Page 1 of 1

Value Averaging 100bp Premium?

Posted: Sun Oct 19, 2008 2:04 pm
by Blue
I just finished reading Edleson's Value Averaging.

The book left me skeptical of the 100+ bp potential benefit described in monte carlo simulations of using the strategy... but I was trying to place my finger on the source of skepticism. I also have inherent respect and hope for any strategy so highly recommended by Bill Bernstein. Edleson's book makes a strong case for the benefit... outpeforming DCA/periodic investing by 100+ bp in ~85-90% of investment periods & simulations.

I've read previous value averaging boglehead threads
Jan 18, 2008
July 17, 2007
March 27, 2007

And I walk away conflicted ... is there a benefit to be captured by value averaging? Why isn't everyone value averaging... where does the book (and Bill Bernstein) error if there is no benefit to the strategy?

Blue

Posted: Sun Oct 19, 2008 3:10 pm
by superlight
I value average, but I can see reasons for it to be unattractive. It's more complicated than lump sum or dollar cost averaging, for what is only a statistical advantage. The real market will do what it wants, and we can't know for years if our market will play into the hands of the value averager or the lump summer.

Posted: Sun Oct 19, 2008 3:15 pm
by InertiaMan
I also read the book, and was not so much skeptical as I was challenged with how to apply the concepts to my unique situation.

I think that VA applies more to people earlier in their accumulation phase than folks like me who are nearing the end of it. Or in the case of windfalls, I could imagine deploying a lump sum by value averaging in once a month over a year, but he didn't really cover the benefits of such short-range examples.

I also think that it could be challenging to deploy consistently, depending on the nature of market volatility. It seems like you could have LONG periods where you can't possibly meet your value line goals (during severe or sustained market underperformance) or long periods holding big amounts of cash (during prolonged market bull runs).

Posted: Sun Oct 19, 2008 3:19 pm
by PatrickS
Frequent trading policies make it difficult to implement a true value-averaging strategy if you're using funds instead of ETFs, but I use a modified strategy for reits. Essentially, I buy the required amount each month to get to my target, but don't sell when I'm over target.

Posted: Sun Oct 19, 2008 9:17 pm
by Blue
So I spent the past few hours accumulating data and backtesting the value averaging strategy over the past 20 years on Vanguards 500 index fund VFINX.

Assumed a scenario of $1000 monthly contributions beginning 3/27/1987 through Oct 2008.

I backtested buy and hold vs value averaging strategies at different expected growth rates r. Most scenarios V.A. suffered a deficit relative to Buy and hold in the neigborhood of 10-30 bp. xirr of 3.61 for buy and hold vs 3.44 for value averaging over the ~21 year period. I did not include dividends to make the modelling simpler but I do not forsee this substantially influencing the distinction between the two strategies. I also did not include transaction costs or taxes which should have favored V.A.

My results are consistent with JohnB's posts in one of the previous threads and discordant with the results portrayed in the book. Not sure I have the energy to sort out the source(s) for the discordance, but suffice it to say, I am skeptical of Bernstein's forward from the book,

Any investor fortunate enough to have come across Value Averaging during the 1990s and absorb its message was amply rewarded...
Baloney!

Blue

Re: Value Averaging 100bp Premium?

Posted: Sun Oct 19, 2008 10:47 pm
by yobria
Blue wrote:And I walk away conflicted ... is there a benefit to be captured by value averaging? Why isn't everyone value averaging... where does the book (and Bill Bernstein) error if there is no benefit to the strategy?Blue
Walking away is probably the best strategy IMO. Value averaging is kind of like alchemy - intriguing theory, waste of time in practice. In a random walk world you're not going to get an extra 100 BPS out of anything except taking more risk. Your time is far better spent keeping it simple, living w/in your means, rebalancing, etc. Even with a large lump sum I wouldn't bother w/ the complexity, txn costs, varying amount of stock exposure, etc.

Nick

Posted: Sun Oct 19, 2008 11:01 pm
by captain3d
Wether VA is good or not becomes some what irrelevant when market swings like we are seeing recently dictate hugh amounts to be invested to hit the target. Pretty soon you run out of cash.

I know because I have been doing it ;-) two weeks running I had to add 10x the normal amount to hit the target.

I still like the idea but the bigger your portfolio becomes the bigger cash reserve you will need on hand to hit the numbers.

phil

Posted: Sun Oct 19, 2008 11:20 pm
by tfb
If you include the cash reserve as part of the AA, VA is a story of over-balancing. In a down market, it wants you to dump the cash reserve into the market. In a up market, the market is hitting the target for you without new contributions. So your contributions are added to the cash reserve. You are buying more stocks in a down period, less in an up period, not much different from DCA, only magnified. If you are not too religious about a constant AA, I don't think there's anything wrong with VA. It prompts you to buy when markets are down.

Re: Value Averaging 100bp Premium?

Posted: Mon Oct 20, 2008 5:40 am
by superlight
yobria wrote:Walking away is probably the best strategy IMO. Value averaging is kind of like alchemy - intriguing theory, waste of time in practice. In a random walk world you're not going to get an extra 100 BPS out of anything except taking more risk. Your time is far better spent keeping it simple, living w/in your means, rebalancing, etc. Even with a large lump sum I wouldn't bother w/ the complexity, txn costs, varying amount of stock exposure, etc.
It doesn't strike me as complicated enough to qualify as alchemy. It's just a mechanical way to add more in down periods and less in up periods. It removes judgment and emotion from that process (once it's set up).

Given the common, and counterproductive, tendency to run from a declining market, or chase a rising market, it seems reasonable.

(The lump some alternative might work in general, over a century, but I only have to remember it was suggested to us in 1st Qtr 2008 to feel better about less random entry points.)

Posted: Mon Oct 20, 2008 5:44 am
by superlight
Blue wrote:My results are consistent with JohnB's posts in one of the previous threads and discordant with the results portrayed in the book. Not sure I have the energy to sort out the source(s) for the discordance, but suffice it to say, I am skeptical of Bernstein's forward from the book,
Here's the paper that convinced me:

http://www.studyfinance.com/jfsd/pdffil ... rshall.pdf

Do you see any error in their method, or is it just their choice of interval (monthly/quarterly) and period?

Also, did you look at mean outcomes or the distribution of outcomes?

Posted: Mon Oct 20, 2008 9:39 pm
by Blue
superlight wrote:
Blue wrote:My results are consistent with JohnB's posts in one of the previous threads and discordant with the results portrayed in the book. Not sure I have the energy to sort out the source(s) for the discordance, but suffice it to say, I am skeptical of Bernstein's forward from the book,
Here's the paper that convinced me:

http://www.studyfinance.com/jfsd/pdffil ... rshall.pdf

Do you see any error in their method, or is it just their choice of interval (monthly/quarterly) and period?

Also, did you look at mean outcomes or the distribution of outcomes?
It is a crazy few days with family and work commitments. I am looking forward to studying this article in detail later this week... my initial thought is that I restricted the monthly contributions along the value path to a fixed amount (i.e., unable to make up a complete shortfall unless there was adequate cash from previously being above the value path).... perhaps the higher internal rate of return is a function of a having a pile of cash ready to buy as much as needed on substantial dips. This is great in theory but I wonder if these articles/book consider the opportunity cost of having that much cash on hand?

Thanks for the link.... looks to be an interesting read.
Blue

Posted: Tue Oct 21, 2008 7:42 am
by superlight
I think VA is most useful to people who have a lump sum that want to put in the market/bonds. They can start the sum in money markets [or bank CDs], and move it over from there. I suppose it's a little trickier for someone contributing from monthly earnings. There one might need to build up a money market fund a bit to start with ....

Because in down times, we do need to contribute more. In June I invested "2x" in stocks. At these prices I'll need to put in "5x-6x" [in December]to stay on the formula.

Posted: Tue Oct 21, 2008 8:04 am
by jeffyscott
captain3d wrote:Wether VA is good or not becomes some what irrelevant when market swings like we are seeing recently dictate hugh amounts to be invested to hit the target. Pretty soon you run out of cash.

I know because I have been doing it ;-) two weeks running I had to add 10x the normal amount to hit the target.

I still like the idea but the bigger your portfolio becomes the bigger cash reserve you will need on hand to hit the numbers.

phil

Yes, while I did not quite completely run out of cash, I did run out of cash that I felt I could prudently add to stocks as of Oct. 6. I decided I wanted to keep some cash in reserve to add to TIPS as those suddenly became attractive. I also decided I wanted to have some left, just in case we were to go to, say, 800 on the S&P 500.

So...I now set new purchases to get back to target in about one year (sooner if the market goes up, of course).

Posted: Tue Oct 21, 2008 10:08 pm
by Blue
Interesting that both of you keep significant cash on hand to value average in on the dips. That has an opportunity cost that I do not think is factored in the Marshall article or Edleson's book.

I think it is interesting that the Marshall article compares DCA and VA, but does not include lump sum investing in the comparison.

Edleson's book also leaves lump sum out as a comparison, instead choosing DCA and constant shares as a comparison.

My impression is that it is generally agreed with the overall upward long-term trend of the market that lump sum would have higher expected returns on average over DCA.

Sure VA looks good if you have a pile of money which you can put in at any point, but if you look at the opportunity cost of having that pile of money sitting there.... then VA doesn't look so good anymore.... especially when compared to lump sum for example.... unless I am missing something?

Blue

Posted: Tue Oct 21, 2008 10:35 pm
by tfb
Blue wrote:Sure VA looks good if you have a pile of money which you can put in at any point, but if you look at the opportunity cost of having that pile of money sitting there.... then VA doesn't look so good anymore.... especially when compared to lump sum for example.... unless I am missing something?
If you take into consideration of the "side fund" you have to adjust the AA of the "main fund" upward. Say you would do 60/40 lump sum, then you keep 10% in side fund and do 70/30 in the investment portfolio. Under normal conditions you have the same AA as the lump sum. When market is down, your side fund has less cash and your AA has more than 60% in stocks. When it's up, your side fund has more cash and your AA has less than 60% in stocks. That's the idea.

Posted: Wed Oct 22, 2008 5:23 am
by superlight
Blue wrote:Interesting that both of you keep significant cash on hand to value average in on the dips. That has an opportunity cost that I do not think is factored in the Marshall article or Edleson's book.
This is easier to take when bank CDs are paying 5+% ;-)

Posted: Wed Oct 22, 2008 6:44 am
by Blue
tfb wrote:
Blue wrote:Sure VA looks good if you have a pile of money which you can put in at any point, but if you look at the opportunity cost of having that pile of money sitting there.... then VA doesn't look so good anymore.... especially when compared to lump sum for example.... unless I am missing something?
If you take into consideration of the "side fund" you have to adjust the AA of the "main fund" upward. Say you would do 60/40 lump sum, then you keep 10% in side fund and do 70/30 in the investment portfolio. Under normal conditions you have the same AA as the lump sum. When market is down, your side fund has less cash and your AA has more than 60% in stocks. When it's up, your side fund has more cash and your AA has less than 60% in stocks. That's the idea.
interesting approach... of course the question then becomes whether this approach increases risk adjusted returns beyond a fixed equity-bond ratio.

Posted: Wed Oct 22, 2008 7:09 am
by jeffyscott
Blue wrote:Interesting that both of you keep significant cash on hand to value average in on the dips. That has an opportunity cost that I do not think is factored in the Marshall article or Edleson's book.

What I was trying do is not, I think, exactly by the book. We had decided that the maximum risk level we wanted amounted to having 2X annual income in equities. Having reached that level some time ago, we were not keeping money in "cash" that we would have otherwise wanted in stocks. I am aiming for a, gradually rising, target dollar amount in each equity fund. So the principle is essentially value averaging, as I understand it.

Also our "cash" has been earning 5-5.5% in a stable value fund.

Posted: Wed Oct 22, 2008 3:21 pm
by snray02
I don't think there is a portfolio of any size that could have maintained a value averaging strategy in this market. Even a smallish $50K portfolio could easily be down 40% at this point. This would require a $20K infusion of new cash to get back to nominal value. How many $50K (now $30K) portfolios are going to have $20K "idle cash" sitting around?

I consider value averaging to be a slightly interesting academic exercise with little practical relevance.
Sam

Posted: Wed Oct 22, 2008 3:39 pm
by superlight
I think I mentioned (somewhere) that my nominal quarterly investment is "2x" and that I was looking at "5x-6x" right now. I have cash for either ... but I think I'll find it difficult psychologically to beyond 5x.