Hypothetical Y2K retiree update

[raddr]

I've finally gotten around to calculating the 2004 year end balance of the portfolio of a hypothetical investor who retired at the end of 1999 with a 75:25 mix of S&P500 stocks and 6 mo. commercial paper. This was the type of portfolio that was being touted as "100% safe" on some investment boards based on historical backtesting at the time. The following numbers assume a 0.2%/yr. expense ratio, 4%/yr. withdrawal, and are stated in constant (real) dollar amounts:

Code: Select all

    Year    Return  CPI  Withdrawal Balance
    1999                            1000
    2000    -9.1     3.4      40     869
    2001    -9.9     1.6      40     743
    2002   -18.8     2.4      40     564
    2003    19.7     1.8      40     635
    2004     5.3       3      40     628


It's clear that this person would be in real trouble, particularly if he is an early retiree with possibly 30-50 years of life remaining. He would be taking withdrawals of about 6.4% at the start of 2005 because of the severe hit that the portfolio has taken. Using MC analysis this portfolio would have a greater than 50:50 chance of failure during the next 40 years IF you assume that returns will be as good as in the past. Using Gordon equation predictions the chance of failure is about 80%.

[ben]

VBINX guy ended 2004 with $883k in todays USD. There is a good reason for using Sp500 and 6mth commercial papers for the longterm historical SWR studies - but when doing it in our modern age it does not make any sense I think. VBINX (tot equ/bond) guy would have to pull around $45k 1st Jan 2005.

I also think you are doing a mistake by BOTH adjusting nest egg size AND withdrawals for inflation. Either or. Am I missing something?

Cheers!

[raddr]

Hi Ben,

VBINX guy ended 2004 with $883k in todays USD. There is a good reason for using Sp500 and 6mth commercial papers for the longterm historical SWR studies - but when doing it in our modern age it does not make any sense I think. VBINX (tot equ/bond) guy would have to pull around $45k 1st Jan 2005.


Cheers!

The reason I used a 75:25 mix rather than the more bond-dominated VBINX is that at the height of the bubble everyone was saying that you were crazy to have less than 70-75% stocks in your portfolio and because the historically optimized allocation was about 75% stocks. Also, the lower equity allocation found in VBINX is way conservative for a very long retirement, IMO.

BTW, I get an inflation adjusted portfolio value of $767 rather than $883 for the VBINX portfolio. I think that is because I am using constant 2000 dollars and am adjusting for inflation whereas you are not.

I also think you are doing a mistake by BOTH adjusting nest egg size AND withdrawals for inflation. Either or. Am I missing something?

I am adjusting both the balance and withdrawals for inflation. The table above shows constant dollar amounts for both. The
current withdrawal rate of about 6.4% is due to the 37% hit the portfolio has taken (0.04/0.628 = 6.4%).

[ben]

Yes my $883k is in todays $ as mentioned. Also agree that we are not comparing apples to apples with this fund - but 60% equity is plenty for me and most FIREd I think. The 40% could be split out more: commodities/PM/foreign bonds/reits come to mind.

Anyway; just proves what we all say here: diversify into asset classes you think are not too overvalued.

[raddr]

Anyway; just proves what we all say here: diversify into asset classes you think are not too overvalued.

Exactly. That's really the point of the whole exercise, isn't it? It shows the dangers of loading up on the hottest asset class (S&P500 in this case) and thinking that you'll be okay because of it's historical performance.

[ataloss]

628 k? I would be working especially with current valuations

[raddr]

628 k? I would be working especially with current valuations

Yeah, me too. A portfolio that takes that kind of hit early on is pretty much doomed to failure.

[lazyday]

Very interesting and scary analysis.

I imagine if one did something like the Trinity study in 1965, the 100% safe withdrawal rate would be higher than 4%. (I guess we can calculate the swr to 1965, I haven't). Then, a 1965 or 1973 retiree using that maximum swr would have run out of money.

I think the same might be true for 1999/2000, only more so. Year end 1999 might turn out to have been worse than any point in history to retire on an S&P500/shortbond portfolio.

I agree that the hypothetical retiree is in trouble, and may easily run out of money in the years or decades ahead. He/she might make things even worse IMO by trying too hard to fix past mistakes, and gobbling up lots of small cap value today. (After all, if S&P500 was a hot, very overpriced asset in 1/2000, maybe today SV is a hot, somewhat overpriced asset, compard to other assets.)

[raddr]

Very interesting and scary analysis.

I imagine if one did something like the Trinity study in 1965, the 100% safe withdrawal rate would be higher than 4%.

It would've been just a bit higher but, yes, the more history you have behind you the lower the SWR will go. Just because it was around 4% for the last century or so doesn't mean it can't can't be a lot lower in the next century, particularly when dividend yields are less than half of what they were up till now.

[tmeri]

raddr, what are the details for the returns, or can you post the source for those numbers?

I am adjusting both the balance and withdrawals for inflation.

That doesn't seem right to me, but I have not looked at your numbers in detail yet.

[raddr]

Hi Tmeri,

raddr, what are the details for the returns, or can you post the source for those numbers?

I am adjusting both the balance and withdrawals for inflation.

That doesn't seem right to me, but I have not looked at your numbers in detail yet.

The S&P500 returns can be found at a number of places. M* (VFINX) is as good a place as any:

http://quicktake.morningstar.com/Fund/T ... ab=returns

6 mo. commercial paper is more problematic. In fact, looking back at my database I have been using 6-mo. CD rates since commercial paper series was discontinued. Both series, as well as CPI data, can be found here:

http://research.stlouisfed.org/fred2/

The table I posted is a bit confusing. The "Return" column is actually the real return, i.e. a 75:25 mix of S&P500:6 mo. CD minus inflation and minus 0.2% ER. The gross return can be had by adding back the CPI plus the ER. Thus, for 2000 the real return of -9.1%, net of expenses, in the the table would be be 5.5% nominal, before expenses. That's probably why the numbers look bad to you. After that calculation I subtracted 4%/yr. adjusted for inflation as the withdrawal rate.

Does this clear things up at all?

[tmeri]

How does 5.5% nominal become -9.1% real after you subtract 3.4% inflation and 0.2% expenses? I get: 5.5% - 3.4% - 0.2% = 1.9%. Obviously, I'm missing something.

[raddr]

Not yet. How does 5.5% nominal become -9.1% real after you subtract 3.4% inflation and 0.2% expenses? I get: 5.5% - 3.4% - 0.2% = 1.9%. Obviously, I'm missing something.

Oops, I'm sorry. I left off the minus sign. It's -5.5 not 5.5. Here's my stock/cash data:

Code: Select all

   6 mo. CD           S&P500
2000   6.8               -9.1
2001   3.7               -11.9
2002   1.8               -22.1
2003   1.2               28.68
2004   1.7               10.88

[wanderer]

ben -

I also think you are doing a mistake by BOTH adjusting nest egg size AND withdrawals for inflation.

I suppose he could inflate up the $40 w/drawal and take out of current dollar nest egg but wouldn't the result be the same?

Agree with raddr, VBINX bond holdings too high for 50 year horizon.

[wanderer]

hmm. I come out with 7k-12k more for ending balances. Assumed draw at beginning so 'bad' sequence affects balance a bit less... Still, not a good sitch.