rates in those cases (some even quoting as high as 4.2% yield).

Those failure cases came from the variation in starting retirement amount explained earlier. Stop thinking that all the 10,000 cases have starting retirements amounts that are the average.

added bolded parts

[simpletron]

Jim, I have a hundred questions for you.

How do you invest your money that you need in 1 year? 2 years? 3years? 4years? 5 years? 6 years? 7 years? … 100 years? (never know how long you’ll live)

Your answer to these questions will determine how you should invest your retirement accounts during retirement. You should invest your money based on a weighted average of those asset allocations because it matches your risk tolerance. The weighting is your life expectancy for living that many more years.

Off topic: I remember reading an article of this prediction that at the rate the life expectancy is increasing that you and me will live forever…I thought it was crazy at first, but the more I thought about the more plausible it became.