Many readers were surprised when I answered a recent Ask the Spud question by suggesting you’re usually better off investing a lump sum rather than using dollar-cost averaging (DCA).
DCA is popular technique that crops up all the time in personal finance books—David Chilton’s The Wealthy Barber was one of the first to popularize it back in the late 1980s. If you need a refresher, DCA is a strategy for investing a large sum gradually. For example, if you have $100,000 you might invest $25,000 today and the same amount in each of the next three quarters, or perhaps $5,000 at a time over 20 months.
The idea sounds appealing: if the markets plummet after you invest a lump sum, you’ll suffer a major loss and be filled with regret. However, by investing a little at a time you avoid putting all your money at risk immediately, and if markets decline you’ll benefit by making some of your purchases when prices are low.
Roy, the eponymous hero of The Wealthy Barber, made DCA seem like magic: “Dollar cost averaging is as close to infallible investing as you can get. It genuinely slants the odds in the investor’s favour.” Unfortunately, that’s simply not true. Academics have been studying dollar-cost averaging for more than 30 years, and the results are clear: DCA usually results in lower long-term returns.
The evidence is clear
There have been many studies comparing the alternatives, beginning with a landmark 1979 paper that declared DCA “suboptimal.” That result was confirmed in a 1993 paper that found DCA produced higher returns in just 27% to 39% of the scenarios it tested.
A much more recent examination of this question appears in a Vanguard paper called “Dollar-cost averaging just means taking risk later,” released last July. Using data from the U.S., United Kingdom and Australia, the researchers compared the results of investing a $1 million lump sum with using DCA over 12 months. In all three countries, over rolling 10-year periods, the lump-sum strategy came out ahead almost exactly two-thirds of the time for a portfolio of 60% equities and 40% bonds. They also ran the numbers using DCA periods from six to 36 months, and various mixes of stocks and bonds, with similar results.
So let’s be clear: using DCA does not stack the odds in your favour: on the contrary, lump-sum investing is at least a two-to-one favorite.
When DCA makes sense
Of course, people don’t make decisions based only on the odds of success, but also on the consequences of failure: you have a five in six chance of winning at Russian roulette, but you won’t find many takers. In the worst 5% of results in Vanguard’s U.S. simulation, the investor who went all-in with $1 million ended up with $200,000 less after 10 years, compared with DCA. Imagine how you would have felt if you invested a lump sum just before Black Monday in 1987, or in September 2008.
The fear of regret can’t be brushed off, so I share Larry Swedroe’s take on DCA. If you’re extremely anxious about going all-in, and you’re willing to accept the likelihood of lower returns, then a gradual investment is entirely reasonable. But stick to a strict timetable, investing equal amounts each month or each quarter, for example. And don’t spread it out longer than a year, as you’ll lower the odds even more. This simulation (hat tip to reader John for the link) found that using DCA over 36 months resulted in higher returns in just 30 cases out of 493—a dreadful 6% success rate.
You may also be interested in Rick Ferri’s thoughts on lump sum investing, which contradicts some of what I’ve argued here. Ferri suggests readers consider the source of the windfall (a pension payout is treated differently from a lottery win) and the size of the lump sum relative to your existing portfolio.
In the end, there’s no right or wrong answer, since the key factors are emotional. Just make sure you understand the trade-off. Investing the lump sum might result in higher blood pressure, but it’s also likely to deliver higher returns.
Dan,
While I agree that when investing a lump sum it is best to just invest everything immediately. I do think your quote from The Wealth Barber is somewhat out of context. Although admittedly it has been a while since I’ve read the book. I believe Chilton recommends using DCA to ‘pay yourself first’ by setting up a pre-authorized contribution plan to automatically deduct money from your paycheck and invest it. Which should have a higher expected return than the alternative of saving up cash throughout the year then investing it all at once right before the RRSP deadline.
I agree that you shouldn’t invest a large lump sum that way. But when Chilton refers to dollar cost averaging, I see it as buying at regular intervals with a fixed dollar amount. As in, investing a portion of your paycheque every month – say, $200/month.
I don’t recall him saying you should invest a large lump sum that way. The average investor isn’t dealing with lump sums, they put a couple hundred bucks away at regular intervals. That’s the DCA most people are accustomed to, isn’t it?
->If you need a refresher, DCA is a strategy for investing a large sum gradually.
I disagree, DCA is also a strategy for investing over the long term for an individual investor who does not have $200k sitting around or receive an unexpected windfall.
The question an average person is more likely to face is, do they invest $200 a month or do they wait until 12 months have passed and invest $2400? I believe the math will be reversed in those cases and DCA is very advantageous both from getting invested first and from minimizing downside risk.
The phrase “dollar-cost averaging” is used in two different contexts: (1) investing a lump sum over time, and (2) continuous automatic investing (usually from an income). Arguing over which of these is the correct definition is pointless — there are too many people in each camp to have any hope of unifying them.
Robb is right that Chilton meant definition (2). By definition (1), DCA is a losing choice but is more palatable to many people. By definition (2), DCA is good because it means that you’re not engaged in market timing (or worse, you’re not saving at all).
The last paragraph of the Wikipedia page (http://en.wikipedia.org/wiki/Dollar_cost_averaging) on DCA explains why criticisms of DCA by definition (1) can confuse (and do a disservice to) investors with definition (2) in mind.
DCA, when applied to a lump sum, is a lottery ticket. Most people will lose money, but about once a decade if you do it at just the right time you’ll save 10-30%. Many people play the game in hopes of winning the prize.
I would argue that lump sum methodology is exactly what (2) above, is. When you have money, you invest it. You don’t save up for the end of the year, etc. While this does average the cost of what you buy, it also means that you buy as soon as you can, contradicting what (1) is.
If I were to have a large sum of money to invest, risk mitigation would be a larger concern than maximizing my return. When I look at these studies, the focus seems to be on return (mean) rather than the risk (standard deviation). The emprical data in the ‘1993’ study seems to show a lower standard deviation for the DCA approach. This seems to be particularly true as the percentage of risky assets increases. I am no math wizard, but in my opinion, these studies seem to focus more on reward than on risk.
I am not sure how to assess a decrease in return when risk is also decreasing. A small decrease in return may be worth it when accompanied by a large decrease in risk. How do these studies incorporate the risk with the return? Is there a metric that combines risk and return together?
Jeff, uncertainty is not the same thing as risk. If I offer you two investments, one of which is guaranteed to lose 10% in the next year and one which will give you positive returns of 0 – 10%, the first one has no volatility. And yet you would probably not take it.
Thanks for another great post! I would be interested in hearing your opinion on value averaging, which according to some scholars does provide slightly higher returns over time.
My point is that returns should be looked at on a risk-adjusted basis instead of an absolute basis. A portfolio that is mostly cash for the first year is expected to be less risky than one that is not. Conversely, a portfolio that has little to no cash in the first year is expected to generate a higher return. The question is: for the same level of risk, which portfolio provides a better return? I am pretty that the academics somehow deal with this (a sharpe ratio?), but I am not sure how and hope for further explanation.
Value Indexer, your example seems to have two examples of risk because the distributions of the outcomes are known in both cases. In the first case, the risk is zero, while in the second it’s anywhere 0-10%. But it does demonstrate why risk and return should be looked at together.
An all-cash portfolio will be less volatile. If volatility was all that mattered no one would invest in stock markets. They would then have a very high risk of not being able to retire despite having a low volatility (remember that opportunity cost is a risk too). Choosing what to do with a lump sum is essentially the same thing at a smaller scale. It’s not something that’s easy to fine-tune because there are a lot of unknowns, so I am comfortable with the choice that has the best result most often. Anything else would be a gamble.
Regardless of that there are two other factors that override this. First, most people aren’t 100% in stocks so their normal asset allocation will already act to reduce volatility even if they don’t keep anything in cash. And second, if you receive a large lump sum that may put you in a position where you actually want a more conservative allocation. In either case we’re not talking about someone putting it all in the stock market.
Agree with those above – I think there are two definitions of DCA being used and confused here. My thought when I read The Wealthy Barber was not that he was speaking of receiving a lump sum and then stretching it out in contributions over a period of time, but rather speaking of making regular contributions from every paycheck instead of either saving them up to make a contribution, or spending it and planning on investing a bonus or tax return or such at a later time (and then having to spend it on car repairs or other emergency and never doing any investments at all). From that perspective, both “camps” are actually saying the same thing – invest it when you have it, don’t wait and don’t try to market time.
I will add my voice to the crowd. I use DCA to invest a percentage of my pay every 2 weeks – and I beleive that this is a better option than doing it once a year when (or IF!) I have the funds saved. But when I ditched my high cost mutual funds for the couch potato – I held my breath and did it all at once.
Thanks to everyone for the comments, and sorry for being late to respond.
I’m big fan of David Chilton and the work he does as an investor advocate, so I hope I didn’t come across as critical of him personally. And it’s true that The Wealthy Barber does not discuss lump sums, but I do think it oversells the idea of DCA in the context of monthly contributions. Not a huge deal, but it probably contributed to the myths about DCA.
There’s no question that investing a portion of your paycheque every month is a marvelous strategy, but what makes it so powerful is the fact that you’re saving every month, period. Even if you’re using a savings account (same price every month) or a fund that goes up a little every month, it’s still a great idea, even though DCA would offer zero benefit in these cases.
In all cases, the key point is you invest money when you have it.
“Even if you’re using a savings account (same price every month) or a fund that goes up a little every month, it’s still a great idea, even though DCA would offer zero benefit in these cases.”
Actually, I think if you literally buy the same dollar amount every month (of a fund with non-zero volatility,) over a sufficiently long period of time your performance before costs should beat that of the fund. The reason is that on average you will be buying more shares when the price is high and fewer when the price is low.
That said, while kind of cool, there’s no tangible benefit to this fact. As you say, investing funds immediately is the optimal strategy regardless, so any theoretical benefit of DCA over time is really irrelevant. (Although I have found it to be a handy tool when ‘selling’ indexing to someone who likes the idea of beating the market. :) )
See my write up on this very topic at
http://www.statisticool.com/DollarCostAveragingVsLumpSum.htm
In the end, “it depends”, on market conditions, as well as costs, and some other factors.
Cheers,
Justin
I think an interesting related study would be whether, given a set amount, would higher frequency purchases of smaller amounts at a slightly higher MER outperform less frequent larger purchases of a slightly lower MER fund.
As an example (I’m picking on these two just for the simple math): Say a person had $200 per paycheck to invest, and got paid bi-weekly. That means they could set up an automatic investment plan to buy Global Couch Potato option #3 (with the RBC bond fund swapped in for the TD) once a week (the RBC funds allow $25 minimum purchases, $25 X 4 = $100), or they could buy Global Couch Potato option #2 approximately once a month (TD funds require $100 minimum purchase per fund). GCP#2 is a lower MER, but with GCP#3 the money is getting invested “sooner”.
In my work RRSP/DPSP, I have a limited fund list to choose from, but a set amount is taken each paycheque. While it’s currently going to three different funds and gets rebalanced every few months automatically, I’m tempted to start contributing everything going forward into the money market fund and once a year, “rebalance” and put the money into the other funds. This way my regular contributions continue, but I can simulate a one time addition to funds I actually prefer.
I always thought 0f DCA by definition (2) in the way of making consistent monthly purchases based on an ongoing income. Last year at this time I was in a position to chose a large lump sum purchase or DCA the money in and after reading a post by Larry Swedroe on the Bogleheads site about the misconception of ‘sell in May’ thinking I decided to go for it with the lump sum. I felt honoured to have Mr Swedroe actually PM me a couple of times to discuss it. Certainly, he stressed the psychological peril of the lump sum approach but in this case I was glad to have gone all in as the return on my VTI buy has been excellent and would have been much less with the DCA approach. It could have been different of course as taking the 70% side of any bet is far from a sure thing.
@GabyYYZ, why not put the contributions from each paycheque directly into the funds you prefer?
Because that’s essentially DCA. Building up the savings in something static or low-interest simulates piling up the cash, and once (or twice) a year, I move to the funds I want , simulating one-time contributions.
@Gaby: I’m afraid you’ve got it backwards. Remember, the idea is to invest money as soon as you have it, rather than to let it languish in cash. If you save up a year’s worth of contributions, you shouldn’t think of it as a lump sum you received on December 31. It’s actually 12 lump sums you received during the year, all of which you left uninvested for a period of one to twelve months.
This is why there is no logical inconsistency to recommending a monthly investment plan (what has been called DCA definition #2 in this thread) while also recommending an inheritance or other lump sum be invested all at once.
I think its important to stress psychological effects here. Young people are often told to invest most if not all of their savings in stocks because their time horizon is so long. This may be the optimal theoretical path but I would argue this has long run potentially very damaging effects because if equity markets are initially in a bearish phase the saving behaviour is being negatively reinforced. The attitude toward taking risk may be permanently damaged. Better advice may be for the young to save periodically (“paying yourself first” as Chilton puts it) tilting toward safer investments, to get ones feet wet so to speak, and shift to higher risk when it is clear saving has its rewards.
There still is the potential for lump sums to have negative returns through bad luck in timing the investment, so periodically investing in a regular schedule may result in higher savings rates overall.
Final point: a lump sum investment of $10,000, from say an inheritance, to a person of average income has a certain risk reward profile that is I would imagine is psychologically different than a $100,000 or multi six figures lump sum, because the eventual loss or gain potential will always be compared to income levels.
. . . or the answer is neither DCA or lump sum, it’s pay off the 19% interest credit card debt . . .
Thanks Dan and everyone else…I stand duly corrected. :)
@Andrew: Sounds like Rick Ferri’s “Flight Path”: http://www.forbes.com/sites/rickferri/2012/10/29/the-flight-path-approach-to-age-based-asset-allocation/
(As opposed to the more common “Glide Path”, of which “age in bonds” is the most common variant.)
It would be really interesting to see a study that measures the actual results of real investors, comparing those who lump sum to those who DCA. That would be a very difficult study to pull off, but I wonder if for the average investor DCA would actually be a better strategy simply for the psychological benefits. Clearly lump sum is a technically superior strategy, but we all know that psychology can have just as much if not more to do with investing.
Great post! You learn something new everyday! This makes so much sense, but I’ve never really questioned DCA before. Thanks for such a well researched an informative post.
@Justin Smith
Really liked this – would suggest adding interest gained from the lump sum, while it is not invested as this may make it a bit more realistic.
Below is how i did this, where e4 is a cell where i typed in interest… could make this a rand too between 0-5% if you wish …
=IF(MOD(C8,30)=0,(1-C8/365)*$E$4,0)
The report concluded that if an investor expects such trends to continue and is satisfied with their target asset allocations and the risk/return characteristics of their strategy they should invest the lump sum immediately to gain exposure to the markets as soon as possible. If the investor is primarily concerned with minimizing downside risk and potential feelings of regret because they invested a lump sum just before a market downturn, then lump-sum investing is not the method for them.
@CCP:
For complex reasons, regarding the way I manage my professional corporation, I am not able to invest regularly through the year. I usually have a big amount of money to invest between January and June and nothing to invest between June and December. Do you see any problems investing money as soon as it become available to invest during the first half of the year (2-4 transactions) even if then I have nothing left to invest for the second half of the year?
@Jas: I don’t think there is ever any harm simply investing cash whenever you have it.
@CCP: It makes sense, thanks for the advice !
If you’re freaking out, invest half now and half a couple of months later.
I’m curious if there is any research comparing DCA and lump sum investing relative to whether the market is considered “overvalued” (eg. http://www.currentmarketvaluation.com/). I know that moves into timing the market, but one would expect the DCA vs lump sum comparison to change depending on such metrics