A view on Dollar Cost Averaging vs Lump Sum

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
bertilak
Posts: 7847
Joined: Tue Aug 02, 2011 5:23 pm
Location: East of the Pecos, West of the Mississippi

A view on Dollar Cost Averaging vs Lump Sum

Post by bertilak »

The AAII puts out a booklet* that has an interesting quote, a hypothetical bet:
Suppose someone offers to flip a fair coin with you. If you loose you pay X dollars but if you win he will pay you 2X dollars.
Would you take that bet? Sounds like a no-brainer, but what if you were required to bet so that 2X is all of your net worth? The odds are very much in your favor (two to one) but I think you would, rightly, turn down the bet.

I think the above argues strongly in favor of Dollar Cost Averaging. For example, you should almost certainly take the bet if you could do so at, say, 10% of your net worth for each successive coin flip.

The above argument still holds when there is reason to believe that, on average, the coin will be weighted slightly in your favor for the earlier flips. (This represents the fact that, on average, the sooner you get into the market the better you will do, an argument in favor of lump sum investing I have heard right here on boglheads.org.)

Any comments?

* The booklet: Maximum Return Minimum Risk: A Practical Approach Second edition (2005) by James B. Cloonan, Ph.D.
May neither drought nor rain nor blizzard disturb the joy juice in your gizzard. -- Squire Omar Barker (aka S.O.B.), the Cowboy Poet
User avatar
Don Christy
Posts: 391
Joined: Sun Oct 11, 2009 10:33 pm

Post by Don Christy »

The problem with that hypothetical bet is that it is a poor model for lump sum versus DCA in the context of a well designed, diversified portfolio. If you are buying a single stock or bond or undiversied asset, I would probably agree with DCA into that "bet," but a typical BH portfolio is not equivalent to betting 50% of your net worth on a coin toss.

Try this thought experiment instead. Let's say you inherit a large portfolio, and by the grace of the gods, on the day you inherit it, it happens to match your preferred AA perfectly.

Would you:
1) immediately liquidate to cash and start DCAing back to your AA.
2) sleep well knowing that you have a well diversified portfolio appropriate to your risk tolerance and goals.

I would wager that almost everyone would choose 2.

IMO, the only reason to DCA is in the context of regular savings (i.e. funds available over time). If you receive a lump sum that you want to invest, DCA just keeps you in a less diversified AA longer, increases the risk of not meeting your goals (e.g. inflation risk), reduces your expected returns, and creates more work to ultimately get you to your desired AA.

Any other reason seems to be based on the psychology of loss aversion. And as good BHs, we should recognize these irrational psychological responses and not let them dictate our investment decisions.
“Speak only if it improves upon the silence." Mahatma Gandhi
User avatar
Topic Author
bertilak
Posts: 7847
Joined: Tue Aug 02, 2011 5:23 pm
Location: East of the Pecos, West of the Mississippi

Post by bertilak »

Don Christy wrote: Would you:
1) immediately liquidate to cash and start DCAing back to your AA.
2) sleep well knowing that you have a well diversified portfolio appropriate to your risk tolerance and goals.

I would wager that almost everyone would choose 2.
Don, thanks for the comments. I have another thread where I said that I was going with your second choice, so I guess I would have to agree that's a decent choice!

In that other thread, DCA vs. Lump Sum was mostly an aside to the main question in the thread, but since posting it I read the booklet I quoted above. That quote was in a section of the booklet not even discussing DCA, but the concept of risk aversion in general. Even so, it jumped out at me as relevant to DCA vs LS.

The hypothetical bet (X vs 2X) is an extreme, but looking at extremes is a valid way of analyzing things. So even if we are never in an all or nothing game, the hypothetical extreme does point out which way things lean.

In other words, I don't think LS vs DCA involves an all-or-nothing risk, and even if it DID, you could still get hit by a total loss the day after you had DCA'd everything in! BUT -- there truth here may have to do with volatility and its relationship to the overall market trend. I suspect there probably is some "best bet" rate at which money should be invested. Investing in a well-diversified portfolio should push things towards LS as the best choice, but push it all the way? I don't know.

It is a bit like insurance. Based on averages and long term trends, insurance is a suckers bet -- the house always wins in the end. But we buy insurance to shift risk around -- we "lose" (spend) what we can afford to lose in order to avoid the loss we can't afford (or prefer not to pay). So DCA is a bit like insurance. The math is against you if you don't take into account one-time events that could be devastating. The difference might be that the guarantee isn't as solid as real insurance.
May neither drought nor rain nor blizzard disturb the joy juice in your gizzard. -- Squire Omar Barker (aka S.O.B.), the Cowboy Poet
User avatar
Don Christy
Posts: 391
Joined: Sun Oct 11, 2009 10:33 pm

Post by Don Christy »

I agree with most of what you say, though perhaps we depart in one important aspect. I don't see DCA as providing any insurance. Whatever you are in, instead of your preferred AA, is not risk free (if risk is appropriately defined). Certainly being in cash is not risk free (e.g. inflation). Perhaps TIPS are risk free you say. Well, only if your goals can be met given the expected returns of TIPS.

A well diversified portfolio with appropriate levels of risk IS THE INSURANCE. IMO DCA is less efficient and increases overall portfolio risk. If you have an appropriately selected AA, the only benefit I see of DCA is purely psychological, and that's only if you ignore other risks.
bertilak wrote:
In other words, I don't think LS vs DCA involves an all-or-nothing risk, and even if it DID, you could still get hit by a total loss the day after you had DCA'd everything in! BUT -- there truth here may have to do with volatility and its relationship to the overall market trend. I suspect there probably is some "best bet" rate at which money should be invested. Investing in a well-diversified portfolio should push things towards LS as the best choice, but push it all the way? I don't know.

It is a bit like insurance. Based on averages and long term trends, insurance is a suckers bet -- the house always wins in the end. But we buy insurance to shift risk around -- we "lose" (spend) what we can afford to lose in order to avoid the loss we can't afford (or prefer not to pay). So DCA is a bit like insurance. The math is against you if you don't take into account one-time events that could be devastating. The difference might be that the guarantee isn't as solid as real insurance.
“Speak only if it improves upon the silence." Mahatma Gandhi
dbr
Posts: 33842
Joined: Sun Mar 04, 2007 9:50 am

Post by dbr »

So answer this question. If one prefers DCA, does that mean that at completion of the DCA process one should withdraw the money and start the DCA process over again?
User avatar
Aptenodytes
Posts: 3768
Joined: Tue Feb 08, 2011 8:39 pm

Re: A view on Dollar Cost Averaging vs Lump Sum

Post by Aptenodytes »

bertilak wrote:The AAII puts out a booklet* that has an interesting quote, a hypothetical bet:
Suppose someone offers to flip a fair coin with you. If you loose you pay X dollars but if you win he will pay you 2X dollars.
Would you take that bet? Sounds like a no-brainer, but what if you were required to bet so that 2X is all of your net worth? The odds are very much in your favor (two to one) but I think you would, rightly, turn down the bet.

I think the above argues strongly in favor of Dollar Cost Averaging.
I don't see it that way at all. The terms of the "bet" aren't fixed. You adjust them through your asset allocation. A 20/80 portfolio has different risk than 80/20. The best response to the situation is to modify the risk/return relationship to the point you are comfortable with it and then carry on.

If you have money in the market you are making a "bet" every hour of every day. You can't get around that, and if you have somehow tricked yourself into thinking that once money is in the market the risk evaporates there is a good chance you will end up unhappy down the line.
dbr
Posts: 33842
Joined: Sun Mar 04, 2007 9:50 am

Re: A view on Dollar Cost Averaging vs Lump Sum

Post by dbr »

Aptenodytes wrote:
If you have money in the market you are making a "bet" every hour of every day. You can't get around that, and if you have somehow tricked yourself into thinking that once money is in the market the risk evaporates there is a good chance you will end up unhappy down the line.
Exactly. I am convinced that at the heart of lump sum fear is an investor who is targeting an AA that is too risky. It is easy for that to happen when a person suddenly finds themselves with assets much larger than they are accustomed to.
dkturner
Posts: 1608
Joined: Sun Feb 25, 2007 7:58 pm

Post by dkturner »

dbr wrote:So answer this question. If one prefers DCA, does that mean that at completion of the DCA process one should withdraw the money and start the DCA process over again?
I have asked this very question over the years to individuals who have windfalls they would like to invest, but are uneasy about making an all in investment. None of them ever get it.

FWIW, for purely psychological reasons, DCA is probably not a bad idea, since a large loss from an all in investment at a market top is far more painful than the missed opportunity of a larger profit from DCAing into a rising market.
User avatar
dmcmahon
Posts: 2335
Joined: Fri Mar 21, 2008 10:29 pm

Post by dmcmahon »

The issue is investor psychology, not expected return. People are inclined to feel pain at losses more keenly than feel joy at gains. A better thought experiment is this:

Suppose you had $X to invest in stocks. In the first month you put the entire amount into the market. The market plunges immediately afterward - how are you feeling? Now suppose instead the market skyrockets - how are you feeling? Finally, consider both of those same outcomes after putting $X/10 into the market, with a plan to dribble the other 9/10 out over some period of time. Which is more likely to result in a panic response, and which is more likely to result in staying the course? Each person has to answer this for him/her self.

I think a more interesting question is, if you're going to DCA, over what period of time should you do so? Is a year long enough to spread out the investments and avoid the "gut check" issue? Suppose, for example, that you followed a DCA strategy over 2007...
User avatar
Aptenodytes
Posts: 3768
Joined: Tue Feb 08, 2011 8:39 pm

Post by Aptenodytes »

dmcmahon wrote:The issue is investor psychology, not expected return. People are inclined to feel pain at losses more keenly than feel joy at gains. A better thought experiment is this:

Suppose you had $X to invest in stocks. In the first month you put the entire amount into the market. The market plunges immediately afterward - how are you feeling? Now suppose instead the market skyrockets - how are you feeling? Finally, consider both of those same outcomes after putting $X/10 into the market, with a plan to dribble the other 9/10 out over some period of time. Which is more likely to result in a panic response, and which is more likely to result in staying the course? Each person has to answer this for him/her self.
I agree. For me the goal is to invest as fast as your emotions let you.
Dandy
Posts: 6357
Joined: Sun Apr 25, 2010 7:42 pm

Post by Dandy »

The arguments for lump sum are pretty solid. It is strange how we treat money differently e.g. gambling winnings are "found money" and are often spent differently than money already in our wallets.

I wonder how large a percentage of your portfolio would the lump sum have to be before the difference in returns between lump sum and DCA (over a year's time) would be significant? If the lump sum is 25% of your current portfolio then it might make a big difference if it is 5% then maybe not.

If it is a large lump sum then maybe your risk tolerance and resulting asset allocation might need to be reviewed. Do you still need to take your current level of risk to attain your goals? I find that the larger the portfolio the more risk averse I become.

Most of us have gradually accumulated our current portfolio and have become used to the ups and downs associated with it. So, in a sense, we are used to DCA. So, in spite of the great arguments for lump sum investing I still favor DCA for most significant large lump sum situations. Being comfortable with your investing approach is important - as long as it doesn't kill your potential returns.
User avatar
Topic Author
bertilak
Posts: 7847
Joined: Tue Aug 02, 2011 5:23 pm
Location: East of the Pecos, West of the Mississippi

Post by bertilak »

dmcmahon wrote:The issue is investor psychology, not expected return. People are inclined to feel pain at losses more keenly than feel joy at gains.
And rightly so, in my opinion. It is not a matter of some irrational (aka psychological) fear. There are losses that simply cannot be tolerated. Expected return does not take into account all relevant factors, including the extremes away from the average. Once something happens, you cannot take it back.

The key to DCA is the "A" -- "Averaging". The average is not as extreme as individual events, and it is the extremes that are at issue.

Asset allocation (AA) is one way to "average out" the extremes. DCA is another. Just because one is effective (AA) does not mean the other (DCA) isn't. Also, neither AA nor DCA is guaranteed to be 100% effective. Someday someone will get hurt in the market despite following best practices. Still, we strive for best practices. Anyone can get hit by a bus, even if they look both ways!

AA = average across space (sectors).
DCA = average across time.

You get DCA for "free" as you invest funds that become available periodically. If a lump sum of funds becomes available all at once, should we give up on DCA?
May neither drought nor rain nor blizzard disturb the joy juice in your gizzard. -- Squire Omar Barker (aka S.O.B.), the Cowboy Poet
User avatar
dmcmahon
Posts: 2335
Joined: Fri Mar 21, 2008 10:29 pm

Post by dmcmahon »

dkturner wrote:
dbr wrote:So answer this question. If one prefers DCA, does that mean that at completion of the DCA process one should withdraw the money and start the DCA process over again?
I have asked this very question over the years to individuals who have windfalls they would like to invest, but are uneasy about making an all in investment. None of them ever get it.
I would answer "no". There's a difference between having $X built up patiently over the years with a large fraction of $X being accumulated gains, and having $X saved out of the sweat of your brow and none of it being accumulated gains. It is easier psychologically (at least for me) to see losses versus past gains than to see losses versus "cash in". There are also practical tax consequences, since at most $3k of capital losses can be deducted per year. This can lead to some very depressing thoughts at market bottoms. When you have 100+ years of losses, and you've lost hope of ever having sufficient gains to erase the losses, the tax system adds insult to injury.
User avatar
Don Christy
Posts: 391
Joined: Sun Oct 11, 2009 10:33 pm

Post by Don Christy »

bertilak wrote:
dmcmahon wrote:The issue is investor psychology, not expected return. People are inclined to feel pain at losses more keenly than feel joy at gains.
And rightly so, in my opinion. It is not a matter of some irrational (aka psychological) fear. There are losses that simply cannot be tolerated. Expected return does not take into account all relevant factors, including the extremes away from the average. Once something happens, you cannot take it back.

The key to DCA is the "A" -- "Averaging". The average is not as extreme as individual events, and it is the extremes that are at issue.
If those losses can't be tolerated, the proposed AA is arguably not appropriate. As you and others have noted, that risk doesn't go away as a result of DCA.

If you have a wad of cash and choose to DCA, what you are doing is averaging one AA with another AA. This indicates a fear of the ultimate AA.

If you have periodically available cash, DCA of asset prices is the result of the activity, but represents a different kind of averaging. In this context you are always adding to your portfolio at the target AA.
Asset allocation (AA) is one way to "average out" the extremes. DCA is another. Just because one is effective (AA) does not mean the other (DCA) isn't.
IMO they are not the same at all.
Also, neither AA nor DCA is guaranteed to be 100% effective. Someday someone will get hurt in the market despite following best practices. Still, we strive for best practices. Anyone can get hit by a bus, even if they look both ways!
Well of course. If it was guaranteed, there would be no risk and therefore no risk premium. But if you define risk appropriately, I don't think DCA helps to reduce risk.

AA = average across space (sectors).
DCA = average across time.

You get DCA for "free" as you invest funds that become available periodically.
Funds you don't have are not really being risked and therefore are not the same as funds you keep in cash as you slowly DCA into your AA. DCA as part of periodic investment RESULTS in DCA, but is not a justification for DCA.
If a lump sum of funds becomes available all at once, should we give up on DCA?
Yes. :lol: Develop an AA that is appropriate and then (as someone else said) invest as fast as your emotions will allow.
“Speak only if it improves upon the silence." Mahatma Gandhi
User avatar
grabiner
Advisory Board
Posts: 27964
Joined: Tue Feb 20, 2007 11:58 pm
Location: Columbia, MD

Re: A view on Dollar Cost Averaging vs Lump Sum

Post by grabiner »

bertilak wrote:The AAII puts out a booklet* that has an interesting quote, a hypothetical bet:
Suppose someone offers to flip a fair coin with you. If you loose you pay X dollars but if you win he will pay you 2X dollars.
Would you take that bet? Sounds like a no-brainer, but what if you were required to bet so that 2X is all of your net worth? The odds are very much in your favor (two to one) but I think you would, rightly, turn down the bet.

I think the above argues strongly in favor of Dollar Cost Averaging. For example, you should almost certainly take the bet if you could do so at, say, 10% of your net worth for each successive coin flip.
This has nothing to do with dollar-cost averaging. If the stock market is expected to earn 10% (but might lose 50%) and the bond market is expected to earn 5% (and is unlikely to lose much), then you probably don't want 100% of your money in stock, but this is just as true after ten months of dollar-cost averaging as it is when you get started.

Rather, it is an argument for diversification; if you can invest your stock money in 10 stocks rather than 1 stock, you get the same expected return (and more after rebalancing) with a lot less risk.
Wiki David Grabiner
Post Reply