When Does Dollar Cost Averaging Work?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
alluringreality
Posts: 1511
Joined: Tue Nov 12, 2019 9:59 am

Re: When Does Dollar Cost Averaging Work?

Post by alluringreality »

tomsense76 wrote: Thu Jul 15, 2021 7:11 pm If trying to jump in with both feet is too hard, DCAing works in that you are now invested. Is your performance worse than if you lump summed? Most likely. Is it better than continuing to sit in cash for the entire time period? Almost certainly! What's the best behaviorally? The one you can follow through with.
Thank you for posting. In my opinion you did a far better job of interjecting the sort of practical or behavioral perspective that I attempted. I'm not going to reply with a point by point restatement of my position, because it's simply not worth the time to note that DCA can potentially end up being a useful alternative to a one time lump-sum investment for some people or situations. It is possible that price might happen to rise or fall across or beyond a DCA period, yet a basic point of using DCA is to invest assets regardless of how the future plays out, which some people choose to accomplish with a periodic investing plan. Someone that chooses to invest a significant portion of assets with DCA may simply have a different personal intent for "best" or "risk" than another person that chooses a lump-sum for their own considerations.
45% US Indexes, 25% Ex-US Indexes, 30% Fixed Income - Buy & Hold
Thesaints
Posts: 5108
Joined: Tue Jun 20, 2017 12:25 am

Re: When Does Dollar Cost Averaging Work?

Post by Thesaints »

NiceUnparticularMan wrote: Fri Jul 16, 2021 5:36 am You're just changing the hypothetical. Maybe this will help.

Assets are priced at $90 at Time 1. $100 at T2. $110 at T3.

Investor A invests $100 at Time 1. Investor B invests $33.33 at each of T1, T2, and T3. Investor C invests $100 at T2.
...and Investor D invests $100 at T3. Or do you think lump-sum investors have a way to avoid market peaks ?

Do you think there is a material difference between B and C? But I think you have now said no.
No, there isn't. But while the DCA investor is assured to get a near-average price the lump-sum investor maybe he gets it, maybe he gets a much lower price, or maybe he gets a much higher price; nobody can tel in advance which is going to be. That is the source of the increased volatility in his return.
Right, but my point is this is true in the following case too:

Investors A and B both have $100 in cash.

Investor A invests at T1 at price P1. Investor B keeps his money in cash for 3 years. Then Investor B invests at T2 at price P2.

It is just as true that compared to A, B maybe gets a higher price, and maybe a lower price. So it is just as true that spending 3 years invested in cash will reduce the volatility of B's returns relative to A's.
That is not correct. Volatility of Investor B is identical. He too picks a single entry point, which is exposed to the full volatility of the underlying investment.
You are placing all the emphasis on the DCA getting an "average" price, but that's not where the lower volatility of returns comes from! The lower volatility comes from the fact that WHENEVER you hold cash instead of assets with price uncertainty, you reduce the volatility of returns. DCA is just one way to hold cash, but any other way has the same sort of effect.
Of course is true that if you always hold cash, your return volatility is lower, but here we are talking about investors who want the same asset allocation, eventually. The issue is how should one achieve that AA.


Right, so changing the example above, now imagine the price sequence is:

T1 $90
T2 $150
T3 $100

Investor A invests $100 at T1. Investor B invests $33.33 at each period. Investor C invests $100 at T3.

Investor B doesn't pay an average price between A and C, they pay a higher average price than either. What this shows is just that by holding cash even longer before investing, Investor C has further reduced the volatility of returns even versus the DCA investor.
you forgot about Investor D, who invests $100 at T2. You have to take into account all the LS investors, not just the luckier ones.

Of course there is. Holding aside tax consequences, Investor B can withdraw the $100K from his Target fund.

Again, you're just resisting the hypothetical, but it is very much an option for many investors, and in fact people do it all the time--sell off into cash when they are getting concerned about short term losses.
Focussing on the residual 100k still to be invested, he now has the choice between rolling the dice and pick a single entry point, or distribute his investment amongst several different entry points.
Not that hard of a concept, is it ?
And Investor A has that option too! Pull out the same $100K, and then reinvest it at several different reentry points.

I think on some level you know the financial consequences of Investor A doing that (again holding aside things like tax issues) are identical to the financial consequences of Investor B doing that.

So if the financial consequences are the benefit to B, the same logic should apply to A.
[/quote]

Do you think losses incurred previously can be neglected ? When you say "Investor A can sell his stocks and DCA back in" means that when he sells his stocks he will realize a loss (or a gain) depending on his entry and his exit point. That too counts; in finance past is not fiorgotten.
You are correct that going forward they will have the same volatility of returns.
blackball
Posts: 107
Joined: Thu May 01, 2008 9:36 pm

Re: When Does Dollar Cost Averaging Work?

Post by blackball »

Does dollar cost averaging work when selling? Is the answer the opposite of dca for buying?
dbr
Posts: 46181
Joined: Sun Mar 04, 2007 8:50 am

Re: When Does Dollar Cost Averaging Work?

Post by dbr »

blackball wrote: Fri Jul 16, 2021 10:03 am Does dollar cost averaging work when selling? Is the answer the opposite of dca for buying?
If the debate is DCA vs lump sum the transaction is selling one thing and buying another. People usually assume the thing sold is a low risk holding, such as cash, which is not usually thought of as selling but it is. What is actually going on is that a person is changing their asset allocation and doesn't know if they should do it all at once now or in steps over time. The situation is symmetric except that the risk is opposite. People get very upset about the proposition of increasing risk but never seem to ask if it is a problem to reduce risk all at once. People bail from a risky market all the time and don't stop to worry that they might miss a large market gain. The asymmetry is that there is more aversion to losing money than there is to failing to gain money.
Thesaints
Posts: 5108
Joined: Tue Jun 20, 2017 12:25 am

Re: When Does Dollar Cost Averaging Work?

Post by Thesaints »

blackball wrote: Fri Jul 16, 2021 10:03 am Does dollar cost averaging work when selling? Is the answer the opposite of dca for buying?
When buying one can take advantage of the asset volatility by investing the same amount repeatedly. His average acquisition price will be lower than the average asset price; more so the higher the asset volatility is.
When selling, unfortunately it is the opposite. Selling to obtain equal receipts in money, causes the average selling price to be lower than the average price of the underlying investment. One should instead sell equal number of shares, which unfortunately causes the receipt to be variable.
Last edited by Thesaints on Sat Jul 17, 2021 2:06 pm, edited 1 time in total.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

Thesaints wrote: Fri Jul 16, 2021 9:45 am Or do you think lump-sum investors have a way to avoid market peaks ?
No, of course not. The only point I was making is once you are done buying, whether the price you bought at was one price or the average of several prices does not matter.
That is not correct. Volatility of Investor B is identical. He too picks a single entry point, which is exposed to the full volatility of the underlying investment.
I then don't understand how you are purporting to calculate long-term volatility of returns.

Let's do another simplified example.

We are looking at volatility of returns at time T.

Investors A and B both get $100K in cash 20 years before T. Investor A invests in a risky portfolio immediately. Investor B keeps it in cash for 19 years and 364 days, then invests it in the risky portfolio the day before T.

Do you not think Investor B will have a lower volatility of returns at T than A?

OK, now Investor C also gets $100K in cash 20 years before T. Investor C does a DCA between the beginning and end of the first year.

Do you think Investor C will have a lower volatility of returns at T than B?

What is true is Investor C will have a lower volatility of returns at T than A. That is because Investor C is a mild variant on Investor B--Investor C held some of their investment in cash for some period of time, and that reduces the volatility of returns at any later point in time. But since they only held cash for a limited period of time, it will not do as much to reduce the volatility of returns as holding cash for longer periods of time.
Of course is true that if you always hold cash, your return volatility is lower, but here we are talking about investors who want the same asset allocation, eventually. The issue is how should one achieve that AA.
Well, in the example above, we met this condition! Investor B also got to the same AA "eventually"! Sure, just the day before, but they still did it.

And we can do all sorts of variants in between. Another investor holds cash for 10 years then buys. That investor will have a higher volatility of returns than an investor who waits until the day before, but a lower volatility of returns than an investor who DCA over the first year.

Again, all we are showing here is holding cash for a while reduces the volatility of returns.
Do you think losses incurred previously can be neglected ? When you say "Investor A can sell his stocks and DCA back in" means that when he sells his stocks he will realize a loss (or a gain) depending on his entry and his exit point. That too counts; in finance past is not fiorgotten.
You are correct that going forward they will have the same volatility of returns.
Neglected for what purpose? Counts for what purpose?

For tax purposes, it might matter, but I am holding those aside.

For the sorts of financial purposes we are discussing, it is irrelevant. Whether on the day in question the investor got the cash from an employment bonus, inheritance, bank robbery, selling risky assets, or so on won't make any difference.

So, we can imagine we are talking about someone with a Target fund in something like a 401K, such that there are no tax consequences.

And so if holding aside taxes, there are financial reasons to prefer DCA over that period, the same financial reasons should apply to converting to cash and then reinvesting over that period.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

blackball wrote: Fri Jul 16, 2021 10:03 am Does dollar cost averaging work when selling? Is the answer the opposite of dca for buying?
So the basic reason DCA doesn't "work" when buying is assets with a positive expected return are more likely to increase in price over time. So, holding onto cash and waiting to buy tends to just mean buying at a higher price.

What happens when selling depends on what you mean by doing the opposite. But, suppose that means something like you would ordinarily plan to sell at T, but instead you will start gradually selling some period of time before T.

In that case, you will tend to be selling at a lower price, so it is equally likely not to "work". On the other hand, doing that will reduce volatility in terms of how much you have in the end.

Again, all this is really doing is illustrating that when a risky asset has a positive expected return, more time invested tends to mean higher returns.

But if you want to sacrifice expected returns for lower volatility, one option is to allocate some of those funds to cash instead. DCA, or the opposite in selling form, are particular ways of doing that. But you could do a lot of other things to increase your cash holdings.
seajay
Posts: 1656
Joined: Sat May 01, 2021 3:26 pm
Contact:

Re: When Does Dollar Cost Averaging Work?

Post by seajay »

Lump sum into all-stock at the start of April 2000, and lump out at the end of March 2009 and you'd have achieved a -7% annualised real total return. Average in/out around those peak-to-trough dates/points and the losses would have been lower. Holding stock and bonds and rebalancing between them is yet another form of time/cost averaging/DCA'ing. 50/50 tends not be the most rewarding nor the worst case. More often you'd do better with all-stock reward wise, but many are more concerned about reducing the risk of the worst cases than maximizing the upside gains, when so averaging in/out, and 50/50 (whatever) rebalancing is more inclined to avoid the worst case lump-all-in/lump-all-out/concentration-risk outcome.
alluringreality
Posts: 1511
Joined: Tue Nov 12, 2019 9:59 am

Re: When Does Dollar Cost Averaging Work?

Post by alluringreality »

NiceUnparticularMan wrote: Sat Jul 17, 2021 3:58 am For the sorts of financial purposes we are discussing, it is irrelevant. Whether on the day in question the investor got the cash from an employment bonus, inheritance, bank robbery, selling risky assets, or so on won't make any difference.
Your replies read like you may not have actual experience, or lack a fear of loss, investing a considerable portion of new assets. For example, DCA discussions often concern considerably more assets than the 11% new investment example. Some people simply will not agree with you that near-term losses are completely irrelevant, and emotions can be involved with a rare event like investing a large portion of new assets. Everyone will not agree on which risk and reward choices they prefer.
http://www.efficientfrontier.com/ef/997/dca.htm

A portion of DCA focus relates with behavioral concepts, and there are various ways that someone might approach how they personally choose to deal with emotions or potential regret. This forum seems to heavily favor lump-sum for all situations and people. My main concern is that often individuals essentially giving advice are clearly operating from a position where the "house money effect" is in their own favor, yet someone looking for guidance may not be in that same position. The following is a restatement of my comment.

"Many new investors are more interested in minimizing their potential loss, and it's important that an ill-timed market drop not scare them off from investing in the future."
https://www.bogleheads.org/wiki/Dollar_cost_averaging
45% US Indexes, 25% Ex-US Indexes, 30% Fixed Income - Buy & Hold
dbr
Posts: 46181
Joined: Sun Mar 04, 2007 8:50 am

Re: When Does Dollar Cost Averaging Work?

Post by dbr »

alluringreality wrote: Sat Jul 17, 2021 9:40 am
"Many new investors are more interested in minimizing their potential loss, and it's important that an ill-timed market drop not scare them off from investing in the future."
https://www.bogleheads.org/wiki/Dollar_cost_averaging
This has always been a reasonable consideration. I personally think that behind this concern is a lack of certainty and understanding what long term asset allocation the investor wants and also lack of experience or understanding of the uncertain and volatile nature of investment returns on any time scale, during or after placing an investment. The idea that DCA somehow actually gets you more money is a different discussion.

In any case it is often that dilemmas at one level are actually symptoms of not resolving something more fundamental at a higher level, or at an underlying level depending on which direction one draws one's charts.

I keep trying in these conversations to see if I can get the OP in each case to explain why the alternative of DCA even occurs but have had little success getting a coherent answer from most people who are asking the question.
bantam222
Posts: 136
Joined: Sun Nov 23, 2014 4:57 pm

Re: When Does Dollar Cost Averaging Work?

Post by bantam222 »

The only pro of DCA if it gives the investor better peace of mind.

If you are considering DCA from x —> y, that means you have determined y is the ideal long term portfolio.

Every day you are performing the DCA, you are in an non optimal portfolio.

Regarding comments about the markets being overvalued (btw people were saying this a few years ago too - look where we are at now, no one can predict the future) - if you feel strongly about this, this means your end state portfolio (y in above example) is too aggressive and needs to be revisited. DCA will not help as it basically slows the time for you to build into a final asset allocation you do not think is ideal. Maybe you would be better off lump sum into an asset allocation you do think is ideal.

Net net - you want to optimize the amount of time you are in the optimal asset allocation.
Thesaints
Posts: 5108
Joined: Tue Jun 20, 2017 12:25 am

Re: When Does Dollar Cost Averaging Work?

Post by Thesaints »

NiceUnparticularMan wrote: Sat Jul 17, 2021 3:58 am
Thesaints wrote: Fri Jul 16, 2021 9:45 am Or do you think lump-sum investors have a way to avoid market peaks ?
No, of course not. The only point I was making is once you are done buying, whether the price you bought at was one price or the average of several prices does not matter.
Since your return is calculated as "present price/acquisition price" the above statement doesn't look quite right.
I then don't understand how you are purporting to calculate long-term volatility of returns.

Let's do another simplified example.

We are looking at volatility of returns at time T.

Investors A and B both get $100K in cash 20 years before T. Investor A invests in a risky portfolio immediately. Investor B keeps it in cash for 19 years and 364 days, then invests it in the risky portfolio the day before T.

Do you not think Investor B will have a lower volatility of returns at T than A?
It depends on which volatility you have in mind. If it is the volatility calculated starting 20 years ago, surely B's volatility is much lower (in fact, it is zero) than A's.
But if you have in mind the volatility going forward, i.e. for t > T, then the two investors return has the same volatility.
OK, now Investor C also gets $100K in cash 20 years before T. Investor C does a DCA between the beginning and end of the first year.

Do you think Investor C will have a lower volatility of returns at T than B?
It depends on which volatlity you have in mind, right ? :)
If it is the volatility counted starting 20 years ago, C's volatility is of course higher than B's, but it is lower than A's.
If it is instead the volatility going forward, then C's volatility is indeed lower than B's.

What is true is Investor C will have a lower volatility of returns at T than A. That is because Investor C is a mild variant on Investor B--Investor C held some of their investment in cash for some period of time, and that reduces the volatility of returns at any later point in time. But since they only held cash for a limited period of time, it will not do as much to reduce the volatility of returns as holding cash for longer periods of time.
I think your confusion is in not realizing that you are assessing volatility for different asset allocations.
Once you focus on investors with the same asset allocation, then you should understand immediately that how they got into that asset allocation does count in terms of variability of future returns.
Well, in the example above, we met this condition! Investor B also got to the same AA "eventually"! Sure, just the day before, but they still did it.
Excellent! Now we can discuss the volatility of their returns going forward, once they have the same AA.
And we can do all sorts of variants in between. Another investor holds cash for 10 years then buys. That investor will have a higher volatility of returns than an investor who waits until the day before, but a lower volatility of returns than an investor who DCA over the first year.

Again, all we are showing here is holding cash for a while reduces the volatility of returns.
And by now you know the answer: "It depends on which volatility you have in mind"
Do you think losses incurred previously can be neglected ? When you say "Investor A can sell his stocks and DCA back in" means that when he sells his stocks he will realize a loss (or a gain) depending on his entry and his exit point. That too counts; in finance past is not fiorgotten.
You are correct that going forward they will have the same volatility of returns.
Neglected for what purpose? Counts for what purpose?
Counts for assessing the volatility of one's return by holding a certain portfolio. What else ?

And so if holding aside taxes, there are financial reasons to prefer DCA over that period, the same financial reasons should apply to converting to cash and then reinvesting over that period.
When you sell to buy back and realize this "pseudo-DCA" you have in mind, you incur a loss (or a gain). That has to be carried forward in order to assess the investment return. One can't simply forget about it.
Thesaints
Posts: 5108
Joined: Tue Jun 20, 2017 12:25 am

Re: When Does Dollar Cost Averaging Work?

Post by Thesaints »

NiceUnparticularMan wrote: Sat Jul 17, 2021 4:06 am
blackball wrote: Fri Jul 16, 2021 10:03 am Does dollar cost averaging work when selling? Is the answer the opposite of dca for buying?
So the basic reason DCA doesn't "work" when buying is assets with a positive expected return are more likely to increase in price over time. So, holding onto cash and waiting to buy tends to just mean buying at a higher price.
The part in bold is the key ! Future financial returns are not known with certainty. You may read that a 60/40 portfolio has an expected return of 6.5% annualized, but that only means that it "tends" to have that return and there are in fact many other different returns which are possible. At the end of the day you will experience a single one of those possible returns, not their average, not the expected return. DCA allows you to restrict the span of those possible return, in particular eliminating the very worst and the very best, at the price of lowering somewhat the expected value.
What happens when selling depends on what you mean by doing the opposite. But, suppose that means something like you would ordinarily plan to sell at T, but instead you will start gradually selling some period of time before T.
No. It is not possible to do the opposite of DCA. What I mean by "it is the opposite" is that while when buying the asset volatility works in your favor when DCA-ing, that same volatility works against you when selling for equal $ amounts.

But if you want to sacrifice expected returns for lower volatility, one option is to allocate some of those funds to cash instead. DCA, or the opposite in selling form, are particular ways of doing that. But you could do a lot of other things to increase your cash holdings.
If you reduce your stock allocation, then you are bound to have lower expected returns in perpetuity. If you DCA into the asset allocation instead, your expected return is lower, but as time goes by it tends to asymptotically to the same return a LS investor with the same AA enjoys
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: When Does Dollar Cost Averaging Work?

Post by nigel_ht »

NiceUnparticularMan wrote: Thu Jul 15, 2021 5:29 am
What is your answer to the hypothetical? Investor A and Investor B live next door to each other. Investor A hears Investor B is going to DCA his $100K into his existing $900K Target fund over the next 2-3 years.

If that is a good idea for Investor B, shouldn't Investor A take $100,000 out of his $1M Target fund and do the same thing?

Yes or no?
Never surrender the premise…because they are designed to drive to a desired logical conclusion.

Example:

Investor A is 35 and has a $3M portfolio allocated 90/10. He’s been thinking about FIRE but valuations are high even though he okay at a conservative 3% SWR.

investor B is 35, has a $2M portfolio allocated 90/10 and receives a $1M windfall. She decides this is enough to retire on and elects to DCA at a rate of 0.4%/month (96 months) into VT while keeping the rest in a combination of cash, short and med bonds during the DCA period and spending first from this pool.

Good so far? The numbers are a little different from the above but still pretty much the same scenario right?

Investor A hears what Investor B is doing and likes what she’s doing. Should he take out $1M, retire and DCA it back in the same way?

The answer is YES if there isn’t a huge tax liability and he’s going to retire right now as well.

But why go to around 60/40 and ramp back up to around 90/10?

Because glide paths mitigate SORR to some degree. Compare failure rates of 3.5% SWR when CAPE > 20:

Image

By using a upward glide path of 60% equities going up 100 Investor B can reduce the risk of failure or increase the SWR or some degree of both because 60/40 has a SWR of only 3.03% when cape is > 20 for 100% success over 60 years. She can do 3.42% for 100% and up to 3.75 for 95%.

That’s a good enough improvement over 60/40 for Investor A to pull the trigger on early retirement rather than keep working. That doing so has a penalty when things are going really well is immaterial…it just means whatever charity he likes get less money when he passes.

So the logical conclusion is that YES!, he should pull $1M out of his equities into a combination of cash, bills and bonds and DCA back in over 8 years just like Investor B.

Never surrender the premise of a hypothetical as it intentionally drives you to the logical conclusion the person proposing the premise wants.
Thesaints
Posts: 5108
Joined: Tue Jun 20, 2017 12:25 am

Re: When Does Dollar Cost Averaging Work?

Post by Thesaints »

It seems to me that selling at a single price is as risky as buying at a single price, though...
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: When Does Dollar Cost Averaging Work?

Post by nigel_ht »

Thesaints wrote: Sat Jul 17, 2021 5:07 pm It seems to me that selling at a single price is as risky as buying at a single price, though...
If the assumption is that the market is rising 2/3rds of the time then generally selling as late as possible gives you the best outcome. The 1/3 risk doesn't change much...
Thesaints
Posts: 5108
Joined: Tue Jun 20, 2017 12:25 am

Re: When Does Dollar Cost Averaging Work?

Post by Thesaints »

nigel_ht wrote: Sat Jul 17, 2021 5:18 pm
Thesaints wrote: Sat Jul 17, 2021 5:07 pm It seems to me that selling at a single price is as risky as buying at a single price, though...
If the assumption is that the market is rising 2/3rds of the time then generally selling as late as possible gives you the best outcome. The 1/3 risk doesn't change much...
But the objective of DCA is precisely avoiding those times when the market is not “about average”. If we only cared about average values, everybody should be 100% in stocks all the time, right ?
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: When Does Dollar Cost Averaging Work?

Post by nigel_ht »

Thesaints wrote: Sat Jul 17, 2021 5:28 pm
nigel_ht wrote: Sat Jul 17, 2021 5:18 pm
Thesaints wrote: Sat Jul 17, 2021 5:07 pm It seems to me that selling at a single price is as risky as buying at a single price, though...
If the assumption is that the market is rising 2/3rds of the time then generally selling as late as possible gives you the best outcome. The 1/3 risk doesn't change much...
But the objective of DCA is precisely avoiding those times when the market is not “about average”. If we only cared about average values, everybody should be 100% in stocks all the time, right ?
My opinion is that DCA is an AA glide path.

The primary time we talk about glide paths on BH is near retirement when we fear "not average" biting us in the rear. The rest of the time we figure that the average will eventually drown out any interim bad luck.
FireProof
Posts: 960
Joined: Thu May 05, 2011 12:15 pm

Re: When Does Dollar Cost Averaging Work?

Post by FireProof »

nzahir wrote: Wed Jul 07, 2021 1:30 am Hey guys,

So most people here would recommend to lump sum and Vanguards study said that 2/3 of the time Lump Sum investing would outperform Dollar Cost Averaging.

But when is that 1/3 of the time where DCA would work out better?

Is that usually when market valuations are high? The study I believe was a bit vague and not very specific iirc.

And how long is the DCA timeframe? 12 months?

Thank you in advance!
It reduced path dependency by dividing your investment into multiple equally weighted trials. Therefore, you don't run the risk of putting all your money in at the worst possible time. Given the tendency of the market to go up over time, your average expected return will be worse, but it can still be valuable for reducing maximum downside, which is especially valuable WHEN YOU ARE SIGNIFICANTLY LEVERAGED, and a reducing maximum downside could be the difference between being wiped out or not.
Thesaints
Posts: 5108
Joined: Tue Jun 20, 2017 12:25 am

Re: When Does Dollar Cost Averaging Work?

Post by Thesaints »

nigel_ht wrote: Sat Jul 17, 2021 5:34 pm
Thesaints wrote: Sat Jul 17, 2021 5:28 pm
nigel_ht wrote: Sat Jul 17, 2021 5:18 pm
Thesaints wrote: Sat Jul 17, 2021 5:07 pm It seems to me that selling at a single price is as risky as buying at a single price, though...
If the assumption is that the market is rising 2/3rds of the time then generally selling as late as possible gives you the best outcome. The 1/3 risk doesn't change much...
But the objective of DCA is precisely avoiding those times when the market is not “about average”. If we only cared about average values, everybody should be 100% in stocks all the time, right ?
My opinion is that DCA is an AA glide path.

The primary time we talk about glide paths on BH is near retirement when we fear "not average" biting us in the rear. The rest of the time we figure that the average will eventually drown out any interim bad luck.
That’s certainly a way to look at it, keeping in mind that it is a very short path. I’d never advocate a 10 year long DCA.
HootingSloth
Posts: 1050
Joined: Mon Jan 28, 2019 2:38 pm

Re: When Does Dollar Cost Averaging Work?

Post by HootingSloth »

nigel_ht wrote: Sat Jul 17, 2021 3:13 pm
NiceUnparticularMan wrote: Thu Jul 15, 2021 5:29 am
What is your answer to the hypothetical? Investor A and Investor B live next door to each other. Investor A hears Investor B is going to DCA his $100K into his existing $900K Target fund over the next 2-3 years.

If that is a good idea for Investor B, shouldn't Investor A take $100,000 out of his $1M Target fund and do the same thing?

Yes or no?
Never surrender the premise…because they are designed to drive to a desired logical conclusion.

Example:

Investor A is 35 and has a $3M portfolio allocated 90/10. He’s been thinking about FIRE but valuations are high even though he okay at a conservative 3% SWR.

investor B is 35, has a $2M portfolio allocated 90/10 and receives a $1M windfall. She decides this is enough to retire on and elects to DCA at a rate of 0.4%/month (96 months) into VT while keeping the rest in a combination of cash, short and med bonds during the DCA period and spending first from this pool.

Good so far? The numbers are a little different from the above but still pretty much the same scenario right?

Investor A hears what Investor B is doing and likes what she’s doing. Should he take out $1M, retire and DCA it back in the same way?

The answer is YES if there isn’t a huge tax liability and he’s going to retire right now as well.

But why go to around 60/40 and ramp back up to around 90/10?

Because glide paths mitigate SORR to some degree. Compare failure rates of 3.5% SWR when CAPE > 20:

Image

By using a upward glide path of 60% equities going up 100 Investor B can reduce the risk of failure or increase the SWR or some degree of both because 60/40 has a SWR of only 3.03% when cape is > 20 for 100% success over 60 years. She can do 3.42% for 100% and up to 3.75 for 95%.

That’s a good enough improvement over 60/40 for Investor A to pull the trigger on early retirement rather than keep working. That doing so has a penalty when things are going really well is immaterial…it just means whatever charity he likes get less money when he passes.

So the logical conclusion is that YES!, he should pull $1M out of his equities into a combination of cash, bills and bonds and DCA back in over 8 years just like Investor B.

Never surrender the premise of a hypothetical as it intentionally drives you to the logical conclusion the person proposing the premise wants.
Nigel,

I believe that chart is from Big ERN. Isn't it illustrating a chunk put into bonds (rather than cash) and then transitioned into stocks quite gradually? For example, going from 60/40 to 100/0 at a rate of 0.3% per month (one of the more successful strategies depicted in the chart), takes over 11 years. I don't think most people would call that a DCA strategy. They would call it a bond tent or a rising equity glidepath instead. From what I have seen, DCA is used to refer to a much sharper (maybe over the course of a year) transition from cash (rather than bonds) to stocks, and not necessarily at the same time as retirement.
Global Market Portfolio + modest tilt towards volatility (80/20->60/40 as approach FI) + modest tilt away from exchange rate risk (80% global+20% U.S. stocks; currency-hedge bonds) + tax optimization
Pepper11
Posts: 195
Joined: Fri Jun 13, 2014 9:22 pm

Re: When Does Dollar Cost Averaging Work?

Post by Pepper11 »

The argument of DCA vs Lump Sum is very simple, but that doesn't mean its clear.

Lump sum will come out ahead the majority of the time.

Lump sum also has a higher risk of large losses.

If the possibility of large losses would be catastrophic to your financial situation, you should consider DCA.

Not sure why there is so much dogma for lump sum on this forum.
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: When Does Dollar Cost Averaging Work?

Post by nigel_ht »

HootingSloth wrote: Sat Jul 17, 2021 9:15 pm
nigel_ht wrote: Sat Jul 17, 2021 3:13 pm
NiceUnparticularMan wrote: Thu Jul 15, 2021 5:29 am
What is your answer to the hypothetical? Investor A and Investor B live next door to each other. Investor A hears Investor B is going to DCA his $100K into his existing $900K Target fund over the next 2-3 years.

If that is a good idea for Investor B, shouldn't Investor A take $100,000 out of his $1M Target fund and do the same thing?

Yes or no?
Never surrender the premise…because they are designed to drive to a desired logical conclusion.

Example:

Investor A is 35 and has a $3M portfolio allocated 90/10. He’s been thinking about FIRE but valuations are high even though he okay at a conservative 3% SWR.

investor B is 35, has a $2M portfolio allocated 90/10 and receives a $1M windfall. She decides this is enough to retire on and elects to DCA at a rate of 0.4%/month (96 months) into VT while keeping the rest in a combination of cash, short and med bonds during the DCA period and spending first from this pool.

Good so far? The numbers are a little different from the above but still pretty much the same scenario right?

Investor A hears what Investor B is doing and likes what she’s doing. Should he take out $1M, retire and DCA it back in the same way?

The answer is YES if there isn’t a huge tax liability and he’s going to retire right now as well.

But why go to around 60/40 and ramp back up to around 90/10?

Because glide paths mitigate SORR to some degree. Compare failure rates of 3.5% SWR when CAPE > 20:

Image

By using a upward glide path of 60% equities going up 100 Investor B can reduce the risk of failure or increase the SWR or some degree of both because 60/40 has a SWR of only 3.03% when cape is > 20 for 100% success over 60 years. She can do 3.42% for 100% and up to 3.75 for 95%.

That’s a good enough improvement over 60/40 for Investor A to pull the trigger on early retirement rather than keep working. That doing so has a penalty when things are going really well is immaterial…it just means whatever charity he likes get less money when he passes.

So the logical conclusion is that YES!, he should pull $1M out of his equities into a combination of cash, bills and bonds and DCA back in over 8 years just like Investor B.

Never surrender the premise of a hypothetical as it intentionally drives you to the logical conclusion the person proposing the premise wants.
Nigel,

I believe that chart is from Big ERN. Isn't it illustrating a chunk put into bonds (rather than cash) and then transitioned into stocks quite gradually? For example, going from 60/40 to 100/0 at a rate of 0.3% per month (one of the more successful strategies depicted in the chart), takes over 11 years. I don't think most people would call that a DCA strategy. They would call it a bond tent or a rising equity glidepath instead. From what I have seen, DCA is used to refer to a much sharper (maybe over the course of a year) transition from cash (rather than bonds) to stocks, and not necessarily at the same time as retirement.
Cash performs less well but if you stuff your cash into bills at least it more or less keeps up with inflation...

Yes, it's from big ERN...hopefully the chart itself is labeled as such...they usually are so I didn't bother doing it since I wrote that post on my phone...
User avatar
dogagility
Posts: 3237
Joined: Fri Feb 24, 2017 5:41 am

Re: When Does Dollar Cost Averaging Work?

Post by dogagility »

Pepper11 wrote: Sat Jul 17, 2021 9:33 pm The argument of DCA vs Lump Sum is very simple, but that doesn't mean its clear.
It is simple from a financial probability standpoint.
Pepper11 wrote: Sat Jul 17, 2021 9:33 pm Lump sum will come out ahead the majority of the time.
Agree. The market trajectory over time is up. We all expect this, or we wouldn't be investing in it.
Pepper11 wrote: Sat Jul 17, 2021 9:33 pm Lump sum also has a higher risk of large losses.
Agree. But then if a person is worried about large losses, they should choose an asset allocation that has less exposure to equities.
Pepper11 wrote: Sat Jul 17, 2021 9:33 pm If the possibility of large losses would be catastrophic to your financial situation, you should consider DCA.
I disagree; the person should choose a different asset allocation.
Pepper11 wrote: Sat Jul 17, 2021 9:33 pm Not sure why there is so much dogma for lump sum on this forum.
Lump sum is suggested because it has the highest probability of maximizing portfolio size.
Make sure you check out my list of certifications. The list is short, and there aren't any. - Eric 0. from SMA
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

Thesaints wrote: Sat Jul 17, 2021 2:16 pm DCA allows you to restrict the span of those possible return, in particular eliminating the very worst and the very best, at the price of lowering somewhat the expected value.
Again, so does any scheme where you hold more cash for some period. For example, so does every three years, converting your portfolio to cash, then gradually reinvesting over those three years, then doing it all again over and over.

I'm not denying holding more cash reduces both volatility and expected returns. I'm just pointing out this particular way of holding more cash does nothing more than that. And it is available to everyone (at least with funds in tax-deferred accounts), not just people who happened to recently receive new funds.
But if you want to sacrifice expected returns for lower volatility, one option is to allocate some of those funds to cash instead. DCA, or the opposite in selling form, are particular ways of doing that. But you could do a lot of other things to increase your cash holdings.
If you reduce your stock allocation, then you are bound to have lower expected returns in perpetuity. If you DCA into the asset allocation instead, your expected return is lower, but as time goes by it tends to asymptotically to the same return a LS investor with the same AA enjoys
I'm not quite sure what you are saying here, but it is true that if you hold 50% of your portfolio in cash for say a year, then drop your cash to 0%, then the impact of that temporary cash holding measured in certain relative terms goes down the more years you imagine occurring after you did this.

So, hold 50% cash for a year, and there is only one day after with 0% cash, that can have a very large impact by certain relative measures. Hold 50% cash for a year, and there is 30 years after with 0% cash, and that can have less impact by those relative measures.

But again, there is nothing special about DCA in particular versus any other temporary scheme for holding cash. And again holding aside tax, it doesn't matter what path came before--whether you got those funds through, say, 10 prior years of saving, or an inheritance from Granny, the future financial consequences are the same if you then decide to go with a temporary allocation to cash.
Last edited by NiceUnparticularMan on Sun Jul 18, 2021 9:09 am, edited 1 time in total.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

nigel_ht wrote: Sat Jul 17, 2021 3:13 pm
NiceUnparticularMan wrote: Thu Jul 15, 2021 5:29 am
What is your answer to the hypothetical? Investor A and Investor B live next door to each other. Investor A hears Investor B is going to DCA his $100K into his existing $900K Target fund over the next 2-3 years.

If that is a good idea for Investor B, shouldn't Investor A take $100,000 out of his $1M Target fund and do the same thing?

Yes or no?
Never surrender the premise…because they are designed to drive to a desired logical conclusion.

Example:

Investor A is 35 and has a $3M portfolio allocated 90/10. He’s been thinking about FIRE but valuations are high even though he okay at a conservative 3% SWR.

investor B is 35, has a $2M portfolio allocated 90/10 and receives a $1M windfall. She decides this is enough to retire on and elects to DCA at a rate of 0.4%/month (96 months) into VT while keeping the rest in a combination of cash, short and med bonds during the DCA period and spending first from this pool.

Good so far? The numbers are a little different from the above but still pretty much the same scenario right?

Investor A hears what Investor B is doing and likes what she’s doing. Should he take out $1M, retire and DCA it back in the same way?

The answer is YES if there isn’t a huge tax liability and he’s going to retire right now as well.

But why go to around 60/40 and ramp back up to around 90/10?

Because glide paths mitigate SORR to some degree. Compare failure rates of 3.5% SWR when CAPE > 20:

Image

By using a upward glide path of 60% equities going up 100 Investor B can reduce the risk of failure or increase the SWR or some degree of both because 60/40 has a SWR of only 3.03% when cape is > 20 for 100% success over 60 years. She can do 3.42% for 100% and up to 3.75 for 95%.

That’s a good enough improvement over 60/40 for Investor A to pull the trigger on early retirement rather than keep working. That doing so has a penalty when things are going really well is immaterial…it just means whatever charity he likes get less money when he passes.

So the logical conclusion is that YES!, he should pull $1M out of his equities into a combination of cash, bills and bonds and DCA back in over 8 years just like Investor B.

Never surrender the premise of a hypothetical as it intentionally drives you to the logical conclusion the person proposing the premise wants.
Right, this is the same logic as behind the "bond tent" theory.

And so this isn't really about what to do with a sudden increase in available funds. It is just a question of whether or not a bond tent is a good idea in general, however you got the funds you would be using for your bond tent.

And then the second question is assuming you believed a "bond tent" was a good idea, how much, starting when, ending when, and in what exact assets? Again, the optimal answers to those questions don't really depend on how you get the relevant funds.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

Pepper11 wrote: Sat Jul 17, 2021 9:33 pm The argument of DCA vs Lump Sum is very simple, but that doesn't mean its clear.

Lump sum will come out ahead the majority of the time.

Lump sum also has a higher risk of large losses.

If the possibility of large losses would be catastrophic to your financial situation, you should consider DCA.

Not sure why there is so much dogma for lump sum on this forum.
The answer to your last sentence is:

Because there is an obvious alternative to dealing with the problem assumed in the prior sentence, the one that starts with "If the possibility of large losses would be catastrophic to your financial situation . . . ." An alternative which almost everyone here is very familiar with.

If you are in that position, you should hold a higher percentage of low-risk assets.

If you are in that position but it would be temporary, maybe you only do that for some period of time--e.g., a bond tent strategy.

And it is true DCA is a "cash tent" strategy, but it is a very specific and frankly pretty odd one. So it is quite unlikely it would happen to turn out that a DCA scheme specifically was the right sort of strategy to deal with the problem that was the premise of that penultimate question.
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: When Does Dollar Cost Averaging Work?

Post by nigel_ht »

NiceUnparticularMan wrote: Sun Jul 18, 2021 9:00 am
nigel_ht wrote: Sat Jul 17, 2021 3:13 pm
NiceUnparticularMan wrote: Thu Jul 15, 2021 5:29 am
What is your answer to the hypothetical? Investor A and Investor B live next door to each other. Investor A hears Investor B is going to DCA his $100K into his existing $900K Target fund over the next 2-3 years.

If that is a good idea for Investor B, shouldn't Investor A take $100,000 out of his $1M Target fund and do the same thing?

Yes or no?
Never surrender the premise…because they are designed to drive to a desired logical conclusion.

Example:

Investor A is 35 and has a $3M portfolio allocated 90/10. He’s been thinking about FIRE but valuations are high even though he okay at a conservative 3% SWR.

investor B is 35, has a $2M portfolio allocated 90/10 and receives a $1M windfall. She decides this is enough to retire on and elects to DCA at a rate of 0.4%/month (96 months) into VT while keeping the rest in a combination of cash, short and med bonds during the DCA period and spending first from this pool.

Good so far? The numbers are a little different from the above but still pretty much the same scenario right?

Investor A hears what Investor B is doing and likes what she’s doing. Should he take out $1M, retire and DCA it back in the same way?

The answer is YES if there isn’t a huge tax liability and he’s going to retire right now as well.

But why go to around 60/40 and ramp back up to around 90/10?

Because glide paths mitigate SORR to some degree. Compare failure rates of 3.5% SWR when CAPE > 20:

Image

By using a upward glide path of 60% equities going up 100 Investor B can reduce the risk of failure or increase the SWR or some degree of both because 60/40 has a SWR of only 3.03% when cape is > 20 for 100% success over 60 years. She can do 3.42% for 100% and up to 3.75 for 95%.

That’s a good enough improvement over 60/40 for Investor A to pull the trigger on early retirement rather than keep working. That doing so has a penalty when things are going really well is immaterial…it just means whatever charity he likes get less money when he passes.

So the logical conclusion is that YES!, he should pull $1M out of his equities into a combination of cash, bills and bonds and DCA back in over 8 years just like Investor B.

Never surrender the premise of a hypothetical as it intentionally drives you to the logical conclusion the person proposing the premise wants.
Right, this is the same logic as behind the "bond tent" theory.

And so this isn't really about what to do with a sudden increase in available funds. It is just a question of whether or not a bond tent is a good idea in general, however you got the funds you would be using for your bond tent.

And then the second question is assuming you believed a "bond tent" was a good idea, how much, starting when, ending when, and in what exact assets? Again, the optimal answers to those questions don't really depend on how you get the relevant funds.
Yes, I agree but so what?

The leading question was whether investor A should ever do what Investor B is doing with a DCA.

In this scenario the answer is yes…which is the opposite of the “desired” answer.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

HootingSloth wrote: Sat Jul 17, 2021 9:15 pm Nigel,

I believe that chart is from Big ERN. Isn't it illustrating a chunk put into bonds (rather than cash) and then transitioned into stocks quite gradually? For example, going from 60/40 to 100/0 at a rate of 0.3% per month (one of the more successful strategies depicted in the chart), takes over 11 years. I don't think most people would call that a DCA strategy. They would call it a bond tent or a rising equity glidepath instead. From what I have seen, DCA is used to refer to a much sharper (maybe over the course of a year) transition from cash (rather than bonds) to stocks, and not necessarily at the same time as retirement.
Indeed.

I note the debate over the bond tent strategy, as I understand it, basically boils down to the observation that it is generally going to be much more powerful to show some plausible flexibility in terms of things like the exact timing of retirement, the exact size of withdrawals, and so on. So, the case for the bond tent being a good idea in practice really depends on assuming away most such flexibility.

There is also a bit of an issue for the bond tent strategy in that SPIAs, particularly under the same assumptions of inflexibility, might actually work a bit better.
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

nigel_ht wrote: Sun Jul 18, 2021 9:13 am
NiceUnparticularMan wrote: Sun Jul 18, 2021 9:00 am
Right, this is the same logic as behind the "bond tent" theory.

And so this isn't really about what to do with a sudden increase in available funds. It is just a question of whether or not a bond tent is a good idea in general, however you got the funds you would be using for your bond tent.

And then the second question is assuming you believed a "bond tent" was a good idea, how much, starting when, ending when, and in what exact assets? Again, the optimal answers to those questions don't really depend on how you get the relevant funds.
Yes, I agree but so what?

The leading question was whether investor A should ever do what Investor B is doing with a DCA.

In this scenario the answer is yes…which is the opposite of the “desired” answer.
I'm not sure I agree with your framing of the leading question, or questions, nor what answers are "desired".

And actually, I think it is quite important that we seem to have reached the conclusion that it is not really DCA of new funds that matters, it is a well-timed bond tent that (might) matter, and therefore one really should not think of this issue as being about what to do with new funds whenever they happen to show up. And even if new funds did happen to show up right when you should be beginning a bond tent, that bond tent wouldn't necessarily look like DCA anyway.

I might even--gasp--say you should "lump sum" those funds into your bond tent . . . .
HootingSloth
Posts: 1050
Joined: Mon Jan 28, 2019 2:38 pm

Re: When Does Dollar Cost Averaging Work?

Post by HootingSloth »

NiceUnparticularMan wrote: Sun Jul 18, 2021 9:15 am
HootingSloth wrote: Sat Jul 17, 2021 9:15 pm Nigel,

I believe that chart is from Big ERN. Isn't it illustrating a chunk put into bonds (rather than cash) and then transitioned into stocks quite gradually? For example, going from 60/40 to 100/0 at a rate of 0.3% per month (one of the more successful strategies depicted in the chart), takes over 11 years. I don't think most people would call that a DCA strategy. They would call it a bond tent or a rising equity glidepath instead. From what I have seen, DCA is used to refer to a much sharper (maybe over the course of a year) transition from cash (rather than bonds) to stocks, and not necessarily at the same time as retirement.
Indeed.

I note the debate over the bond tent strategy, as I understand it, basically boils down to the observation that it is generally going to be much more powerful to show some plausible flexibility in terms of things like the exact timing of retirement, the exact size of withdrawals, and so on. So, the case for the bond tent being a good idea in practice really depends on assuming away most such flexibility.

There is also a bit of an issue for the bond tent strategy in that SPIAs, particularly under the same assumptions of inflexibility, might actually work a bit better.
How much flexibility can save you probably depends on your assumptions about the plausible range of what the stock market might do, no? If you think future stock market performance might match that of an early 20th-century Germany, then it seems like the bond tent (or, probably better, some kind of LMP for baseline expenditures) would probably win out over being flexible.
Global Market Portfolio + modest tilt towards volatility (80/20->60/40 as approach FI) + modest tilt away from exchange rate risk (80% global+20% U.S. stocks; currency-hedge bonds) + tax optimization
HootingSloth
Posts: 1050
Joined: Mon Jan 28, 2019 2:38 pm

Re: When Does Dollar Cost Averaging Work?

Post by HootingSloth »

NiceUnparticularMan wrote: Sun Jul 18, 2021 9:19 am I might even--gasp--say you should "lump sum" those funds into your bond tent . . . .
Yes, it seems to me that you should definitely lump sump your funds into your bond tent. This was the conclusion of my earlier posts looking at someone going from $2M invested 80/20 to $3M invested 60/40 following a $1M windfall. When you are doing this, DCAing actually makes you (a bit) worse off in the very worst sequences because stocks are plummeting and you are "catching the falling knife" with your DCA.
Global Market Portfolio + modest tilt towards volatility (80/20->60/40 as approach FI) + modest tilt away from exchange rate risk (80% global+20% U.S. stocks; currency-hedge bonds) + tax optimization
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

HootingSloth wrote: Sun Jul 18, 2021 11:24 am
NiceUnparticularMan wrote: Sun Jul 18, 2021 9:15 am
HootingSloth wrote: Sat Jul 17, 2021 9:15 pm Nigel,

I believe that chart is from Big ERN. Isn't it illustrating a chunk put into bonds (rather than cash) and then transitioned into stocks quite gradually? For example, going from 60/40 to 100/0 at a rate of 0.3% per month (one of the more successful strategies depicted in the chart), takes over 11 years. I don't think most people would call that a DCA strategy. They would call it a bond tent or a rising equity glidepath instead. From what I have seen, DCA is used to refer to a much sharper (maybe over the course of a year) transition from cash (rather than bonds) to stocks, and not necessarily at the same time as retirement.
Indeed.

I note the debate over the bond tent strategy, as I understand it, basically boils down to the observation that it is generally going to be much more powerful to show some plausible flexibility in terms of things like the exact timing of retirement, the exact size of withdrawals, and so on. So, the case for the bond tent being a good idea in practice really depends on assuming away most such flexibility.

There is also a bit of an issue for the bond tent strategy in that SPIAs, particularly under the same assumptions of inflexibility, might actually work a bit better.
How much flexibility can save you probably depends on your assumptions about the plausible range of what the stock market might do, no? If you think future stock market performance might match that of an early 20th-century Germany, then it seems like the bond tent (or, probably better, some kind of LMP for baseline expenditures) would probably win out over being flexible.
My first thought for such black swan scenarios is it would be good to have a lot in non-US assets generally, possibly including non-US bonds.

My second thought is you might NEED to be very flexible in your withdrawals if some major portion of your portfolio is going to zero, or near enough.

But yes, I think it is true such scenarios exceed the limits of what a "normal" flexible withdrawal plan is designed to handle.
HootingSloth
Posts: 1050
Joined: Mon Jan 28, 2019 2:38 pm

Re: When Does Dollar Cost Averaging Work?

Post by HootingSloth »

NiceUnparticularMan wrote: Sun Jul 18, 2021 11:33 am
HootingSloth wrote: Sun Jul 18, 2021 11:24 am
NiceUnparticularMan wrote: Sun Jul 18, 2021 9:15 am
HootingSloth wrote: Sat Jul 17, 2021 9:15 pm Nigel,

I believe that chart is from Big ERN. Isn't it illustrating a chunk put into bonds (rather than cash) and then transitioned into stocks quite gradually? For example, going from 60/40 to 100/0 at a rate of 0.3% per month (one of the more successful strategies depicted in the chart), takes over 11 years. I don't think most people would call that a DCA strategy. They would call it a bond tent or a rising equity glidepath instead. From what I have seen, DCA is used to refer to a much sharper (maybe over the course of a year) transition from cash (rather than bonds) to stocks, and not necessarily at the same time as retirement.
Indeed.

I note the debate over the bond tent strategy, as I understand it, basically boils down to the observation that it is generally going to be much more powerful to show some plausible flexibility in terms of things like the exact timing of retirement, the exact size of withdrawals, and so on. So, the case for the bond tent being a good idea in practice really depends on assuming away most such flexibility.

There is also a bit of an issue for the bond tent strategy in that SPIAs, particularly under the same assumptions of inflexibility, might actually work a bit better.
How much flexibility can save you probably depends on your assumptions about the plausible range of what the stock market might do, no? If you think future stock market performance might match that of an early 20th-century Germany, then it seems like the bond tent (or, probably better, some kind of LMP for baseline expenditures) would probably win out over being flexible.
My first thought for such black swan scenarios is it would be good to have a lot in non-US assets generally, possibly including non-US bonds.

My second thought is you might NEED to be very flexible in your withdrawals if some major portion of your portfolio is going to zero, or near enough.

But yes, I think it is true such scenarios exceed the limits of what a "normal" flexible withdrawal plan is designed to handle.
I agree on all of those points. I will note that, even if they had invested in a global portfolio of stocks, that German investor would have had a 57-year long drawdown in stocks. I think it is really a combination of diversification, flexibility, and an LMP (or something similar) that would help most for black swans. Of course, that can get expensive to implement, but can be worth it for some.
Global Market Portfolio + modest tilt towards volatility (80/20->60/40 as approach FI) + modest tilt away from exchange rate risk (80% global+20% U.S. stocks; currency-hedge bonds) + tax optimization
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

HootingSloth wrote: Sun Jul 18, 2021 11:38 am
NiceUnparticularMan wrote: Sun Jul 18, 2021 11:33 am
HootingSloth wrote: Sun Jul 18, 2021 11:24 am
NiceUnparticularMan wrote: Sun Jul 18, 2021 9:15 am
HootingSloth wrote: Sat Jul 17, 2021 9:15 pm Nigel,

I believe that chart is from Big ERN. Isn't it illustrating a chunk put into bonds (rather than cash) and then transitioned into stocks quite gradually? For example, going from 60/40 to 100/0 at a rate of 0.3% per month (one of the more successful strategies depicted in the chart), takes over 11 years. I don't think most people would call that a DCA strategy. They would call it a bond tent or a rising equity glidepath instead. From what I have seen, DCA is used to refer to a much sharper (maybe over the course of a year) transition from cash (rather than bonds) to stocks, and not necessarily at the same time as retirement.
Indeed.

I note the debate over the bond tent strategy, as I understand it, basically boils down to the observation that it is generally going to be much more powerful to show some plausible flexibility in terms of things like the exact timing of retirement, the exact size of withdrawals, and so on. So, the case for the bond tent being a good idea in practice really depends on assuming away most such flexibility.

There is also a bit of an issue for the bond tent strategy in that SPIAs, particularly under the same assumptions of inflexibility, might actually work a bit better.
How much flexibility can save you probably depends on your assumptions about the plausible range of what the stock market might do, no? If you think future stock market performance might match that of an early 20th-century Germany, then it seems like the bond tent (or, probably better, some kind of LMP for baseline expenditures) would probably win out over being flexible.
My first thought for such black swan scenarios is it would be good to have a lot in non-US assets generally, possibly including non-US bonds.

My second thought is you might NEED to be very flexible in your withdrawals if some major portion of your portfolio is going to zero, or near enough.

But yes, I think it is true such scenarios exceed the limits of what a "normal" flexible withdrawal plan is designed to handle.
I agree on all of those points. I will note that, even if they had invested in a global portfolio of stocks, that German investor would have had a 57-year long drawdown in stocks. I think it is really a combination of diversification, flexibility, and an LMP (or something similar) that would help most for black swans. Of course, that can get expensive to implement, but can be worth it for some.
Yeah, particularly at today's valuations, the most helpful thing is to be really wealthy. That way you can diversify and hedge in all sorts of ways, get lame returns if nothing bad happens, or a mix if something bad happens, and still be OK when the dust settles.
Thesaints
Posts: 5108
Joined: Tue Jun 20, 2017 12:25 am

Re: When Does Dollar Cost Averaging Work?

Post by Thesaints »

dogagility wrote: Sun Jul 18, 2021 6:03 am I disagree; the person should choose a different asset allocation.
Easy example:

Let's say stocks' expected return is 10% and bonds' is zero, to have round figures.
A 70/30 AA has a 7% expected return.
If I DCA into that AA over, let's say, 2 years, I give up one year worth of return from a 70% stock allocation, i.e. 7%.
Therefore, after N years, the cumulative expected return of the LS and DCA investors are

N=5
LS investor: 40%
DCA investor: 31%

N=10
LS investor: 97%
DCA investor: 84%

N=20
LS investor: 387%
DCA investor: 362%

But you are saying I could go to a, let's say 60/40 AA to lower my risk ?
It's expected return is 6%, let's see what happens after N years:

N=5
60/30 AA: 34%

N=10
60/30 AA: 79%

N=20
60/30 AA: 321%

As you can see, expected return falls more and more behind the 70/30 AA, even when DCA is taken into account.
Whereas the DCA return keeps up with the LS and , in fact,the two returns converge asymptotically.
Thesaints
Posts: 5108
Joined: Tue Jun 20, 2017 12:25 am

Re: When Does Dollar Cost Averaging Work?

Post by Thesaints »

NiceUnparticularMan wrote: Sun Jul 18, 2021 8:54 am
Thesaints wrote: Sat Jul 17, 2021 2:16 pm DCA allows you to restrict the span of those possible return, in particular eliminating the very worst and the very best, at the price of lowering somewhat the expected value.
Again, so does any scheme where you hold more cash for some period. For example, so does every three years, converting your portfolio to cash, then gradually reinvesting over those three years, then doing it all again over and over.
When you sell at a single price you are actually adding volatility. Furthermore, the strategy you suggest has only a little more than half the expected return of the DCA strategy (see my previous post).
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

Thesaints wrote: Mon Jul 19, 2021 2:55 pmWhen you sell at a single price you are actually adding volatility.
I am not sure what you are claiming here.

Investor A: Holds 100% stock at T0, holds it for 20 years.

Investor B: Holds 100% stock at T0. Converts to 100% cash. Converts back to 100% stock three years later. Holds it for the remaining 17 years.

Investor B will have a lower volatility of returns at the end than Investor A. Also, the more years you add on the end past this cash period, the less the difference in their returns (by some measures). Which is just because you are assuming relatively less and less time in cash proportionately the more and more time you add on afterward.
Furthermore, the strategy you suggest has only a little more than half the expected return of the DCA strategy (see my previous post).
Sure, because you are doing more of it. If you only hold a relatively short-lived "cash tent" once during a relatively long investment period, it only has a relatively small financial impact. If you hold that same "cash tent" many times during the investment period, it will have a relatively larger financial impact. Also if you hold a longer-lived "cash tent," and so on. All this is just illustrating the more time you spend invested in cash versus risky assets, the bigger the impact of such strategies.

So, we would have to ask questions like this: Should we hold a "cash tent" at all? Are their better alternatives? If a "cash tent" is indeed best, how many times should we hold a "cash tent"? When should we hold a "cash tent"? How big a "tent"? For how long? And so on.

Again holding aside possible tax issues, the optimal answers to these questions don't have anything in particular to do with when and how you received funds.
nigel_ht
Posts: 4742
Joined: Tue Jan 01, 2019 9:14 am

Re: When Does Dollar Cost Averaging Work?

Post by nigel_ht »

Lol…DCA would have worked last week vs lump sum…
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

nigel_ht wrote: Mon Jul 19, 2021 3:22 pm Lol…DCA would have worked last week vs lump sum…
Tactical asset allocation is definitely easier with the benefit of hindsight!

But as usual, armed with that knowledge, you could have done much better than just DCA.
User avatar
dogagility
Posts: 3237
Joined: Fri Feb 24, 2017 5:41 am

Re: When Does Dollar Cost Averaging Work?

Post by dogagility »

Thesaints wrote: Mon Jul 19, 2021 2:52 pm
dogagility wrote: Sun Jul 18, 2021 6:03 am I disagree; the person should choose a different asset allocation.
Easy example:

Let's say stocks' expected return is 10% and bonds' is zero, to have round figures.
A 70/30 AA has a 7% expected return.
If I DCA into that AA over, let's say, 2 years, I give up one year worth of return from a 70% stock allocation, i.e. 7%.
Therefore, after N years, the cumulative expected return of the LS and DCA investors are

N=5
LS investor: 40%
DCA investor: 31%

N=10
LS investor: 97%
DCA investor: 84%

N=20
LS investor: 387%
DCA investor: 362%

But you are saying I could go to a, let's say 60/40 AA to lower my risk ?
It's expected return is 6%, let's see what happens after N years:

N=5
60/30 AA: 34%

N=10
60/30 AA: 79%

N=20
60/30 AA: 321%

As you can see, expected return falls more and more behind the 70/30 AA, even when DCA is taken into account.
Whereas the DCA return keeps up with the LS and , in fact,the two returns converge asymptotically.
You took what I wrote out of context. Your reply above doesn't address the exchange between the other poster and myself.

Here's the entire exchange:
If the possibility of large losses would be catastrophic to your financial situation, you should consider DCA.
dogagility wrote: Sun Jul 18, 2021 6:03 am I disagree; the person should choose a different asset allocation.
The solution to avoiding "large losses" that "would be catastrophic" is not to DCA money over time to achieve a final asset allocation. The solution is to choose a different asset allocation that would not be "catastrophic" and then lump sum into that chosen asset allocation.
Make sure you check out my list of certifications. The list is short, and there aren't any. - Eric 0. from SMA
HootingSloth
Posts: 1050
Joined: Mon Jan 28, 2019 2:38 pm

Re: When Does Dollar Cost Averaging Work?

Post by HootingSloth »

Thesaints wrote: Mon Jul 19, 2021 2:52 pm
dogagility wrote: Sun Jul 18, 2021 6:03 am I disagree; the person should choose a different asset allocation.
Easy example:

Let's say stocks' expected return is 10% and bonds' is zero, to have round figures.
A 70/30 AA has a 7% expected return.
If I DCA into that AA over, let's say, 2 years, I give up one year worth of return from a 70% stock allocation, i.e. 7%.
Therefore, after N years, the cumulative expected return of the LS and DCA investors are

N=5
LS investor: 40%
DCA investor: 31%

N=10
LS investor: 97%
DCA investor: 84%

N=20
LS investor: 387%
DCA investor: 362%

But you are saying I could go to a, let's say 60/40 AA to lower my risk ?
It's expected return is 6%, let's see what happens after N years:

N=5
60/30 AA: 34%

N=10
60/30 AA: 79%

N=20
60/30 AA: 321%

As you can see, expected return falls more and more behind the 70/30 AA, even when DCA is taken into account.
Whereas the DCA return keeps up with the LS and , in fact,the two returns converge asymptotically.
Thesaints,

Presumably you think that the DCA investor is getting something valuable in exchange for giving up some of the expected return relative to the LS investor. As you explained it above, the benefit seems to be a narrower dispersion of returns. When the DCA investor gets to the end of the 2-year period, could they not decide to sell what they had just purchased over the last two years, and DCA it back in over the next 2-years, in order to get the same exchange--narrower dispersion in exchange for lower expected returns? If they thought this was a good deal the first time, why is it not a good deal the second time?
Global Market Portfolio + modest tilt towards volatility (80/20->60/40 as approach FI) + modest tilt away from exchange rate risk (80% global+20% U.S. stocks; currency-hedge bonds) + tax optimization
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

HootingSloth wrote: Mon Jul 19, 2021 5:06 pm
Thesaints wrote: Mon Jul 19, 2021 2:52 pm
dogagility wrote: Sun Jul 18, 2021 6:03 am I disagree; the person should choose a different asset allocation.
Easy example:

Let's say stocks' expected return is 10% and bonds' is zero, to have round figures.
A 70/30 AA has a 7% expected return.
If I DCA into that AA over, let's say, 2 years, I give up one year worth of return from a 70% stock allocation, i.e. 7%.
Therefore, after N years, the cumulative expected return of the LS and DCA investors are

N=5
LS investor: 40%
DCA investor: 31%

N=10
LS investor: 97%
DCA investor: 84%

N=20
LS investor: 387%
DCA investor: 362%

But you are saying I could go to a, let's say 60/40 AA to lower my risk ?
It's expected return is 6%, let's see what happens after N years:

N=5
60/30 AA: 34%

N=10
60/30 AA: 79%

N=20
60/30 AA: 321%

As you can see, expected return falls more and more behind the 70/30 AA, even when DCA is taken into account.
Whereas the DCA return keeps up with the LS and , in fact,the two returns converge asymptotically.
Thesaints,

Presumably you think that the DCA investor is getting something valuable in exchange for giving up some of the expected return relative to the LS investor. As you explained it above, the benefit seems to be a narrower dispersion of returns. When the DCA investor gets to the end of the 2-year period, could they not decide to sell what they had just purchased over the last two years, and DCA it back in over the next 2-years, in order to get the same exchange--narrower dispersion in exchange for lower expected returns? If they thought this was a good deal the first time, why is it not a good deal the second time?
Yeah, I am puzzled about that too. The arguments against even more aggressive or larger scale forms of holding cash tents or cash positions generally are that they will only reduce volatility of returns at the cost of lower expected returns. But that is true of DCA too--its claimed financial advantage over LSI was that it reduced volatility of returns at the cost of lower expected returns, so why is that now a bad not good tradeoff as applied to other cash strategies?

Again I know this takes a bit of a framing shift, but once you see cash and bond tents as a strategy broadly available to investors, not just to investors with newly-received funds, and you also see those tent strategies as variable in terms of asset composition, timing, scale, frequency, and so on, it just seems odd to me to insist that DCA of newly-received funds is necessarily the optimal tent strategy, and all the others must be worse.

I mean maybe it will happen to work out like that in some individual case, but it seems to me the odds are that for most investors, if any tent strategy at all is warranted, a different one is likely to be optimal.
HootingSloth
Posts: 1050
Joined: Mon Jan 28, 2019 2:38 pm

Re: When Does Dollar Cost Averaging Work?

Post by HootingSloth »

NiceUnparticularMan wrote: Tue Jul 20, 2021 1:55 am
HootingSloth wrote: Mon Jul 19, 2021 5:06 pm
Thesaints wrote: Mon Jul 19, 2021 2:52 pm
dogagility wrote: Sun Jul 18, 2021 6:03 am I disagree; the person should choose a different asset allocation.
Easy example:

Let's say stocks' expected return is 10% and bonds' is zero, to have round figures.
A 70/30 AA has a 7% expected return.
If I DCA into that AA over, let's say, 2 years, I give up one year worth of return from a 70% stock allocation, i.e. 7%.
Therefore, after N years, the cumulative expected return of the LS and DCA investors are

N=5
LS investor: 40%
DCA investor: 31%

N=10
LS investor: 97%
DCA investor: 84%

N=20
LS investor: 387%
DCA investor: 362%

But you are saying I could go to a, let's say 60/40 AA to lower my risk ?
It's expected return is 6%, let's see what happens after N years:

N=5
60/30 AA: 34%

N=10
60/30 AA: 79%

N=20
60/30 AA: 321%

As you can see, expected return falls more and more behind the 70/30 AA, even when DCA is taken into account.
Whereas the DCA return keeps up with the LS and , in fact,the two returns converge asymptotically.
Thesaints,

Presumably you think that the DCA investor is getting something valuable in exchange for giving up some of the expected return relative to the LS investor. As you explained it above, the benefit seems to be a narrower dispersion of returns. When the DCA investor gets to the end of the 2-year period, could they not decide to sell what they had just purchased over the last two years, and DCA it back in over the next 2-years, in order to get the same exchange--narrower dispersion in exchange for lower expected returns? If they thought this was a good deal the first time, why is it not a good deal the second time?
Yeah, I am puzzled about that too. The arguments against even more aggressive or larger scale forms of holding cash tents or cash positions generally are that they will only reduce volatility of returns at the cost of lower expected returns. But that is true of DCA too--its claimed financial advantage over LSI was that it reduced volatility of returns at the cost of lower expected returns, so why is that now a bad not good tradeoff as applied to other cash strategies?

Again I know this takes a bit of a framing shift, but once you see cash and bond tents as a strategy broadly available to investors, not just to investors with newly-received funds, and you also see those tent strategies as variable in terms of asset composition, timing, scale, frequency, and so on, it just seems odd to me to insist that DCA of newly-received funds is necessarily the optimal tent strategy, and all the others must be worse.

I mean maybe it will happen to work out like that in some individual case, but it seems to me the odds are that for most investors, if any tent strategy at all is warranted, a different one is likely to be optimal.
I think there are at least a few possible answers to the question I have posed, and they can lead someone in different directions, based on their circumstances.

One possible answer is that the tradeoff after 2 years is equally valuable to the investor, so they should take it again (and again and again) every 2 years. This would make them keep zigzagging their AA back and forth in a sawtooth pattern. I think this answer typically leads in the direction of saying that they should, instead, adopt a constant AA that is somewhere in between the most conservative and most aggressive points of this sawtooth AA. It will be simpler for them to maintain and get a similar equilibrium between risk and reward.

Another possible answer is that the tradeoff is not as valuable to them now because they retired at the beginning of the first one, and now their exposure to SORR is reduced. I think this answer typically leads in the direction of a longer lasting bond tent. The Big ERN study, and other examinations of SORR, show that this can be a benefit for people, but that something lasting along the order of a decade, rather than 2 years, is going to have a more consistent effect. SORR depends most on the cumulative return over the first decade, rather than the very early returns (which might, for example, be a crash with a sharp recovery or a short run up that is followed by a lost decade). A 2-year cash tent will not give you as much bang for your buck in reducing SORR.

A third possible answer is that the tradeoff is not as valuable to them now because they ended up being wealthier at the end of the 2 years, and so they now have a greater ability to take on risk. I think this answer typically leads in the direction of a portfolio-size-based glidepath like this one. This kind of glidepath will allow the investor to match their ability to take risk with an AA on a more principled basis, rather than by relying on the ad hoc effect of a DCA strategy.

DCA often seems to be appealing because a person can't quite figure out what their AA should be now that the windfall has changed their circumstances significantly. If they think about what is motivating them to do DCA carefully, I think there might be different justifications. Once the individual investor understands the particular justifications that apply in their personal situation, then they can find an AA glidepath that is better tailored to that situation. The AA employed in a DCA strategy is driven almost by random chance. You got this windfall, and you are letting the timing and size of the windfall determine the timing and height of your "tent" (and the fact that you have cash push you into holding cash, rather than bonds). But if you think about what the tent is doing for you (which can vary from person to person), then you can design a glidepath that is made just for you.
Global Market Portfolio + modest tilt towards volatility (80/20->60/40 as approach FI) + modest tilt away from exchange rate risk (80% global+20% U.S. stocks; currency-hedge bonds) + tax optimization
NiceUnparticularMan
Posts: 6103
Joined: Sat Mar 11, 2017 6:51 am

Re: When Does Dollar Cost Averaging Work?

Post by NiceUnparticularMan »

HootingSloth wrote: Tue Jul 20, 2021 6:24 am
NiceUnparticularMan wrote: Tue Jul 20, 2021 1:55 am
HootingSloth wrote: Mon Jul 19, 2021 5:06 pm
Thesaints wrote: Mon Jul 19, 2021 2:52 pm
dogagility wrote: Sun Jul 18, 2021 6:03 am I disagree; the person should choose a different asset allocation.
Easy example:

Let's say stocks' expected return is 10% and bonds' is zero, to have round figures.
A 70/30 AA has a 7% expected return.
If I DCA into that AA over, let's say, 2 years, I give up one year worth of return from a 70% stock allocation, i.e. 7%.
Therefore, after N years, the cumulative expected return of the LS and DCA investors are

N=5
LS investor: 40%
DCA investor: 31%

N=10
LS investor: 97%
DCA investor: 84%

N=20
LS investor: 387%
DCA investor: 362%

But you are saying I could go to a, let's say 60/40 AA to lower my risk ?
It's expected return is 6%, let's see what happens after N years:

N=5
60/30 AA: 34%

N=10
60/30 AA: 79%

N=20
60/30 AA: 321%

As you can see, expected return falls more and more behind the 70/30 AA, even when DCA is taken into account.
Whereas the DCA return keeps up with the LS and , in fact,the two returns converge asymptotically.
Thesaints,

Presumably you think that the DCA investor is getting something valuable in exchange for giving up some of the expected return relative to the LS investor. As you explained it above, the benefit seems to be a narrower dispersion of returns. When the DCA investor gets to the end of the 2-year period, could they not decide to sell what they had just purchased over the last two years, and DCA it back in over the next 2-years, in order to get the same exchange--narrower dispersion in exchange for lower expected returns? If they thought this was a good deal the first time, why is it not a good deal the second time?
Yeah, I am puzzled about that too. The arguments against even more aggressive or larger scale forms of holding cash tents or cash positions generally are that they will only reduce volatility of returns at the cost of lower expected returns. But that is true of DCA too--its claimed financial advantage over LSI was that it reduced volatility of returns at the cost of lower expected returns, so why is that now a bad not good tradeoff as applied to other cash strategies?

Again I know this takes a bit of a framing shift, but once you see cash and bond tents as a strategy broadly available to investors, not just to investors with newly-received funds, and you also see those tent strategies as variable in terms of asset composition, timing, scale, frequency, and so on, it just seems odd to me to insist that DCA of newly-received funds is necessarily the optimal tent strategy, and all the others must be worse.

I mean maybe it will happen to work out like that in some individual case, but it seems to me the odds are that for most investors, if any tent strategy at all is warranted, a different one is likely to be optimal.
I think there are at least a few possible answers to the question I have posed, and they can lead someone in different directions, based on their circumstances.

One possible answer is that the tradeoff after 2 years is equally valuable to the investor, so they should take it again (and again and again) every 2 years. This would make them keep zigzagging their AA back and forth in a sawtooth pattern. I think this answer typically leads in the direction of saying that they should, instead, adopt a constant AA that is somewhere in between the most conservative and most aggressive points of this sawtooth AA. It will be simpler for them to maintain and get a similar equilibrium between risk and reward.

Another possible answer is that the tradeoff is not as valuable to them now because they retired at the beginning of the first one, and now their exposure to SORR is reduced. I think this answer typically leads in the direction of a longer lasting bond tent. The Big ERN study, and other examinations of SORR, show that this can be a benefit for people, but that something lasting along the order of a decade, rather than 2 years, is going to have a more consistent effect. SORR depends most on the cumulative return over the first decade, rather than the very early returns (which might, for example, be a crash with a sharp recovery or a short run up that is followed by a lost decade). A 2-year cash tent will not give you as much bang for your buck in reducing SORR.

A third possible answer is that the tradeoff is not as valuable to them now because they ended up being wealthier at the end of the 2 years, and so they now have a greater ability to take on risk. I think this answer typically leads in the direction of a portfolio-size-based glidepath like this one. This kind of glidepath will allow the investor to match their ability to take risk with an AA on a more principled basis, rather than by relying on the ad hoc effect of a DCA strategy.

DCA often seems to be appealing because a person can't quite figure out what their AA should be now that the windfall has changed their circumstances significantly. If they think about what is motivating them to do DCA carefully, I think there might be different justifications. Once the individual investor understands the particular justifications that apply in their personal situation, then they can find an AA glidepath that is better tailored to that situation. The AA employed in a DCA strategy is driven almost by random chance. You got this windfall, and you are letting the timing and size of the windfall determine the timing and height of your "tent" (and the fact that you have cash push you into holding cash, rather than bonds). But if you think about what the tent is doing for you (which can vary from person to person), then you can design a glidepath that is made just for you.
Great post. I'll just specifically endorse the bolded part, because I really think that is the right landing point for this entire discussion.
Post Reply