Schiller data upbeat on SWR

[donheff]

On my portfolio spread sheet I run several SWR calculations including the standard initial percent plus inflation, a couple of flat % of portfolios, Guyton, etc. I have consistently taken out less than the smallest of the options but I like looking at things and calculating where would we be now if we did "x" or "y." In any event, one of the calculations I run is the Schiller PE/10 method. The basic idea is that you divide 1 by the Schiller PE/10 to arrive at a safe withdrawal amount. I use the more conservative .83/PE10 because several "authorities" describe it as a better approach. It has been fun to watch my portfolio drop with the DOW and the Schiller SWR go up. At COB today it went up to 4.29% (5% if you use 1). The idea, of course, is that if you are beginning retirement today the market is statistically poised for a good starting run. I haven't withdrawn anything close to 4% but it is nice to see Schiller data encouraging me to smile.

 

[haha]

On my portfolio spread sheet I run several SWR calculations including the standard initial percent plus inflation, a couple of flat % of portfolios, Guyton, etc. I have consistently taken out less than the smallest of the options but I like looking at things and calculating where would we be now if we did "x" or "y." In any event, one of the calculations I run is the Schiller PE/10 method. The basic idea is that you divide 1 by the Schiller PE/10 to arrive at a safe withdrawal amount. I use the more conservative .83/PE10 because several "authorities" describe it as a better approach. It has been fun to watch my portfolio drop with the DOW and the Schiller SWR go up. At COB today it went up to 4.29% (5% if you use 1). The idea, of course, is that if you are beginning retirement today the market is statistically poised for a good starting run. I haven't withdrawn anything close to 4% but it is nice to see Schiller data encouraging me to smile.

can you link to articles discussing this idea?

Thnx,

Ha

 

[arebelspy]

Seems fairly intuitive. If you can afford to retire today even after the drops of the past week with a good SWR, then going forward you should be even better, because the market is near a low.

OTOH, if you just get to a point where you're ready to retire two weeks ago at a good SWR, then the market tanking makes you look shakier (i.e. using FIRECalc you'd be retiring in a worse year).

That is, if you are ready to retire at the beginning of 08, then the market tanking then is scary, and over the long run you might not make it because you withdraw your normal WR and your portfolio is hit harder that year, making it harder to make up those gains. Whereas if you have enough to retire with a good SWR after it's tanked in 08, you're looking good.

That being said, articles with more info would be appreciated.

 

[Brett_Cameron]

Here is his home page link: http://www.econ.yale.edu/~shiller/

It contains many links including links to his data.

 

[donheff]

can you link to articles discussing this idea?

Thnx,

Ha

I thought I read the discussion that recommended using .83 instead of 1 as the dividend here but I can't find it. The links I kept on my spreadsheet to remind myself where I got this are:

A chart showing Schiller data over time with the current daily rate highlighted; and

A discussion of the Schiller data set.

I also found a Bogleheads discussion on using PE10 in SWR with a link to a detailed article.

By the way, I am not advocating using this for setting an SWR. I see it as just another, fairly intuitive way to look at things.

 

[samclem]

. . . because the market is near a low.

Oh, if only we could know that. It's surely nearer a low!

So, if one varies his/her withdrawal rate according the Schiller's data on valuations--is he/she a DMT?

 

[donheff]

Oh, if only we could know that. It's surely nearer a low!

So, if one varies his/her withdrawal rate according the Schiller's data on valuations--is he/she a DMT?

Since I don't plan to use it, I am no expert but the impression I got was that the .83/PE10 calculation was to set an initial SWR that would then be adjusted upward by inflation the standard way (or Guyton 95%ed, or whatever). The concept is simply that the initial "safe" WR (on a balanced portfolio) can be set somewhat higher if you ER when equities are underpriced than when they are overpriced. That makes sense to me. Safer still by all accounts is to start with an even lower SWR than Schiller or anyone else would recommend - provided you can afford it.

Edit: I haven't even considered what it would mean to do your annual withdrawal at a percent of portfolio varying with the Schiller calculation. It would be a sort of reverse timing, taking out a greater percentage when the market is down and a lesser percentage when up. Sounds like a recipe for disaster.

 

[walkinwood]

A PE10 based SWR is used only at the beginning of ER, isn't it? After that, your withdrawals are adjusted for inflation irrespective of the market.

 

[samclem]

A PE10 based SWR is used only at the beginning of ER, isn't it? After that, your withdrawals are adjusted for inflation irrespective of the market.

One could do it either way. There's some discussion of the options in the last link in donheff's post.

Everyone needs to decide if they want to adjust their annual "take" based on their portfolio's size or what inflation is doing. I'm in the first camp, but every situation is different.

 

[DblDoc]

The article I posted last week also addressed this issue: Can We Predict the Sustainable Withdrawal Rate for New Retirees?

It uses a formula that incorporates PE10, current dividend yields and nominal treasury yields for a "60:40" portfolio. It was not particularly rosy looking for recent or about to retirees...

DD