Imagine

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
larryswedroe
Posts: 16023
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Imagine

Post by larryswedroe »

It is often discussed about TLH whether to wait 31 days and reinvest or immediately.

As you know I have been proponent of not waiting for two reasons--every day you are out of market you miss the ERP. But you also have far greater chance of missing a big up then avoiding a big down. Using 5% as the definition of "big" I found 70% more likely to miss the big up than avoid the big down

Imagine harvesting losses a few days ago and sitting it out now. Already missed about 8% --of course we dont know the eventual outcome.
SVariance
Posts: 61
Joined: Fri Apr 04, 2008 12:47 pm

Re: Imagine

Post by SVariance »

larryswedroe wrote:It is often discussed about TLH whether to wait 31 days and reinvest or immediately.

As you know I have been proponent of not waiting for two reasons--every day you are out of market you miss the ERP. But you also have far greater chance of missing a big up then avoiding a big down. Using 5% as the definition of "big" I found 70% more likely to miss the big up than avoid the big down

Imagine harvesting losses a few days ago and sitting it out now. Already missed about 8% --of course we dont know the eventual outcome.
Larry,

It is just silly to wait 31 days to invest the proceeds. Who does that? In many situations, you can buy an ETF that has some but not all of the characteristics of the security that was sold.

SV
User avatar
tfb
Posts: 8411
Joined: Mon Feb 19, 2007 4:46 pm

Post by tfb »

Small ups and downs add up too. So only looking at "big" swings isn't that meaningful. Since you've looked at data, you should really look at the full 31-day period and see what the outcome is in a down market (remember you have to have a loss to begin with before you TLH). Because nobody can time the market, it's unlikely the TLH is done at the exact bottom of the market.
Harry Sit
User avatar
gummy
Posts: 340
Joined: Mon Mar 12, 2007 2:34 pm
Location: Burlington, Ontari-ari-ari-O
Contact:

Post by gummy »

Tho' I've played with missing the best or worst <U>months</U>, I've never played with <U>days</U> ... until now: :lol:
If you use 10-years worth of daily returns, but replace returns greater than 5% by 0%, you get this:
Image
which also includes replacing daily returns less than -5% by 0%, so you "miss" the really bad days, too.
User avatar
tfb
Posts: 8411
Joined: Mon Feb 19, 2007 4:46 pm

Post by tfb »

I just looked at a random fund, Vanguard Total Stock Market ETF (VTI), for 31-day price changes. Here's what I got for between 2/4/2008 and 9/19/2008 (first trading day in 2008 is 1/2/2008; waiting 31 days gets to 2/4/2008).

Code: Select all

Price Change from 31 Days Ago:
Mean             -0.61%
Standard Error    0.36%
Median           -0.86%

Bin      Frequency   Cumulative %
< -5%   26           16%
< -2%   36           39%
< -1%   14           48%
< 0%    11           54%
< 1%    05           58%
< 2%    15           67%
< 5%    33           88%
5%+     20          100%
54% of the time, the price 31 days later was lower. Not that I would do it but the "wait 31 days" method isn't as bad as you think. Spreadsheet is here if anyone is interested.

http://sheet.zoho.com/public/thefinancebuff/tlh31
Harry Sit
Topic Author
larryswedroe
Posts: 16023
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Post by larryswedroe »

this is simple

every day stocks are riskier than bonds, so every day you are out of market you are giving up the ERP. On annual basis ERP has been about 8%. So divide that by 12 and that is the ERP you gave up by being out of market on average, randomly, for 31 days.
User avatar
tfb
Posts: 8411
Joined: Mon Feb 19, 2007 4:46 pm

Post by tfb »

larryswedroe wrote:this is simple
Except when actual data did not support your theory. Remember there is short-term momentum at work.
Harry Sit
User avatar
amp
Posts: 353
Joined: Sun May 04, 2008 9:40 am

Post by amp »

tfb wrote:54% of the time, the price 31 days later was lower. Not that I would do it but the "wait 31 days" method isn't as bad as you think.
The problem is that VTI has been falling all year, so your results reflect that. I think you'd have to do this for a much longer period of time to draw any conclusions.
Sidney
Posts: 6791
Joined: Thu Mar 08, 2007 5:06 pm

Post by Sidney »

Remember there is short-term momentum at work.
How do you know when the momentum starts and stops? What is short term?
I always wanted to be a procrastinator.
User avatar
tfb
Posts: 8411
Joined: Mon Feb 19, 2007 4:46 pm

Post by tfb »

Sidney wrote:
Remember there is short-term momentum at work.
How do you know when the momentum starts and stops? What is short term?
Did I say I know? Just pointing out actual data are not as clear cut as one would think. I added some more calculation to my spreadsheet. I already showed 54% of time if one uses the "wait 30 days" method, one would get a lower price on day 31 this year. If someone extends that method and include "I'll wait for some more days if I don't get a lower price," that percentage increase to:

* 64% if one is willing to wait up to 10 more days
* 78% if one is willing to wait up to 30 more days
* 91% if one is willing to wait up to 60 more days
* 100% if one is willing to wait up to 90 more days

Of course this is done with a random fund in a random year. It's not a "study" but it does show "wait 30 days" isn't that bad.
Harry Sit
User avatar
tfb
Posts: 8411
Joined: Mon Feb 19, 2007 4:46 pm

Post by tfb »

amp wrote:
tfb wrote:54% of the time, the price 31 days later was lower. Not that I would do it but the "wait 31 days" method isn't as bad as you think.
The problem is that VTI has been falling all year, so your results reflect that. I think you'd have to do this for a much longer period of time to draw any conclusions.
I'm not drawing any conclusions. I'm just using it as a counter example. If VTI wasn't falling, one would not TLH because there is no loss to harvest.
Harry Sit
yobria
Posts: 5978
Joined: Mon Feb 19, 2007 10:58 pm
Location: SF CA USA

Re: Imagine

Post by yobria »

While selling/immediately rebuying to TLH is obvious for Boglehards and other investment geeks, I doubt most people even understand the rule.

I told a friend, who has his money with a ML broker, Thursday that I'd TLH'd a couple of ETFs and immediately bought similar ones. He asked why I'd bother to do that, and I explained the tax writeoff. "You can do that?" he asked. Apparently his broker never bothered to TLH his account.

Nick
stan1
Posts: 15559
Joined: Mon Oct 08, 2007 4:35 pm

Post by stan1 »

It's not as obvious as it seems even on this board. There are high profile contributors still advocating the "sit out 31 day" approach (perhaps because they personally don't ever want to hold anything other than Vanguard total market mutual funds).
User avatar
amp
Posts: 353
Joined: Sun May 04, 2008 9:40 am

Post by amp »

tfb wrote:If VTI wasn't falling, one would not TLH because there is no loss to harvest.
Maybe I'm reading too much into this, but you seem to be suggesting that TLH situations only arise in bear markets. Of course, that's not true.

I know this, because no matter how bullish market sentiment may be, I always seem to find myself down 10 to 15% shortly after I make a purchase.
yobria
Posts: 5978
Joined: Mon Feb 19, 2007 10:58 pm
Location: SF CA USA

Post by yobria »

amp wrote:I know this, because no matter how bullish market sentiment may be, I always seem to find myself down 10 to 15% shortly after I make a purchase.
Don't feel bad - a certain Canadian math whiz once calculated (on the M* board) that a stock investment had about an 80% chance of being lower at some point in the future than it was when you bought it.

Nick
User avatar
Taylor Larimore
Posts: 32915
Joined: Tue Feb 27, 2007 7:09 pm
Location: Miami FL

Better to miss bear markets than bull markets?

Post by Taylor Larimore »

Hi Bogleheads:

We had a big 2-day jump in the U.S. Stock Market and it appears that waiting 30 days in a money market fund when tax-loss harvesting is a mistake. No one knows what the next 30-day period will bring, but we do know that the S&P is below its price 30 days ago, so sitting in a money market fund during that period would have provided a gain unstead of a loss (and less worry).

There has been research on the subject of avoiding gains and losses. For example, I found that since 1950, the Dow's best monthly gain was +14.4% (Jan 1976) and its worst monthly loss was -23.2% (Oct 1987).

This is from InvesTech Research:

Measuring from 1928 to 2002, if you started with $10 and you followed the famous buy-and-hold strategy, that $10 would become $10,957.

If you missed the 30 best months, your $10 would only be $154.

However, if you missed the 30 worst months, your $10 would be $1,317,803.

One can see from these numbers that missing the worst periods is very important to long-run compounding."


http://moneycentral.msn.com/content/P36303.asp

When tax-loss harvesting a small part of our portfolio, this is one reason we simply sit in a risk-free money market fund for 30-days.

Best wishes.
Taylor
Topic Author
larryswedroe
Posts: 16023
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Post by larryswedroe »

again this is extremely simple
there is an ERP, every day.
thus every day you wait you are giving up that ERP

doesnt matter if one day or 31 or whatever, and the longer the period the more premium you give up
While it is true you will miss some really bad days as the worst losses are bigger than the biggest gains but that doesnt matter, the gains are more frequent, as I have shown-- big gains 70% more likely over 31 days than big losses. That is big difference
User avatar
gummy
Posts: 340
Joined: Mon Mar 12, 2007 2:34 pm
Location: Burlington, Ontari-ari-ari-O
Contact:

Post by gummy »

Measuring from 1928 to 2002, if you started with $10 and you followed the famous buy-and-hold strategy, that $10 would become $10,957.
If you missed the 30 best months, your $10 would only be $154.
However, if you missed the 30 worst months, your $10 would be $1,317,803.
It's too bad the article didn't mention if the investment was in the S&P500 ... or what (?)
I've found that conclusions based upon historical data can depend dramatically upon where the money goes.

If, instead of the S&P, one considers missing the greater than 5% days and the less than -5% days you get this
(where "missing a day or month" means replacing a return with 0%)
Examples:
Image
ImageImage

If you invest monthly (as opposed to a buy-and-hold scenario), the conclusions can also be quite different.
For example:
Image

Shucks ... so many possible assets ... so many possible time periods ... :?

P.S. I couldn't help myself ... have to post this:
Image

Mebbe we'll feel better when September is over, eh? :lol: :lol:
User avatar
Imperabo
Posts: 1109
Joined: Fri Aug 29, 2008 1:00 am

Post by Imperabo »

"I found that since 1950, the Dow's best monthly gain was +14.4% (Jan 1976) and its worst monthly loss was -23.2% (Oct 1987). "


Perfect examples. In the 6 months preceding Jan 1976 the S&P had lost about 5%. In the 6 months preceding Oct 1987 it had gained about about 8%. Only in the first case would most people have been taking losses, and they would have missed a huge rally if they had sat it out. Small sample size of course, but I personally would rather not be sitting out a market that had suffered large recent declines, and should therefore be a relative bargain. Still, not a huge deal if it's small portion of your portfolio.
User avatar
Roverdog
Posts: 482
Joined: Fri May 02, 2008 11:39 am

Post by Roverdog »

I saw the post title, and my first thought, actually, was to figure out what John Lennon has to do with investing. :)

Anyway, I'll add another consideration-- If it's necessary to sell only a small portion of one's holdings to TLH (i.e., those shares with the highest cost) rather than one's entire position, perhaps one has to consider what fraction is being sold. If it's relatively small, then waiting 31 days might be best. If it's a large portion, then transferring to a similar fund might be better. As with many other matters, it's a matter of degrees. :)

Bob
Topic Author
larryswedroe
Posts: 16023
Joined: Thu Feb 22, 2007 7:28 am
Location: St Louis MO

Post by larryswedroe »

Yes stocks are negatively skewed and exhibit excess kurtosis--so there are possibilities of very large losses

Also note in the case of TLH that you are selling AFTER a big down month. So the ERP is now higher than average.

Also on momentum, while I know it exists on individual stocks not sure it exists on asset classes. Have not seen any evidence on that.

But let's assume you have same momentum as for individual stocks. Well it looks like this. First you get REVERSAL in the short term--30 days--then momentum for months 2-12 and then reversal again. So if going to get out after the first 31 days down doesnt seem to be such a good idea since the next 31 is a reversal.
Easy Rhino
Posts: 3279
Joined: Sun Aug 05, 2007 11:13 am
Location: San Diego

Post by Easy Rhino »

larryswedroe wrote:again this is extremely simple
there is an ERP, every day.
thus every day you wait you are giving up that ERP
Since the volatility of a month of returns is usually substantially different than the ERP, there is a certain amount of risk involved. The strange part is that I can't figure if it's a risk of staying in the market, or of staying out of it.

I think the size of the TLH is going to affect your decision making. For instance, two weeks ago, I was TLHing approximately 33% of my entire portfolio in one fell swoop. Sitting in cash for a month would have been, inadvertently, market-timing.

However, if you're just TLHing a single lot in a large mutual fund position, and you're holding them with Vanguard, I can imagine just sitting in cash out of convenience.
Bimmer
Posts: 234
Joined: Tue Mar 06, 2007 9:33 pm

Behavioral Investing

Post by Bimmer »

From the behavioral perspective, isn't one nice aspect of buying immediately that it taken some of the decision-making out of the equation. Set it and forget it, make TLH an automatic process. Waiting a month almost puts a second decision point in the mix - i.e. buy on day 31 or 32, wait for another drop and be burned, Friday is day 31 so wait 'till Monday to see what happens.

Buying immediately makes this more of an automatic process, and keeps your portfolio at all times at the risk level you decided on in your investment statement.
____________________________ | Bimmer | ____________________________
User avatar
Prokofiev
Posts: 1371
Joined: Mon Feb 19, 2007 8:45 pm
Location: New Orleans

Post by Prokofiev »

I always try and buy immediately.

But the down-side to this is getting stuck in a different mutual fund/ ETF that has a higher ER or is not exactly what you want.

If after TLH, the new ETF goes down after 31 days - no problem. You just switch back to the lower ER VG ETF/fund and TLH again. However, in this case you would have been better to sit out the month in cash.

If the new ETF retraces a significant amount of the former loss (which is what you were hoping for), you now have a ST cap gain in the new ETF/fund. So if the market goes up you get stuck. And give up .2% or .3% each year to expenses. So if the ERP is 8% and you actually get the 8/12% = .75%, you might still lose that much after 2-3 years.
Last edited by Prokofiev on Tue Sep 23, 2008 12:06 am, edited 1 time in total.
Everything should be made as simple as possible, but not simpler - Einstein
User avatar
mudfud
Posts: 1241
Joined: Tue Feb 20, 2007 3:34 pm

Post by mudfud »

tfb wrote:
I added some more calculation to my spreadsheet. I already showed 54% of time if one uses the "wait 30 days" method, one would get a lower price on day 31 this year. If someone extends that method and include "I'll wait for some more days if I don't get a lower price," that percentage increase to:

* 64% if one is willing to wait up to 10 more days
* 78% if one is willing to wait up to 30 more days
* 91% if one is willing to wait up to 60 more days
* 100% if one is willing to wait up to 90 more days

Of course this is done with a random fund in a random year. It's not a "study" but it does show "wait 30 days" isn't that bad.
Excellent analysis, tfb.

I'd also point readers to tfb's outstanding blog post on the "10-best days" fallacy.
http://thefinancebuff.com/2007/08/more- ... -days.html
"Are you sure you have tested an a priori hypothesis?" | | Image
Post Reply