Buying the dips: neither good nor bad, it seems to me

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
grayfox
Posts: 5569
Joined: Sat Sep 15, 2007 4:30 am

Post by grayfox »

grayfox,

Unfortunately, this strategy does sound familiar. It sounds to me like P/E market timing. I'm sure other Diehards on this forum can provide Monte Carlo stress-tested results for that backtest, and it ain't pretty.Embarassed
Actually what I had in mind was the TIPS strategy proposed by one of the highly respected authors that posts here. That is why I used 2.5% and the current 1.8% which was the 10-Year TIPS rate earlier today. You know, buy TIPS above 2.5 and sell below 1.5.
Valuethinker
Posts: 41160
Joined: Fri May 11, 2007 11:07 am

Post by Valuethinker »

grayfox wrote:
grayfox,

Unfortunately, this strategy does sound familiar. It sounds to me like P/E market timing. I'm sure other Diehards on this forum can provide Monte Carlo stress-tested results for that backtest, and it ain't pretty.Embarassed
Actually what I had in mind was the TIPS strategy proposed by one of the highly respected authors that posts here. That is why I used 2.5% and the current 1.8% which was the 10-Year TIPS rate earlier today. You know, buy TIPS above 2.5 and sell below 1.5.
I think this is more or less Larry Swedroe's strategy.

Basically there are macroeconomic reasons why real interest rates can't get too low, and likelihoods why they can't get too high. In extremis we could imagine situations where real rates get to 5%, say, but they are pretty severe ones (Fed raises rates sharply to stop a run on the dollar, in which case TIPS are still likely to do better than almost any other USD asset).

In principle real return bonds should have the lowest return of any asset other than cash, because they are the safest asset. However they should retain portfolio diversification benefits (unique positive correlation with inflation). Historically, they have not had significantly lower returns than nominal bonds, despite the lower risk, a 'free lunch'.
Eric White
Posts: 54
Joined: Fri May 18, 2007 10:09 am

Apology to grayfox; new question of "buying on yield sp

Post by Eric White »

grayfox wrote:
grayfox,

Unfortunately, this strategy does sound familiar. It sounds to me like P/E market timing. I'm sure other Diehards on this forum can provide Monte Carlo stress-tested results for that backtest, and it ain't pretty.Embarassed
Actually what I had in mind was the TIPS strategy proposed by one of the highly respected authors that posts here. That is why I used 2.5% and the current 1.8% which was the 10-Year TIPS rate earlier today. You know, buy TIPS above 2.5 and sell below 1.5.
Grayfox,

Now that I have a little background, I think I may be starting to understand the trajectory of the response. I have not read the TIPs/Swedroe thread yet, so I cannot honestly comment on it until I understand it. If you can respond with a link to the thread, I'll get up to speed on it so we can have a constructive converstation relative to "buying on dips."

I'm runnin' on 3 hours sleep with a newborn waking up every 90 minutes. :shock: Sorry if I overreacted!

Do you think "buying on dips" is fundamentally different for fixed versus equity? Or even various subclasses of fixed? It definitely makes

If so, do you think it would be worthwhile or even possible to model and compare "buying on nominal yield spikes", "buying on real yield spikes" and "buying on equity price dips"? It seems like if forum members think it is an issue, it should be vetted so that we understand the impact across our asset allocations.
diasurfer
Posts: 1845
Joined: Fri Jul 06, 2007 8:33 pm
Location: miami-dade

Post by diasurfer »

Your results are interesting but I agree that you need to compare to a DCA at the beginning of the month. The interesting results is if buying on dips (ie waiting to buy, so your money is in the market for less time) can beat having your money invested for the longest time possible. I think your results could also be improved by re-defining the "first day of the month" as "pay day" which could be any day of the month, repeating the simulation for all possible pay days (or at least the 1st - 28th), and then averaging the results.

So the comparison is:
1. pure DCA: I get paid on the Xth of the month and invest on Xth of the month.
2. DCA with dip: I get paid on the Xth of the month. If Xth of the month is down Y percent from the day before, I buy on the Xth. If not, I wait until it is down Y percent from the day before. If it is not down by Y by the Xth of the next month, I buy on the Xth of the next month.

Repeat for X = 1 ... 28 and average results. For each X, Y is a range of threshold values.

And please post results!
User avatar
United
Posts: 447
Joined: Wed Mar 28, 2007 5:45 pm

Post by United »

When the TIPS rate goes up, the expected returns of TIPS go up. Thus you should increase your allocation.


When the stock market goes down, the expected returns of the stock market do NOT go up. Thus, you should not change your allocation in response to market declines (e.g., moving cash reserves into stocks).

The analogy with TIPS does not hold.
User avatar
United
Posts: 447
Joined: Wed Mar 28, 2007 5:45 pm

Post by United »

Here's how this "tactic" is like the martingale system (characterized by common small positive returns and rare large negative returns):


Let's say you implement this tactic using monthly investments after 3% drops.

Most months, an index level will drop 3% below its starting price. When this occurs, you will gain 3% more by waiting for the drop instead of investing immedietely.

However, during some months, the price level of an index will not drop below 3% of its starting value. On these months, you will miss out on the entire month's worth of gains if you wait for dips (and remember, a month's worth of gains will be larger than average if the index level never drops 3% below its starting value).

The gains that you miss by waiting for 3% drops are rarer (but larger) than the gains you cash. Overall, waiting for drops will reduce your returns.
User avatar
United
Posts: 447
Joined: Wed Mar 28, 2007 5:45 pm

Post by United »

I have one last comment.

If this tactic actually did work, investing after drops would not be the ideal way to implement it. Ideally, you should hold a portfolio of equities and fixed income. The more the market drops, the more you should allocate to equities. Then, at the beginning of each month, reset your allocation. Only applying this tactic to new investment money and not current investments as well seems very illogical to me. The source of the money is irrelevant.
User avatar
Topic Author
nisiprius
Advisory Board
Posts: 41995
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

The martingale

Post by nisiprius »

Eric White wrote:For instance, United, I would really like to understand the Martingale process you were describing and if there is a way to use it instead of this brute-force method.
Uh, the martingale is a well-known gambling system that is so simple that it's fairly easy to understand why it doesn't work. Here's one form of it. Do you see the problem in it?

Let's suppose you're playing roulette. The ordinary gambler loses slowly and steadily at a rate determined by how often the 0 and 00 come up. But the martingale-system gambler wins steadily, guaranteed. It's simple. Suppose you wish to average winnings of $10 per spin.

On spin #1 you bet $10 on red. If you win, you've won $10 in one spin and the sequence is complete.

If spin #1 comes up black, you bet $30 on red for spin #2. If you win, you've now won $30 - $10 = $20 on two spins of the wheel, and the sequence is complete.

If spin #2 comes up black, you bet $70 on red for spin #3. If you win, you've now won $70 - $30 - $10 = $30 on three spins of the wheel, and the sequence is complete.

If spin #3 comes up black, you bet $150 on red for spin #4. If you win, you've now won $150 - $70 - $30 - $10 = $40 on four spins of the wheel, and the sequence is complete

You keep betting red until it comes up, betting (2^N - 1)*$10 on spin #N. I.e. the sequence of bets is the doubling sequence $20, $40, $80, $160, $320, etc. with $10 subtracted from each number. With this sequence of bets, you can show that whenever red does come up, your winnings on spin #N exceed your losses on all the previous spins by $10 times the number of spins. This math is correct.

Since red must always come up eventually, by following the system you can take what is normally a loser's game and achieve guaranteed winnings of $10 per spin of the wheel. Notice even when the odds are slightly against the gambler, as they are in roulette, it still works.

Or not.

Do you see what the problem is? It's not with the math. The doubling sequence really does achieve what it's supposed to achieve.

It's martingale with a small "m." It's not a person's name, it's named for a kind of horse harness in which the straps have a sort of doubling-up pattern. The fact that it's named for a horse harness should give you an idea of how old it is.
Valuethinker
Posts: 41160
Joined: Fri May 11, 2007 11:07 am

Post by Valuethinker »

United wrote:Here's how this "tactic" is like the martingale system (characterized by common small positive returns and rare large negative returns):

.
Brilliant piece in the New Yorker financial column this week about the consequences of this.

Since hedge fund managers (and corporate execs with large stock options) make money if they perform on a 12 month basis, they have every incentive to pursue high risk martingale strategies.

they make money and get paid 20% of the upside. Every so often, they lose the investors 100%, walk away, and get another job (or retire).

Martingales is precisely what many (most, all?) hedge funds are doing.
User avatar
stratton
Posts: 11083
Joined: Sun Mar 04, 2007 5:05 pm
Location: Puget Sound

Post by stratton »

Valuethinker wrote:Since hedge fund managers (and corporate execs with large stock options) make money if they perform on a 12 month basis, they have every incentive to pursue high risk martingale strategies.

they make money and get paid 20% of the upside. Every so often, they lose the investors 100%, walk away, and get another job (or retire).

Martingales is precisely what many (most, all?) hedge funds are doing.
A little irregularity like a credit crunch means lots of hedge funds explode all at once. Because of the excessive leverage we have thousands of little black swan breeding grounds being actively encouraged.

Are those hedge fund folks trying to get themselves regulated like a mutual fund? That would solve a lot of problems without any extra rules other than passing the rule to do it. :twisted:

It doesn't even keep them from engaging in hedge fund type activities. It does force them to be transparent and removes huge amounts of risks because of brakes on leverage etc.

Things are getting tightly coupled to a ridiculous amount in the financial world.

Paul
Valuethinker
Posts: 41160
Joined: Fri May 11, 2007 11:07 am

Post by Valuethinker »

stratton wrote:
Valuethinker wrote:Since hedge fund managers (and corporate execs with large stock options) make money if they perform on a 12 month basis, they have every incentive to pursue high risk martingale strategies.

they make money and get paid 20% of the upside. Every so often, they lose the investors 100%, walk away, and get another job (or retire).

Martingales is precisely what many (most, all?) hedge funds are doing.
A little irregularity like a credit crunch means lots of hedge funds explode all at once. Because of the excessive leverage we have thousands of little black swan breeding grounds being actively encouraged.

Are those hedge fund folks trying to get themselves regulated like a mutual fund? That would solve a lot of problems without any extra rules other than passing the rule to do it. :twisted:

It doesn't even keep them from engaging in hedge fund type activities. It does force them to be transparent and removes huge amounts of risks because of brakes on leverage etc.

Things are getting tightly coupled to a ridiculous amount in the financial world.

Paul
I am not sure if regulation is the answer to the hedge fund problem but I agree transparency would be a good thing in general.

One good thing about hedge funds is the multiplicity of strategies. there are funds *up* 200, 300% this year on the sub prime crisis due to aggressive shorting strategies.

This cross coupling that you cite is a key factor in the spread of financial crashes.

The eternal rule applies (more or less, with caveats). If a bank loses $1 of capital via losses, it will reduce lending by $12.50, and this will ripple down throughout the economy.
User avatar
VictoriaF
Posts: 19431
Joined: Tue Feb 27, 2007 7:27 am
Location: Black Swan Lake

Post by VictoriaF »

Any strategy that requires one to monitor markets closely and make multiple decisions is bound to backfire on the psychological level.

You invest on a dip, and it turns out a great move. Your brain gets a boost of some chemicals (I don't have Jason Zweig's book in front of me), and you get an equivalent of a drug induced high state. You are not realizing it, but you are getting addicted to an equivalent of gambling. Next month's dip is not as deep as last month's, and you don't get an expected high. You start tweaking your strategy. You are glued to market price changes and are waiting for the right moment. The moment comes, and you buy more than you planned initially. The dip becomes a prolonged downturn...

Victoria
Last edited by VictoriaF on Sat Jan 05, 2008 8:53 pm, edited 1 time in total.
diasurfer
Posts: 1845
Joined: Fri Jul 06, 2007 8:33 pm
Location: miami-dade

Post by diasurfer »

United wrote:I have one last comment.

If this tactic actually did work, investing after drops would not be the ideal way to implement it. Ideally, you should hold a portfolio of equities and fixed income. The more the market drops, the more you should allocate to equities. Then, at the beginning of each month, reset your allocation. Only applying this tactic to new investment money and not current investments as well seems very illogical to me. The source of the money is irrelevant.
True, but I think most people just look at this as a way to tweak their overall investment strategy a bit, and not bet the house. A lot of people tilt to small value and emerging markets, but they don't invest everything in small value and emerging markets. Plus, if I bought and sold Total International every month, I would start to violate rules that would incur penalties or forbid buying back in for 60 days.

I expect that if the OP ran the numbers as I described, we should see United's claim that missing out on gains outweighs buying on dips and the scheme reduces returns. Still, it would be nice to see it.

For an even more realistic scenario, the investor would hold his pay in Prime MM until the buy threshold is reached. That will reduce the advantage buying equities on payday has which comes from being in the market longer.
grayfox
Posts: 5569
Joined: Sat Sep 15, 2007 4:30 am

Post by grayfox »

Eric White wrote:
Now that I have a little background, I think I may be starting to understand the trajectory of the response. I have not read the TIPs/Swedroe thread yet, so I cannot honestly comment on it until I understand it. If you can respond with a link to the thread, I'll get up to speed on it so we can have a constructive converstation relative to "buying on dips.
I tried searching for a thread that discusses that the TIPS strategy, but came up with 100s of hits. The strategy is discussed in Larry's bond book and has been discussed here or on M* many many times. The is also a shifting-maturity strategy that locks on higher rates for longer terms. The strategy is best for individual TIPS but can also work with funds.
Do you think "buying on dips" is fundamentally different for fixed versus equity? Or even various subclasses of fixed?
Now if it works for bonds can it work for equities? I don't know. Apparently there is a range of real returns for bonds. Is there a range of real returns for equities?

There was a recent thread which showed that the long term expected return of stocks is e/p, where e is some measure of "permanent" earnings. So if p goes down (dip) with e unchanged, then expected return goes up. So it still seems to me buying at lower prices gives a higher expected return.
livesoft
Posts: 73348
Joined: Thu Mar 01, 2007 8:00 pm

Post by livesoft »

VictoriaF wrote:...
You invest on a dip, and it turns out a great move. Your brain gets a boost of some chemicals (I don't have Jason Zweig's book in front of me), and you get an equivalent of a drug induced high state. You are not realizing it, but you are getting addicted to an equivalent of gambling. Next month's dip is not as deep as last month's, and you don't get an expected high. You start tweaking your strategy. You are glued to market price changes and are waiting for the right moment. The moment comes, and you buy more than you planned initially. The dip becomes a prolonged downturn...
Hi Victoria, I have recently read Mr Zweig's book and you have done a great synopsis of one aspect of his writings.

But you write as if buying is a bad thing. I will say that I have been buying on the dips for more than a decade. And I give credit to the drug-induced high that addicted me to buying in 2000, 2001, and 2002 as the general market kept going lower and lower. Those buys have turned out really well for me. And certainly the years 2000-2002 could be characterized as a prolonged downturn. But they were worthy years to be buying.

Now if I had started investing in summer of 2000 and my brain was not already wired to expect a pop after a dip, then my brain probably would be wired now to expect a bigger dip and disappointment after a dip. And I would be writing a different response: "Don't buy on the dips, the market is only going to go lower! Don't try to catch a falling knife!" And I would not have enjoyed the great returns from late 2002 through 2007 as much either.
User avatar
VictoriaF
Posts: 19431
Joined: Tue Feb 27, 2007 7:27 am
Location: Black Swan Lake

Post by VictoriaF »

livesoft wrote:
VictoriaF wrote:...
You invest on a dip, and it turns out a great move. Your brain gets a boost of some chemicals (I don't have Jason Zweig's book in front of me), and you get an equivalent of a drug induced high state. You are not realizing it, but you are getting addicted to an equivalent of gambling. Next month's dip is not as deep as last month's, and you don't get an expected high. You start tweaking your strategy. You are glued to market price changes and are waiting for the right moment. The moment comes, and you buy more than you planned initially. The dip becomes a prolonged downturn...
Hi Victoria, I have recently read Mr Zweig's book and you have done a great synopsis of one aspect of his writings.

But you write as if buying is a bad thing.
Buying is a good thing, but buying on dips may backfire. The safest approach is to use an algorithm which you cannot change as you please, e.g., a payroll deduction to add new money and annual rebalancing.
livesoft wrote:I will say that I have been buying on the dips for more than a decade. And I give credit to the drug-induced high that addicted me to buying in 2000, 2001, and 2002 as the general market kept going lower and lower. Those buys have turned out really well for me. And certainly the years 2000-2002 could be characterized as a prolonged downturn. But they were worthy years to be buying.

Now if I had started investing in summer of 2000 and my brain was not already wired to expect a pop after a dip, then my brain probably would be wired now to expect a bigger dip and disappointment after a dip. And I would be writing a different response: "Don't buy on the dips, the market is only going to go lower! Don't try to catch a falling knife!" And I would not have enjoyed the great returns from late 2002 through 2007 as much either.
You did well between 2002 and 2007. But there are some dangers. If there is a dip that looks similar to 2000-2002, you may use your old approach, and it may not work the next time around.

Victoria
Last edited by VictoriaF on Sat Jan 05, 2008 8:54 pm, edited 1 time in total.
livesoft
Posts: 73348
Joined: Thu Mar 01, 2007 8:00 pm

Post by livesoft »

Woo-hoo!!!
truepusk
Posts: 38
Joined: Thu Nov 29, 2007 10:51 pm

Post by truepusk »

Two important points - one for and one against this strategy - that I think have largely been missed by other posters:

1. Market psychology and contrarian investing - Overall markets and people tend to be prone to hysteria. Thus, often times when the stock market it is down there tends to be undervaluation (yes, yes, I know, how do you determine the 'real' value...). In general it can be a good time to buy low. I'm new to this site, but I get the impression that because of how heavily people subscribe to the random walk theory that there may be a lot who disagree with this point.

2. General up-treanding - The stock market tends to move up in general thus on average, the longer you wait to get in the more returns you are missing out on.

I've seen some articles and studies recently that argue against dollar value averaging. They say that if you have a lump of money that you want to move into the market, your best returns are by moving it in right away. (I'm not sure about slowly changing your allocations over time though, so I don't know if I totally agree with this). However, if we invest with each paycheck overtime, I believe this is a type of dollar value averaging that does make sense.

Thinking about this, what might be smart is to invest every pay check, but invest a little extra when the stock market has gone down recently, if you can afford it. Or the previous poster who mentioned borrowing against cash to buy earlier in the pay cycle when it goes down. But again, I don't know if these strategies will yield too big of a difference in the long run. I agree with the bogleheads that steady investing over the long haul is the best approach!
buyza
Posts: 417
Joined: Fri Sep 14, 2007 10:16 pm

Post by buyza »

livesoft wrote:Woo-hoo!!!
Ditto. I had to ring the register on some of my gains though. Not sticking around Emerging markets after grabbing almost 10% in the same week I buy.

I am actually in the black overall from buying down into the market. Only by about 1% but still nice. 2% if you count VWO earnings.
pop77
Posts: 466
Joined: Wed Dec 12, 2007 7:57 pm

Idea for Backtesting ATTN: Eric White

Post by pop77 »

First I want congratulate the detailed and original thinkers in this blog. I am new to this blog and this is my first post.
Here is an idea. One of the bloggers have already kind of mentioned it, here is what I do. If I have the potential to save 100 dollars a month I do not DCA all the 100 dollars. I invest only 75 dollars and put the remaining in a reserve (typically money market sweep). If the fund I invest drops by 2% I buy 75 more dollars of the fund. The problem however is if the Fund/Stock you own is falling consistently you run out of the reserve. I think this is one way a balancing act as it prevents you in throwing more money at some thing that is falling consistently (Don't catch a falling knife).
So if one could back test this strategy of investing 75% in regular DCA and remaining 25% (from reserver) when the fund falls 2% in the original amount
we can see whether that helps.
TheEternalVortex
Posts: 2569
Joined: Tue Feb 27, 2007 9:17 pm
Location: San Jose, CA

Re: Idea for Backtesting ATTN: Eric White

Post by TheEternalVortex »

pop77 wrote:First I want congratulate the detailed and original thinkers in this blog. I am new to this blog and this is my first post.
Here is an idea. One of the bloggers have already kind of mentioned it, here is what I do. If I have the potential to save 100 dollars a month I do not DCA all the 100 dollars. I invest only 75 dollars and put the remaining in a reserve (typically money market sweep). If the fund I invest drops by 2% I buy 75 more dollars of the fund. The problem however is if the Fund/Stock you own is falling consistently you run out of the reserve. I think this is one way a balancing act as it prevents you in throwing more money at some thing that is falling consistently (Don't catch a falling knife).
So if one could back test this strategy of investing 75% in regular DCA and remaining 25% (from reserver) when the fund falls 2% in the original amount
we can see whether that helps.
The best strategy is always to invest the money as soon as possible. However the day of the month you invest eventually matters extremely little in either case.
waitforit
Posts: 322
Joined: Wed May 23, 2007 8:28 am

Post by waitforit »

I find this thread very interesting, especially this quote:
what might be smart is to invest every pay check, but invest a little extra when the stock market has gone down recently, if you can afford it.
I was going to experiment doing this in my taxable account.. actually, I was going to study it retrospectively to see what sort of impact it would have.

Scenario One: Invest $500/mo according to allocation on a fixed schedule (ie, 28th of the month or whatever)
Scenario Two: Follow scenario one, but make additional investment $I every time broad market indicator N falls X% over Y time-frame.

If anyone knows of any research / papers on this topic I'd love to read them. I don't even know what to search for - what is this strategy called? Is it valid?
Post Reply