Page 1 of 1

jwr has a good point about switching thresholds

Posted: Sat Apr 24, 2004 2:50 am
by ataloss
from his defense of switching post:
In our recommendations we have repeatedly advised people not to depend on numbers alone. Rather, we have advised people to look for cause and effect. We have searched out sensitivities to reduce our dependence on the numbers themselves.


some of us (e.g. me) tend to be a little literal with regard to numbers. if jwr's switching strategy shows that switching % stock allocation at a pe10 of 15 (or whatever) would have been optimal in the past, we would tend to apply that to the future. any intelligent person would realize that you can't predict optimal future approaches using optimized past approaches. I think that what jwr means about not using his numbers "alone" is that he agrees that optimal (or even good) future switching thresholds are unknown and unknowable. But there is some (unknown) future switching threshold that would be optimal. In principle, one should hold less of an asset class as it becomes over valued (or fully valued, or less undervalued). IOW if expected returns are lower hold less of that asset class. I think that this is broadly consistent with results you would get from mean variance optimizer analysis. In most circles this is probably noncontroversial. Perhaps some would hold that an 80% stock allocation is always optimal based on historical swr from 1871-2000 regardless of current valuation. I think jwr is responding to those individuals. bernstein does mention considering "the price of tomatoes" when making allocation policy (as an alternative to relying strictly on mvo outputs). bernstein makes no effort to quantify this for the s&p or other asset classes- as far as I know.

this generating numbers that are not to be relied upon has confused me and possibly others. I think that all this effort has been directed at supporting another poster's point that an 80% stock allocation isn't always optimal for a retiree-

Re: jwr has a good point about switching thresholds

Posted: Sat Apr 24, 2004 3:28 am
by raddr
ataloss wrote: from his defense of switching post:
In our recommendations we have repeatedly advised people not to depend on numbers alone. Rather, we have advised people to look for cause and effect. We have searched out sensitivities to reduce our dependence on the numbers themselves.


I looked at the "switching" threads over there and the impression I'm left with is that cause and effect is implied but it is not proved. As soon as you look at out of sample data, i.e. data since 1980, the switching strategies fail. Not only that but there are numerous self-congratulatory pats on the back and statements about "breaking new ground" and "we have now proven" this or that when, in fact, all that is being done is selectively using historical data to optimize past history and trying to use if for forecasting the future.

If these PE switching studies were not meant to be used literally then I think that there should be some clear statements to that effect at the top and bottom of the threads. I think that Ataloss is right that most of us here are knowledgeable enough to see these threads for what they are, i.e. optimizations of past market behavior without much future predictive value but there are some out there who might actually try to follow the implied recommendations which IMHO would be a major mistake. It is clear from my post comparing the pre 1980 and post 1980 numbers that some out there were unaware of just how dangerous blindly following this type of allocation strategy is. In fact I myself was surprised at just how bad it would it would have been to bail out of stocks and into cash, bonds, TIPS, etc. in 1993 when the PE10 went above 20.
this generating numbers that are not to be relied upon has confused me and possibly others. I think that all this effort has been directed at supporting another poster's point that an 80% stock allocation isn't always optimal for a retiree-


Right, the implication in the threads I looked at was that the switching strategies were viable for future use which I think is dangerous. If there was a recommendation not to follow them then it was pretty well hidden and I missed it. My whole point in the other thread is not to bash John but to make sure that no aspiring retiree bails out of stocks based on some dubious indicator without rigorously running the numbers himself like I did.

Re: jwr has a good point about switching thresholds

Posted: Sat Apr 24, 2004 5:03 am
by BenSolar
raddr wrote: [Right, the implication in the threads I looked at was that the switching strategies were viable for future use which I think is dangerous. If there was a recommendation not to follow them then it was pretty well hidden and I missed it.

I will note that it seemed to me that pretty much any time John approached an actual recommendation, TIPS were the preferred alternative - not commercial paper. Which of course makes a big difference as long as TIPS are available at a decent rate.

Regards,

Re: jwr has a good point about switching thresholds

Posted: Sat Apr 24, 2004 5:25 am
by raddr
BenSolar wrote:
raddr wrote: [Right, the implication in the threads I looked at was that the switching strategies were viable for future use which I think is dangerous. If there was a recommendation not to follow them then it was pretty well hidden and I missed it.

I will note that it seemed to me that pretty much any time John approached an actual recommendation, TIPS were the preferred alternative - not commercial paper. Which of course makes a big difference as long as TIPS are available at a decent rate.

Regards,


Hi Ben,

Actually, it would've made little difference in the time period I was talking about. A 1993 retiree in starting with $1M TIPS or Ibonds at the current rate of about 2.5% would have a portfolio of about $803,000 vs. $1,594,000 if he'd been in 80:20 stocks:cash the whole time (inflation adjusted values, 4% withdrawal rate). The first portfolio will likely go broke in 30 years or less while the second will survive with > 95% certainty, even if you postulate lower returns for stocks (~3.5%/ year). Again, it would've been a big mistake to dump stocks in favor of lower yielding cash or TIPS when the PE10 went above 20 in late 1992.

Since the PE10 rose above 20 in Dec. 1992 VFINX has had a real return of about 9-10%/yr., well above historical averages and way above theoretical TIPS or cash.

Posted: Sat Apr 24, 2004 8:01 am
by ataloss
as someone said Thursday:
of course an unreasonable person could invent some reason to exclude the post 1980 data (and the pre 1921 data or whatever else didn't "fit")

http://nofeeboards.com/boards/viewtopic ... 540#p19540

and jwr did make a case for excluding the 90s data in the defense of switching thread. his thinking seems reasonable (as did his reasoning for excluding the pre 1921 data) but overall, the more data you have to exclude to make the theory fit the facts the less confidence it inspires

Posted: Sun Apr 25, 2004 2:59 am
by raddr
ataloss wrote: but overall, the more data you have to exclude to make the theory fit the facts the less confidence it inspires


LOL! To say the least. If you throw out enough data you can get some real pretty graphs and spiffy statistics. :wink:

BTW my comments were not aimed at JWR's board specifically. Some of the most egregious missapplications of valuation indicators and switching came from discussions over at TMF. My whole point is that bailing out of all equities based on an S&P500 indicator that worked in the past is very dangerous. That said, it is just as crazy to way overpay for any asset class in the first place. That's why I recommend not abandoning equities in times of high S&P500 valuations but instead look for the more reasonable valued asset classes out there.

Posted: Sun Apr 25, 2004 3:37 am
by ataloss
there is a whole board culture at tmf of discovering past associations and attempting to predict the future on that basis

this is my favorite web page dubunking the dumb 4 strategy

http://www.investorhome.com/dogs.htm
Sharpe (William i.e. the nobel Laureate)makes the point that on one hand, if you search hard enough with a large number of random data sets, you will eventually "find some strategy that would have made you a fortune."


maybe the fact that I invested a little in the dumb 4 strategy makes me sensitive to the fact that a little clear thinking might prevent newbies from putting money into some doomed data mined strategy

Posted: Sun Apr 25, 2004 7:09 am
by Mike
From the article ataloss linked to:
...the tendency for investors following the strategy to drive up the price of the stocks thereby reducing or eliminating excess returns.


This is always a danger when a fixed strategy is widely recommended to a large group. For example, I read articles in the WSJ recommending that all investors put 10% of their portfolio into REITs, without explaining how to determine whether the current valuation of REITs is reasonable. Yet the total market cap of the entire REIT sector is less than the market cap of Microsoft, and couldn't possibly absorb 10% of all investor assets. Any strategy intended for use by the majority of investors needs to include a system for evaluating whether assets are currently valued reasonably. A degree of flexibility should be incorporated into the strategy so that assets flow to their most productive places. Even asset classes as large as the S&P can become overvalued if enough investors chase it without regard to valuation.

Posted: Sun Apr 25, 2004 7:56 am
by raddr
Hi Mike,
Mike wrote: Even asset classes as large as the S&P can become overvalued if enough investors chase it without regard to valuation.


This is exactly what has happened IMHO. Investor psychology amazes me in this regard. For example, some investors fret and worry about paying a few basis points in trading costs or expense ratios but think nothing of buying an index at 2 or 3 times it's normal valuation when there are plenty of better alternatives out there. Go figure. :?

Posted: Sun Apr 25, 2004 4:54 pm
by ataloss
I think large asset classes (s&p 500, tsm) can probably only become wildly overvalued if the demographic situation allows

Posted: Sun Apr 25, 2004 5:17 pm
by bpp
For example, some investors fret and worry about paying a few basis points in trading costs or expense ratios but think nothing of buying an index at 2 or 3 times it's normal valuation when there are plenty of better alternatives out there.


That would be me. :shock:US large-caps became underweight in my portfolio last year, so I loaded up on them to rebalance, without concern for the valuation metrics. I also moved my US holdings to a lower-cost form (going from TD Waterhouse index funds to Vanguard).

Penny wise and pound foolish? Perhaps. We'll see in about 20 years. :)

Bpp

Posted: Mon Apr 26, 2004 2:41 am
by raddr
bpp wrote:
For example, some investors fret and worry about paying a few basis points in trading costs or expense ratios but think nothing of buying an index at 2 or 3 times it's normal valuation when there are plenty of better alternatives out there.


That would be me. :shock:US large-caps became underweight in my portfolio last year, so I loaded up on them to rebalance, without concern for the valuation metrics. I also moved my US holdings to a lower-cost form (going from TD Waterhouse index funds to Vanguard).

Penny wise and pound foolish? Perhaps. We'll see in about 20 years. :)

Bpp


Hi Bpp,

I didn't mean that it's wrong to hold a richly valued asset class, just that it's funny that some index investors worry more about squeezing a couple of hundreths of a percent out of an asset class by obsessing over expenses than they do about whether it is really worth the purchase price in the first place. OTOH, most house buyers correctly worry both about purchase expenses and what they think the house is really worth before they buy. Think about it - very interesting and different psychology, no? My question is, why should the two be treated any different?

What you are doing is rebalancing which is a whole 'nother story. There's a big difference there.

Posted: Mon Apr 26, 2004 3:16 am
by Bookm
ataloss wrote: there is a whole board culture at tmf of discovering past associations and attempting to predict the future on that basis

this is my favorite web page dubunking the dumb 4 strategy

http://www.investorhome.com/dogs.htm
Sharpe (William i.e. the nobel Laureate)makes the point that on one hand, if you search hard enough with a large number of random data sets, you will eventually "find some strategy that would have made you a fortune."


maybe the fact that I invested a little in the dumb 4 strategy makes me sensitive to the fact that a little clear thinking might prevent newbies from putting money into some doomed data mined strategy

Heck, I even bought Robert Sheard's book. Ironically enough, taking the time to read the book prolly saved me from trying the FF. At the time I was reading it was, this was also around the time Datasnooper started posting on the FF board. Our arguements there and the strategy's returns at the time prevented me from pulling the trigger. Good thing the illusion of knowledge didn't stop me from discovering my ignorance. Yet, I still have the book. :oops: Not sure why tho.

Bookm

Posted: Mon Apr 26, 2004 3:21 am
by bpp
Hi raddr,
I didn't mean that it's wrong to hold a richly valued asset class, just that it's funny that some index investors worry more about squeezing a couple of hundreths of a percent out of an asset class by obsessing over expenses than they do about whether it is really worth the purchase price in the first place. OTOH, most house buyers correctly worry both about purchase expenses and what they think the house is really worth before they buy. Think about it - very interesting and different psychology, no? My question is, why should the two be treated any different?


I think this is explained by your observation elsewhere that indexing (and mutual funds in general) make it almost "too easy" to buy stocks. If people had to look at each stock before they bought it, they would agonize as much as they do over a house purchase, I bet. The valuation and risk of an individual stock seem like very palpable (if hard to evaluate) things, whereas for an index these are much more abstract concepts, somehow.

Bpp