Yet another Vanguard study on DCA vs Lump Sum

Have a question about your personal investments? No matter how simple or complex, you can ask it here.
Topic Author
canbonbon
Posts: 88
Joined: Mon Jan 05, 2015 6:15 am

Yet another Vanguard study on DCA vs Lump Sum

Post by canbonbon »

Vanguard did another study of DCA vs Lump Sum (as if the earlier one was not enough: viewtopic.php?t=101965)

https://personal.vanguard.com/pdf/ISGDCA.pdf

Here is the link to the old 2012 Study if you are interested:
https://personal.vanguard.com/pdf/s315.pdf
dbr
Posts: 34195
Joined: Sun Mar 04, 2007 9:50 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by dbr »

I think the methodology of trying to resolve the issue by comparing performance is ridiculous. Do Vanguard and anyone else contemplating this question not understand the meaning of expected return and what that is for different asset classes? I will grant it is mildly interesting to see a quantification for certain scenarios, but the essence of the question is not even a question.
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

dbr wrote:I think the methodology of trying to resolve the issue by comparing performance is ridiculous. Do Vanguard and anyone else contemplating this question not understand the meaning of expected return and what that is for different asset classes? I will grant it is mildly interesting to see a quantification for certain scenarios, but the essence of the question is not even a question.
Over time I've seen a lot of misleading DCA literature which is implying it might increase your expected returns.

Edit: This is typical:

http://www.investopedia.com/terms/d/dol ... raging.asp

Note lines like:
The investor purchases more shares when prices are low and fewer shares when prices are high. The premise is that DCA lowers the average share cost over time, increasing the opportunity to profit.
The example given there is typical. It assumes the stock price now is $20, and the average stock price over the next four months is only $17. This can happen, but odds are if the stock has a positive expected return, the expected average stock price over the next four months should be higher, not lower, than $20. And that's the typical "trick" used to motivate DCA as a profitable strategy, contrary to the actual expectation:
If the investor had invested all $5,000 on one of these days instead of spreading the investment across five months, the total profitability of the position would be higher or lower than $6,857.11 depending on the month chosen for the investment. However, no one can time the market. DCA is a safe strategy to ensure an overall favorable average price per share.
User avatar
Phineas J. Whoopee
Posts: 9675
Joined: Sun Dec 18, 2011 6:18 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by Phineas J. Whoopee »

My view, which I've posted before, is if one's allocation is clearly wrong (including because of receiving a large amount of cash), and another is clearly better, it makes the most sense to go from clearly wrong to clearly better all at once, rather than spend still more time with a less-clearly wrong allocation.

Not everybody agrees.

PJW
Last edited by Phineas J. Whoopee on Fri Apr 14, 2017 12:44 pm, edited 1 time in total.
Northern Flicker
Posts: 6860
Joined: Fri Apr 10, 2015 12:29 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by Northern Flicker »

Do we really even need an empirical study to show that DCA lowers expected return, but also lowers variance of return? If expected return is greater than zero for an asset, then lump sum always has a higher expected return. If variance of return is greater than zero then the variance of an average of returns from several starting points will be lower than from a single one.

What would be more interesting is some way of estimating risk-adjusted returns across both strategies.
Risk is not a guarantor of return.
dbr
Posts: 34195
Joined: Sun Mar 04, 2007 9:50 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by dbr »

NiceUnparticularMan wrote:
dbr wrote:I think the methodology of trying to resolve the issue by comparing performance is ridiculous. Do Vanguard and anyone else contemplating this question not understand the meaning of expected return and what that is for different asset classes? I will grant it is mildly interesting to see a quantification for certain scenarios, but the essence of the question is not even a question.
Over time I've seen a lot of misleading DCA literature which is implying it might increase your expected returns.

Edit: This is typical:

http://www.investopedia.com/terms/d/dol ... raging.asp

Note lines like:
The investor purchases more shares when prices are low and fewer shares when prices are high. The premise is that DCA lowers the average share cost over time, increasing the opportunity to profit.
The example given there is typical. It assumes the stock price now is $20, and the average stock price over the next four months is only $17. This can happen, but odds are if the stock has a positive expected return, the expected average stock price over the next four months should be higher, not lower, than $20. And that's the typical "trick" used to motivate DCA as a profitable strategy, contrary to the actual expectation:
If the investor had invested all $5,000 on one of these days instead of spreading the investment across five months, the total profitability of the position would be higher or lower than $6,857.11 depending on the month chosen for the investment. However, no one can time the market. DCA is a safe strategy to ensure an overall favorable average price per share.
DCA in the original and in my opinion only correct use of the term does increase returns. The original use of the term, which has nothing to do with an alternative of investing a lump sum, is a comparison of periodically investing a fixed dollar amount, called "dollar cost averaging" compared to investing at the same intervals in a fixed number of shares. Simple arithmetic shows that the first produces a lower cost per share than the second. So the benefit is real. This, as I said, has nothing to do with how to handle a lump sum. The original concept of DCA applied when one bought integral numbers of shares in individual stocks and preferably in round lots. For investors in mutual funds in fractional shares down to the thousandths, the concept is outmoded and irrelevant though the terminology has migrated to the lump sum debate.
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

dbr wrote:DCA in the original and in my opinion only correct use of the term does increase returns. The original use of the term, which has nothing to do with an alternative of investing a lump sum, is a comparison of periodically investing a fixed dollar amount, called "dollar cost averaging" compared to investing at the same intervals in a fixed number of shares. Simple arithmetic shows that the first produces a lower cost per share than the second. So the benefit is real.
If you assume an upward tendency to share prices, this is true. But that is only true because it means you are investing more money earlier.
For investors in mutual funds in fractional shares down to the thousandths, the concept is outmoded and irrelevant though the terminology has migrated to the lump sum debate.
No argument here, but that migration happened a long time ago, and it is quite widespread.

So that is what Vanguard is dealing with--an erroneous but widespread argument that retail investors are likely to encounter pretty often.
User avatar
iceport
Posts: 4662
Joined: Sat Apr 07, 2007 4:29 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by iceport »

dbr wrote:I think the methodology of trying to resolve the issue by comparing performance is ridiculous. Do Vanguard and anyone else contemplating this question not understand the meaning of expected return and what that is for different asset classes? I will grant it is mildly interesting to see a quantification for certain scenarios, but the essence of the question is not even a question.
I agree with (most of) this. It's like someone questioning which portfolio has a higher expected return, a 100% equity portfolio, or a 50% equity/50% fixed income portfolio.

That is not to say, however, that we all shouldn't spend some time figuring out what asset allocation — or what method of entry into the market for that matter — is right for each one of us individually. In that respect, the quantification is helpful.

But to focus attention on the crux of the matter (as dbr has helped me define it for me personally), maybe there needs to be an additional line of questioning based more on the human nature side of investors, just as there is when helping investors determine an appropriate AA.

Here's a sample question:
Imagine you have contributed regularly to your portfolio over the past 20 years through small periodic contributions coinciding with your paychecks. Over the course of, say, two months your portfolio decreases in value by 25% due to a severe recession.

Now imagine you had invested an equivalent lump sum all at once just two months before the loss of 25% of your portfolio value. Unlike the above scenario, in which your portfolio was purchased over hundreds of individual entry points over two decades, in this scenario your 25% loss in value was predominantly the result of the unlucky timing of a single entry point into the market, and a mere 60 day difference in timing would have made a huge difference in the outcome.

Would you be willing to accept a lower expected return for the first year to mitigate the effects of a possible sudden downturn by spreading your entry into the market across, say, 12 different price points over the course of the next 12 months through a method called DCA?
In considering the above question, an investor might want to know things like: "How much lower would the first-year expected return be if DCA is used?" and "By how much might DCA attenuate the loss of a sudden downturn in the first year?"

To help answer those questions, the most recent Vanguard paper is a step in the right direction, and an analysis like this is also very helpful: Do Not Dollar-Cost-Average for More than Twelve Months
Last edited by iceport on Fri Apr 14, 2017 12:58 pm, edited 1 time in total.
"Discipline matters more than allocation.” ─William Bernstein
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

jalbert wrote:What would be more interesting is some way of estimating risk-adjusted returns across both strategies.
The risk is being diminished solely because while you are DCA, your asset allocation includes more cash than you would otherwise target. Under standard portfolio theory, holding more cash doesn't improve your risk-adjusted returns, it just reduces risk and returns together.

Of course this is a short-term thing--if you are assuming you DCA over a year then hold for 30 years, it makes very little difference to expected return/risk.
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

iceport wrote:and an analysis like this is also very helpful: Do Not Dollar-Cost-Average for More than Twelve Months
Just to be clear, that is just re-duplicating the expected consequences of holding more cash in your allocation for a longer time before finally getting back to your target allocation.
User avatar
iceport
Posts: 4662
Joined: Sat Apr 07, 2007 4:29 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by iceport »

NiceUnparticularMan wrote:
iceport wrote:and an analysis like this is also very helpful: Do Not Dollar-Cost-Average for More than Twelve Months
Just to be clear, that is just re-duplicating the expected consequences of holding more cash in your allocation for a longer time before finally getting back to your target allocation.
Yes, I see that one could choose to see it that way. But I also see this from a different perspective.

So in the interest of clarity, though it may be an unintended consequence, the intent is *not* to adopt a temporarily more conservative AA. The intent is to spread the entry into the market across many different price points. Unfortunately, that can only be accomplished over a span of time. But, from my perspective, if diversifying entry prices could magically be collapsed to a single event, that would be consistent with the intent of DCA (as we are defining it in this thread).
"Discipline matters more than allocation.” ─William Bernstein
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

iceport wrote:So in the interest of clarity, though it may be an unintended consequence, the intent is *not* to adopt a temporarily more conservative AA. The intent is to spread the entry into the market across many different price points.
I'm sure that is true, but there is no expected risk/return benefit to spreading entry over many different FUTURE price points, because the expectation (presumably) is that market prices are going to average higher in the future than the current price.

I mean, suppose I presented the following options:

Plan A: hold 100% cash for 10 years, then go 100% stock for the next 10 years.

Plan B: go 100% stock for the next 20 years.

I'd hope most of us could understand that Plan A is going to reduce short-term volatility, but at the price of long-term expected returns. And that is because odds are, in 10 years the price of stocks will be higher.

DCA is really just a variation on this theme. That doesn't mean I am against it, but we should have no illusions about it, and that means realizing it really is just a mild form of temporarily deciding to have a higher cash allocation. And it has exactly the expected benefits of such a plan, and no more, regardless of what people may intend.
gkaplan
Posts: 7034
Joined: Sat Mar 03, 2007 8:34 pm
Location: Portland, Oregon

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by gkaplan »

Phineas J. Whoopee wrote:My view, which I've posted before, is if one's allocation is clearly wrong (including because of receiving a large amount of cash), and another is clearly better, it makes the most sense to go from clearly wrong to clearly better all at once, rather than spend still more time with a less-clearly wrong allocation.

Not everybody agrees.

PJW
I, for one, agree with you.
Gordon
GoldenFinch
Posts: 2334
Joined: Mon Nov 10, 2014 11:34 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by GoldenFinch »

gkaplan wrote:
Phineas J. Whoopee wrote:My view, which I've posted before, is if one's allocation is clearly wrong (including because of receiving a large amount of cash), and another is clearly better, it makes the most sense to go from clearly wrong to clearly better all at once, rather than spend still more time with a less-clearly wrong allocation.

Not everybody agrees.

PJW
I, for one, agree with you.
Me too! Especially when stated like that.
User avatar
iceport
Posts: 4662
Joined: Sat Apr 07, 2007 4:29 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by iceport »

NiceUnparticularMan wrote:
iceport wrote:So in the interest of clarity, though it may be an unintended consequence, the intent is *not* to adopt a temporarily more conservative AA. The intent is to spread the entry into the market across many different price points.
I'm sure that is true, but there is no expected risk/return benefit to spreading entry over many different FUTURE price points, because the expectation (presumably) is that market prices are going to average higher in the future than the current price.

I mean, suppose I presented the following options:

Plan A: hold 100% cash for 10 years, then go 100% stock for the next 10 years.

Plan B: go 100% stock for the next 20 years.

I'd hope most of us could understand that Plan A is going to reduce short-term volatility, but at the price of long-term expected returns. And that is because odds are, in 10 years the price of stocks will be higher.

DCA is really just a variation on this theme. That doesn't mean I am against it, but we should have no illusions about it, and that means realizing it really is just a mild form of temporarily deciding to have a higher cash allocation. And it has exactly the expected benefits of such a plan, and no more, regardless of what people may intend.
I can't really disagree with your statements, though your example is rather extreme. I agree with your prior assertion, to put things in proper perspective:
NiceUnparticularMan wrote:Of course this is a short-term thing--if you are assuming you DCA over a year then hold for 30 years, it makes very little difference to expected return/risk.
:beer
"Discipline matters more than allocation.” ─William Bernstein
Northern Flicker
Posts: 6860
Joined: Fri Apr 10, 2015 12:29 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by Northern Flicker »

The risk is being diminished solely because while you are DCA, your asset allocation includes more cash than you would otherwise target. Under standard portfolio theory, holding more cash doesn't improve your risk-adjusted returns, it just reduces risk and returns together.
That's fair. I was more thinking each DCA installnent would bring the portfolio back in balance, which in a tax-qualified account would just be equivalent in risk-adjusted terms to a lump sum investment and rebalancing at the DCA points in time. In a taxable account the tax drag on rebalancing comes into play and the rebalancing capability of DCA might lead to a higher risk-adjusted return from buying the lower returning assets.
Risk is not a guarantor of return.
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

jalbert wrote:That's fair. I was more thinking each DCA installnent would bring the portfolio back in balance, which in a tax-qualified account would just be equivalent in risk-adjusted terms to a lump sum investment and rebalancing at the DCA points in time. In a taxable account the tax drag on rebalancing comes into play and the rebalancing capability of DCA might lead to a higher risk-adjusted return from buying the lower returning assets.
But if you include the cash not invested in your analysis, your portfolio is never actually in balance until the end. And if there is any difference in taxation, it is going to be precisely because your portfolio was off target.

Again, I think it is helpful to consider somewhat extreme examples to clarify what is happening in these hypothetical scenarios at a smaller scale. Your target allocation is 50% stocks and 50% bonds. You've got $10,000.

Plan A: Lump sum at target, then rebalance in 10 years.

Plan B: Hold cash for 10 years, then invest at target.

For simplicity, imagine stocks increase by 100%, bonds increase by 50%, cash increases by 10%.

OK, after 10 years . . .

Plan A: Now we have $10,000 stock, $7,500 bonds. Oh no, we need to sell some stock to rebalance back to 50-50! Capital gains tax, bummer. Of course we may have to pay some tax on those bond returns too (depending on our bond strategy), and probably the stocks had some dividend returns along the way on which there is tax . . . . Still, not a bad result.

Plan B: $11,000 in cash. We' may have to pay some tax on that $1000 (depends on our cash strategy). Whatever is left we use to finally buy our stocks and bonds.

Did we pay less in tax in Plan B? Probably. But that's only because we held cash instead of stocks and bonds, and the return on cash was lower, so we paid less tax.

At the end of the day, when determining your target allocations in a taxable account, ideally you are using post-tax expected returns in your calculus. But once you have done that, DCA isn't going to improve your risk-adjusted post-tax expected returns, it will again just be like for some period you held more cash than target.
Wild Willie
Posts: 155
Joined: Mon Jul 05, 2010 10:15 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by Wild Willie »

This is giving me a headache :happy
lostdog
Posts: 3601
Joined: Thu Feb 04, 2016 2:15 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by lostdog »

I wish we could lump sum. I'll take DCA which is the second best. I think most people can only DCA anyway.
Brokerage: VTI+VXUS || Retirement: VTWAX || Short-Term: Cash+BSV || 33x Expenses
dbr
Posts: 34195
Joined: Sun Mar 04, 2007 9:50 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by dbr »

lostdog wrote:I wish we could lump sum. I'll take DCA which is the second best. I think most people can only DCA anyway.
Yes, you are right. Most people are able to invest a regular amount of money at a regular interval. That is a very good plan.

But, you can lump sum. Save up all your potential investments for a year or two and then lump sum the investment. While one is waiting one won't have to worry about the money being lost in a market crash. Another way to obtain a lump sum is to sell a bunch of stuff and then lump sum reinvest it. Actually that is the effect of choosing to keep everything invested every day. Of course, if one were to prefer DCA, one could sell a bunch of stuff and DCA it back in. Again, the benefit would be that money is not exposed for awhile to a decline in the market.
Last edited by dbr on Sat Apr 15, 2017 8:30 am, edited 1 time in total.
ray333
Posts: 177
Joined: Thu Mar 09, 2017 8:47 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by ray333 »

And if you expect shares to drop isn't DCA the way to go (for the short term at least)

Also why would not investing for a year and then dumping 20K into the market in 2018 be better than DCA'ing 2K starting now?
lostdog
Posts: 3601
Joined: Thu Feb 04, 2016 2:15 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by lostdog »

dbr wrote:
lostdog wrote:I wish we could lump sum. I'll take DCA which is the second best. I think most people can only DCA anyway.
Yes, you are right. Most people are able to invest a regular amount of money at a regular interval. That is a very good plan.

But, you can lump sum. Save up all your potential investments for a year or two and then lump sum the investment. While one is waiting one won't have to worry about the money being lost in a market crash. Another way to obtain a lump sum is to sell a bunch of stuff and then lump sum reinvest it. Actually that is the effect of choosing to keep everything invested every day. Of course, if one were to prefer DCA, one could sell a bunch of stuff and DCA it back in. Again, the benefit would be that money is not exposed for awhile to a decline in the market.
Good point. I never thought of it that way.
Brokerage: VTI+VXUS || Retirement: VTWAX || Short-Term: Cash+BSV || 33x Expenses
dbr
Posts: 34195
Joined: Sun Mar 04, 2007 9:50 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by dbr »

ray333 wrote:And if you expect shares to drop isn't DCA the way to go (for the short term at least)

Also why would not investing for a year and then dumping 20K into the market in 2018 be better than DCA'ing 2K starting now?
Without a doubt if you expect the market to drop, waiting to invest is the best decision. Since the market is always poised for a possible drop the conclusion would be that one is better off never investing. In some cases that actually would be the advice.
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

lostdog wrote:I wish we could lump sum. I'll take DCA which is the second best. I think most people can only DCA anyway.
I'd actually say that making regular investments from your income as it is paid to you would be a series of lump sums (using the definitions of lump sum and DCA we have been working with).
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

ray333 wrote:And if you expect shares to drop isn't DCA the way to go (for the short term at least)
If you knew a drop below current prices was coming, then selling all your shares and buying back in when they dropped would be the way to go.

That's called market timing. It tends not to work out well. If you use the same logic to DCA, it is still market timing.
ray333
Posts: 177
Joined: Thu Mar 09, 2017 8:47 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by ray333 »

Say I plan to invest 25.5K annually (Roth max + taxable) ... it'd be better over time if I dumped that lump sum in January 1st, rather than maxing Roth and DCA the 20K over 5-10 months?
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

dbr wrote:Without a doubt if you expect the market to drop, waiting to invest is the best decision. Since the market is always poised for a possible drop the conclusion would be that one is better off never investing. In some cases that actually would be the advice.
This is where I like to point out that stock prices dropping at some point (in fact many points) in the future is a virtual certainty.

Whether stock prices will ever drop below where they are NOW is a whole different issue.

And that difference is why market timing tends not to work.
dbr
Posts: 34195
Joined: Sun Mar 04, 2007 9:50 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by dbr »

NiceUnparticularMan wrote:
lostdog wrote:I wish we could lump sum. I'll take DCA which is the second best. I think most people can only DCA anyway.
I'd actually say that making regular investments from your income as it is paid to you would be a series of lump sums (using the definitions of lump sum and DCA we have been working with).
Yes, that is actually true even though ironically investing a fixed amount at fixed intervals is the original definition of DCA. I sure wonder who fetched on DCA as the term for the alternative to investing a lump sum right away.
dbr
Posts: 34195
Joined: Sun Mar 04, 2007 9:50 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by dbr »

NiceUnparticularMan wrote:
dbr wrote:Without a doubt if you expect the market to drop, waiting to invest is the best decision. Since the market is always poised for a possible drop the conclusion would be that one is better off never investing. In some cases that actually would be the advice.
This is where I like to point out that stock prices dropping at some point (in fact many points) in the future is a virtual certainty.

Whether stock prices will ever drop below where they are NOW is a whole different issue.

And that difference is why market timing tends not to work.
Another good point, of course.
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

ray333 wrote:Say I plan to invest 25.5K annually (Roth max + taxable) ... it'd be better over time if I dumped that lump sum in January 1st, rather than maxing Roth and DCA the 20K over 5-10 months?
Well, you'd have a somewhat higher expected return (and somewhat higher expected volatility) if you lump sum the $20K. That's because what you would be doing with DCA is equivalent to adding more cash to your overall allocation (and locating it in your taxable account), which lowers expected returns (and lowers expected volatility). For example, I believe within a reasonable degree of approximation, DCA $20K over 6 months each year would be like lump summing the $20K January 1 each year but changing your allocation plan to include $5K more in cash in your taxable account.

Note if you end up having a large enough overall portfolio, $5K more in cash in your taxable account over the years probably wouldn't make much difference in outcomes. That's why in the long run, this tends to be a pretty low stakes issue. So if you get an emotional/psychological/behavioral/relationship/etc. benefit out of DCA, that could be a price worth paying.
dbr
Posts: 34195
Joined: Sun Mar 04, 2007 9:50 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by dbr »

NiceUnparticularMan wrote:
Note if you end up having a large enough overall portfolio, $5K more in cash in your taxable account over the years probably wouldn't make much difference in outcomes. That's why in the long run, this tends to be a pretty low stakes issue. So if you get an emotional/psychological/behavioral/relationship/etc. benefit out of DCA, that could be a price worth paying.
I think there are better techniques for obtaining emotional/psychological/behavioral/relationship/etc. than falling back on irrational investing practices. Depending on irrationality to solve psychological problems is itself a psychological problem.
ray333
Posts: 177
Joined: Thu Mar 09, 2017 8:47 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by ray333 »

Is it fair to assume that for a 20+ year investment horizon, today's seemingly "high priced" share cost isn't that relevant?

I've been DCAing since my initial lump sum and buying back in as prices have dropped, but considering just doing a January 1 lump sum for the foreseeable future now that it's up and running (finally, after several account transfers)

(I'm not risk averse and will only be investing any $ I know I won't need for several years)
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

dbr wrote:Yes, that is actually true even though ironically investing a fixed amount at fixed intervals is the original definition of DCA. I sure wonder who fetched on DCA as the term for the alternative to investing a lump sum right away.
I know that usage was already around circa 2000, when I started investing.

Just speculating, my guess is it started with advisers who had some clients who were sitting on cash they were afraid to invest, and the advisers wanted them to make some transactions (for the usual reasons).
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

dbr wrote:I think there are better techniques for obtaining emotional/psychological/behavioral/relationship/etc. than falling back on irrational investing practices. Depending on irrationality to solve psychological problems is itself a psychological problem.
To be clear, I don't do it myself. But if it works for some people on those dimensions, and it doesn't have much expected cost, then I am not going to give them a hard time about it.

Yeah, that is "irrational," but just telling themselves to be more rational doesn't tend to work for a lot of folks.
ray333
Posts: 177
Joined: Thu Mar 09, 2017 8:47 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by ray333 »

DCA is "irrational"?
dbr
Posts: 34195
Joined: Sun Mar 04, 2007 9:50 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by dbr »

ray333 wrote:DCA is "irrational"?
Yes
User avatar
blaugranamd
Posts: 594
Joined: Wed Apr 11, 2012 1:57 pm
Location: D-lux apt in the sky

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by blaugranamd »

The often overlooked component of this is that, in general, expected return is POSITIVE, so between any Day A and Day B the expectation is the share price will INCREASE. Thus, any DCA strategy is more likely to result in purchasing shares at a progressively higher price. The evidence-based investor, which we should all strive to be, would suggest you should always lump sum to avoid this. DCA is really founded in the concept of loss aversion and nearly every discussion of DCA vs Lump Sum tends to revolve around market crashes, rather than expected returns. Suffering a loss immediately after investing is psychologically harder than missing out on the same equivalent in gains but objectively they are equivalent. Further, is there any difference between Investor A who has arrived today at a $10k value 90/10 portfolio over 5 years of regular investments vs Investor B who has a $10k windfall and wishes to have a 90/10 portfolio on the same day? Both are at equal risk of loss at any time in the future and if DCA were to have a higher expected return for Investor B's "lump sum" then the advice for Investor A should be to sell their entire portfolio and DCA back in over the same time period as Investor B. I have never heard anyone recommend the latter.

Yes, DCA is irrational in the sense that it ignores the above logical conclusions. But it is rational for the psychological implications.
-- Don't mistake more funds for more diversity: Total Int'l + Total Market = 7k to 10k stocks -- | -- Market return does NOT = average nor 50th percentile, rather 80-90th percentile long term ---
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

ray333 wrote:Is it fair to assume that for a 20+ year investment horizon, today's seemingly "high priced" share cost isn't that relevant?
Well, to be fair, if this was the only money you would ever invest, and prices are about to drop a lot, then it could have permanent consequences whether you invested before or after the drop. Of course the reverse is true too--if prices are about to go up a lot, then it could also have permanent consequences whether you invested before or after the increase. This is basically saying if market timing worked, it could be enormously profitable. Unfortunately, no one can really make it work.

But that said, if instead this is just one small part of an overall long term investment plan, with lots more investments to come, then yes, the long term outcome is probably not going to be very sensitive to which you do.
I've been DCAing since my initial lump sum and buying back in as prices have dropped, but considering just doing a January 1 lump sum for the foreseeable future now that it's up and running (finally, after several account transfers)

(I'm not risk averse and will only be investing any $ I know I won't need for several years)
I feel confident telling you your plan for the future is a reasonable one, and there is no timing plan which is likely to do better.

By the way, as a matter of practice, for one reason or another I often let relatively small amounts of cash sit around before I get to investing them. Roughly speaking I try to account for that as part of my cash allocation, but still, I am not rushing to invest everything as soon as I theoretically could. But I have no illusions this is somehow going to help my overall performance.
ray333
Posts: 177
Joined: Thu Mar 09, 2017 8:47 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by ray333 »

Thanks. ive essentially been lump sum in my Roth for years as a set/forget approach, guess I need to view my taxable the same way

Yada yada time in Market > timing the market ... think I got it
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

ray333 wrote:DCA is "irrational"?
I'd say irrationality is characterized by doing something that is not reasonably calculated to address your goals.

If your goal is to improve your risk-adjusted returns, then I would agree that DCA is not reasonably calculated to address that goal.

But if your goal is to sleep better at night, it could be quite rational to DCA. Or not--it depends on your own psychology.

And personally, I think we ignore these psychological issues at our peril.
ray333
Posts: 177
Joined: Thu Mar 09, 2017 8:47 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by ray333 »

guess I'll just stop watching the news / listening to certain political podcasts, and blindly invest in my newly set up 3 fund lazy portfolio as much as I can ... (technically 2 funds at the moment, I'm not bothering with bonds at 32)
User avatar
stemikger
Posts: 4950
Joined: Thu Apr 08, 2010 5:02 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by stemikger »

Why not ask a brain surgeon.

https://www.youtube.com/watch?v=93hdzbB1OFg
Choose Simplicity ~ Stay the Course!! ~ Press on Regardless!!!
deltaneutral83
Posts: 1775
Joined: Tue Mar 07, 2017 4:25 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by deltaneutral83 »

Lump sum vs DCA for a windfall is a lot easier to do when you're not 3% off an all time high and an 8 year bull market, 3 months ago we were at the all time high. While it's obvious that 66% of the time you are better to lump sum, I don't think it's 66% in an all time high situation and the subsequent 12-18 months? I think it's more psychology because honestly, who on the retail side has the discipline to DCA over 12-18 months in a taxable account?
User avatar
iceport
Posts: 4662
Joined: Sat Apr 07, 2007 4:29 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by iceport »

NiceUnparticularMan wrote:
ray333 wrote:DCA is "irrational"?
I'd say irrationality is characterized by doing something that is not reasonably calculated to address your goals.

If your goal is to improve your risk-adjusted returns, then I would agree that DCA is not reasonably calculated to address that goal.

But if your goal is to sleep better at night, it could be quite rational to DCA. Or not--it depends on your own psychology.

And personally, I think we ignore these psychological issues at our peril.
I greatly appreciate your considerate, balanced tone NiceUnparticularMan. I have dbr to thank for finally getting me to realize this is solely a psychological issue (albeit with real-world ramifications), but I have yet to convince him, apparently, that the psychology is not meaningless.

Here I'll reiterate the scenario that would have me looking to DCA a relatively large sum — something on the order of 1/4 of a mid-accumulation stage portfolio or more. (The size of the lump sum matters.)
Imagine you have contributed regularly to your portfolio over the past 20 years through small periodic contributions coinciding with your paychecks. Over the course of, say, two months your portfolio decreases in value by 25% due to a severe recession.

Now imagine you had invested an equivalent lump sum all at once just two months before the loss of 25% of your portfolio value. Unlike the above scenario, in which your portfolio was purchased over hundreds of individual entry points over two decades, in this scenario your 25% loss in value was predominantly the result of the unlucky timing of a single entry point into the market, and a mere 60 day difference in timing would have made a huge difference in the outcome.

Would you be willing to accept a lower expected return for the first year to mitigate the effects of a possible sudden downturn by spreading your entry into the market across, say, 12 different price points over the course of the next 12 months through a method called DCA?
It seems to me, advocates of investing a huge lump sum all at once often also advocate never looking back. Well, more power to anyone capable of doing that! I'm not built that way. Knowing that, I would certainly choose to diversify purchase prices to help mitigate the enormous remorse that would be generated by the above scenario if it involved a significant portion of my life savings. And I see nothing irrational about that.

It continues to mystify me why the benefit of DCA is so often ridiculed, but the concept of "willingness" to withstand risk, in general, is not.
"Discipline matters more than allocation.” ─William Bernstein
ray333
Posts: 177
Joined: Thu Mar 09, 2017 8:47 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by ray333 »

deltaneutral83 wrote:Lump sum vs DCA for a windfall is a lot easier to do when you're not 3% off an all time high and an 8 year bull market, 3 months ago we were at the all time high. While it's obvious that 66% of the time you are better to lump sum, I don't think it's 66% in an all time high situation and the subsequent 12-18 months? I think it's more psychology because honestly, who on the retail side has the discipline to DCA over 12-18 months in a taxable account?
my thoughts exactly ... but Ive been told on several occasions these new all time highs will happen many times over the course of my investment horizon so get money in ASAP and ride out this seemingly inevitable bear/correction around the corner ... the initial lump+DCA made sense for me given these short term concerns (and so far so good as I've been buying more shares at lower costs since my initial lump). now that all of my accounts are in order, I'd like to get back to the set/forget approach I've had the last few years ... guess I'll just DCA the remaining 6K then lump again in 6 months or so
randomguy
Posts: 9208
Joined: Wed Sep 17, 2014 9:00 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by randomguy »

dbr wrote:
ray333 wrote:DCA is "irrational"?
Yes
Is holding bonds irrational? After all on average you would be much better off investing 100% stocks. DCA is the same. On average you will end up with less money. But you avoid the issue of investing all of your cash on say March 2000 or August 2008. Is giving up a a couple of percent of return worth it? That is pretty personal. I have no problem lump summing in say 100k or so (call it 5-10% of net worth). Lump summing in something like 5 million (100-200% of networth) isn't something I think I could do.


In the end choices like this just don't matter. You will probably only have a couple of times to do this over your investing career where you are lump summing in 10%+ of your NW (assuming something like a 100k+ NW. We aren't talking about when a IRA contribution changes your networth by 20%:)). Missing out on 2% for 3 or 4 years isn't going to change much. Do what lets you sleep well at night and move on with living.

One thing to note is that large lump sums can change what is a reasonable AA. With my current networth and expectations, 80/20 seems like a good idea. Doubling my networth tomorrow might make 60/40 more reasonable as I would be hitting a point where I don't need much risk to get a sub 3% SWR. Going up 20x and 90/10 might be a better choice since I could afford to lose 75% of my money and still have enough.
randomguy
Posts: 9208
Joined: Wed Sep 17, 2014 9:00 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by randomguy »

ray333 wrote:
deltaneutral83 wrote:Lump sum vs DCA for a windfall is a lot easier to do when you're not 3% off an all time high and an 8 year bull market, 3 months ago we were at the all time high. While it's obvious that 66% of the time you are better to lump sum, I don't think it's 66% in an all time high situation and the subsequent 12-18 months? I think it's more psychology because honestly, who on the retail side has the discipline to DCA over 12-18 months in a taxable account?
my thoughts exactly ... but Ive been told on several occasions these new all time highs will happen many times over the course of my investment horizon so get money in ASAP and ride out this seemingly inevitable bear/correction around the corner ... the initial lump+DCA made sense for me given these short term concerns (and so far so good as I've been buying more shares at lower costs since my initial lump). now that all of my accounts are in order, I'd like to get back to the set/forget approach I've had the last few years ... guess I'll just DCA the remaining 6K then lump again in 6 months or so
If you look at past history, markets spend a lot of time within 3% of all time highs. It is a pretty normal state. And it isn't remotely obvious that it makes it a good or bad time. Taking 18 months to DCA doesn't help if the correction happens in month 19 after the markets go up another 15%. Lump summing when the markets are down 20% doesn't help if they drop another 30% over the next 24. Market timing is hard.:)
lostdog
Posts: 3601
Joined: Thu Feb 04, 2016 2:15 pm

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by lostdog »

So basically the extra cash we have for the month I'll have it sit until I use it to max out our Roth accounts on Jan 1st?

My current plan is the extra money we have for the month will go towards the Roth accounts until maxed. When Roth's are both maxed we then put the extra money into the total market index in our taxable account. Rinse and repeat for next year.
Brokerage: VTI+VXUS || Retirement: VTWAX || Short-Term: Cash+BSV || 33x Expenses
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

deltaneutral83 wrote:Lump sum vs DCA for a windfall is a lot easier to do when you're not 3% off an all time high and an 8 year bull market, 3 months ago we were at the all time high. While it's obvious that 66% of the time you are better to lump sum, I don't think it's 66% in an all time high situation and the subsequent 12-18 months?
If you expect the trend to be upward on average, very often you will be at or near "all time" highs. That phrase seems to imply there is something special about such an event, when in fact that is expected to happen on a regular basis.

Similarly, we are pretty much always in an X year "bull market", as long as you set X large enough. That's also a consequence of the trend being upward on average.
NiceUnparticularMan
Posts: 1740
Joined: Sat Mar 11, 2017 7:51 am

Re: Yet another Vanguard study on DCA vs Lump Sum

Post by NiceUnparticularMan »

iceport wrote:Knowing that, I would certainly choose to diversify purchase prices to help mitigate the enormous remorse that would be generated by the above scenario if it involved a significant portion of my life savings. And I see nothing irrational about that.
Me neither. See also the Wiki entry on Windfalls. They recommend putting it into something safe for a year before deciding what to do. Is this giving up a year of expected market returns? Sure, maybe. It is also sound advice? Of course! Practical wisdom involves recognizing that if most people do in fact feel things like loss aversion, we should not lightly assume we are exceptions.
It continues to mystify me why the benefit of DCA is so often ridiculed, but the concept of "willingness" to withstand risk, in general, is not.
Same. I think it is useful to point out that things like DCA, rebalancing, large bond allocations during accumulation, and so on might not have much benefit in strict financial terms, ASSUMING you would otherwise feel and behave like a robot. But having thought about that, you need to move back to dealing with real human beings, not robots.

Honestly, from my perspective, this is about putting a price on things. DCA for piece of mind has a price, but in many circumstances the price is small. Big swings in allocation based on market timing, however, can have a huge price. And in fact if doing the former can help you avoid doing the latter, that could be quite . . . rational.
Post Reply