Dollar Cost Averaging: A Poor Way To Reduce Risk?

Dollar Cost Averaging (DCA) involves investing a fixed amount at a regular interval. Lump-Sum Investing (LSI) involves putting in all the money you have available to invest at once. These are not mutually exclusive! If you are investing a portion of your paycheck every month, you are both Dollar Cost Averaging and Lump Sum Investing. The following is not about such habitual savings.

However, a different situation arises if you have a larger amount of money. Maybe you received an inheritance, an early retirement payout, or you just sold your house. Do you invest the entire amount immediately, or buy a little at a time? Due to the overall upward trend of the markets, lump-sum investing outperforms DCA about 2/3rd of the time. The argument then, is that DCA is a risk-reduction mechanism; You get less performance, but also less exposure to those ups and downs. But is DCA the best way to lower risk?

This question was examined in this academic paper titled Nobody Gains from Dollar Cost Averaging by Knight and Mandell. Here’s a sample of their results. Let’s say you have $100,000 to invest, and you want to achieve a portfolio of 90% stocks (modeled as the S&P 500) and 10% bonds (T-Bills). But that sounds risky to you. You decide to instead invest gradually over 10 years, every month putting a little bit more in, until you finally put $90,000 into stocks.

But what if you instead put everything at once into 50% stocks and 50% bonds, and kept those 50/50 proportions for the entire 10 years instead? That would also reduce your risk. You may be surprised to find out that historically the 50/50 rebalanced portfolio actually had the same amount of volatility than the 90/10 dollar cost averaged portfolio, but with a higher average return (8.37% vs. 8.05%).

So if you are keeping money out of the market because you don’t want to be exposed to a crash, it may simply be better to invest in a less aggressive investment mix. But if you are already regularly investing what you can each month, keep it up! This doesn’t apply to you.

For more academic papers on why DCA is not the best way to reduce risk, see this AltruistFA reading list. Thanks to reader Craig for sending me this article.

For my overall thoughts on investing for beginners, please see my Rough Guide to Investing.

Comments

  1. I guess for many people DCA is not just to reduce risk, but a little hard to put down say, $10,000, at any time of the year. Spreading it throughout year is much easier to do. Of course, if one already has $100,000 ready to invest, it doesn’t make too much sense to invest a little at a time for 10 years. What’s the chance that the market will be worst 10 years later? As for the performance difference, I don’t feel it is a direct result of the 50/50 or 90/10 allocation, though I didn’t read the paper. The market change also in 10 years can generate that kind of difference. Actually, over long period of time, bond is not safer than stock (unless the bond is held to maturity).

  2. Yes, to reiterate, if you are already investing what you can, then you should keep doing it as there is no other better way to do it. This is specifically for those people who do have larger sums.

    The idea is not about absolute average returns, it’s about optimizing your risk-to-reward ratio. In that case, these papers suggest that DCA seems to be a poor way to reduce risk given the alternatives.

    There is also a more numerical approach in the paper, which is not based on historical numbers. But it was really hard to summarize that ;)

  3. When I saw the title of this post, I thought to myself, “Oh no, here we go again. Another rehash of an age-old debate.” But you really put forth a different perspective on the topic. As a fan of DCA, I can now truthfully say that I am investing lump sum, but 50% in stocks and 50% in cash. =)

  4. Yes, I love the altruistFA reading room. It’s a step above the chaff and a step below finanical journals.

  5. Not sure I agree.

    If you were to inheirit $100,000 and wanted to invest it all slowly, it’s not as though the chunk of money you’re not investing isn’t making any money.

    The first year you put $10k in the market, and the rest ($90k) in a CD or savings account. Sure, you’re not making the returns that you could make in the market, but you’d make at least 5% wth no risk.

    In your example, you’re taking advantage of compounding ALL your money over 10 years at market rate returns. The guy who invests his money in chunks over 10 years still compounds his money over the same time frame, just at an averaged lower rate (assuming the market will return more than the bank).

    If you’re looking for risk mitigation, the bank is a sure thing.

    If you’re looking for maximum return and don’t care about the down side… put all your chips on black and spin the wheel!

    -Grant

  6. I use DCA when purchasing stocks because you’ll almost always pay less when the trades are spread over the course of a week.

    With lump sum investing, you have no way to hedge any risk, especially if the stock takes a short term dip after you make a trade.

    For example, I purchases shares of EFJI at around $7, making up the 1st part of the trades. About 2 days later, the company reported negative earnings guidance, and the stock fell almost $2 in one day.

    If I had invested a lump sum, I would’ve been down 25% in less than 2 days. But since I only bought 1/4 of my position, I got the remaining 3/4 at a huge discount. A week later, the stock returned to $6.75.

    So I use DCA to reduce risk because you never know when bad news will come. By spreading your trades out over the course of a week/month, you can benefit from short-term losses, and close your position at a nice discount.

  7. Grant, TJP:

    You guys are missing the point of the study & the post. Read it again, a little slower. If you still disagree, hire a CFP.

    TJP – you put in a lump sum in a single company? Unless that’s your play/gambling money, stop now, hire a CFP.

    To summarize some key points: markets trend upward. Invest the most that you can at any one time while keeping transaction costs in mind. Market timing does not work – you will lose.

    -Wes

  8. Thanks for this post. I am looing to invest some money this year and your info will help.

  9. I think it depends on your goals and what you will be investing in.

    If the market is fairly valued or under valued and you are investing in a diversified portfolio – absolutely Lump Sum.

    If the market is highly valued, DCA may have merits if there is a correction. (Investing in the Oil Sector last year would be a good example of this)

    If you are investing in individual stocks with high P/Es then DCA can be a very effective risk reduction tool. Now the example of over 10 years period of time is extreme, DCA to me is over a small period of time… The market goes up 7 out of 10 years and averages much higher than T-bills, so of course on that time frame lump sum is going to beat DCA. A comparison over 3 years would be more interesting.

    I think an interesing question to look at is your yearly IRA contribution and whether DCA or Lump Sum (in January) on that 1.3 year time frame makes much of a difference.

  10. As another person has mentioned, why would you not keep the rest in savings. This research is only to prove a point. Nobody would have 10k and DCA for 10 years.
    If you do have a lump sum then DCA could be used in one year, not 10 years. It has nothing to do with risk or return, but how long the money is in the market. With DCA for 10 years it’s like saying who would do better an 18 year old who started to invest in a roth IRA or a person who was 28. We all know the younger even lump sum would do better, because they were in the market longer. There’s a person at the diehard forum I know who really bashes DCA and all the responses are comparing 10 years or something that’s not realistic.

    Most people get paid biweeky, so they should invest either biweekly or monthly. If you get a lump sum, then I’d suggest DCA up to a year, if you believe it will lower your risk or if you are going to do a lump sum investing to buy an ETF over an index fund, especially if you’ll hold for the long-term

  11. The 10 year time frame and $100,000 example used in the study is not typical for most investors. So I agree with JT. Looking at making an IRA contribution in January v. over 1.3 years would be much more useful.

  12. If you are interested in IRA contribution Lump-Sum vs. spread across 1 year, please see the first link in the post. Most of the time, Lump-Sum will beat DCA. By doing DCA you will lower your expected long-term return, while also lowering your risk. How the exact risk/return ratio plots I am not sure.

    But as someone else pointed out, if you put $4k every January vs. $300 every month, that’s still DCA with wider interval. In the end, I expect it not to matter very much either way.

  13. If you plan to lump sum 100k in 1 purchase, consider lump sum in 2 purchases (a DCA lump sum) during a short period (say 2 consecutive days, weeks or months). Let’s assume stock price of X and Y on the two purchases. Which one would you prefer?
    a) lump sum in 1 purchase: 50% chance of paying price X and 50% chance of paying price Y
    b) lump sum in 2 purchses: paying average price of slightly lower than (X+Y)/2

    I would take b) for my 100k. The point of DCA is not about the performance or volativity of a portfolio “in average in 10 years”. It is used to reduce the risk for my portfolio on that particular purchase(s) of 100k.

  14. Thanks for the link to the article. It’s been cited many times in passing, but this is the first time I’ve been able to actually find the original source. I look forward to reading it.

    As an aside, probably the only time “most people” have a large lump sum to invest is right when they shouldn’t be doing it. For example, the late 90s’ bubble. Plenty of my coworkers suddenly found themselves with large amounts of money from their options but little understanding of investing. People were buying and shorting dot-coms left and right, with or without margin, with the money they got from their options. (It was a crazy time.)

    Even if they had just taken their money and made a straight lump-sum buy, I’d bet that most of the companies they would have bought would still be out of the money even today.

  15. My 2 cents:

    The study is interesting but doesn’t seem to mimic real investor decisions. You’d have to be one heck of a wimpy investor to be considering spreading $10k over 10 years.

    Where I actually do make the lump sum vs. DCA decision is year by year. If I’m optimistic for the year, I lump sum my Roth contribution. If I’m pessimistic, I DCA over the year to make myself feel better. This decision constitutes my only conscious form of market timing, and I don’t think any of the passive investing gurus would disagree with it. If my gut is wrong for a particular year, at least I still invested.

  16. I agree with a comment said a few times already… 10 years is simply WAY too long. In fact, it’s such a oddly long amount of time that it makes me suspicious that the books might be cooked! However, having not read the article, I won’t go that far.

    Needless to say, I think most people agree, a better study would have been to invest this over 6 months, not 10 years.

    The idea in DCA is to avoid the super high volatility (daily or weekly up’s and down’s)… not the net gain of 10 years worth of investing!

  17. If you read the paper, you will see that it is assumed that all your money is invested in the risk-free asset (T-bills or CDs for example) and then is moved into the risky asset (stocks) either all at once or over a period of time (DCA).

    So it is not true that they have 90% of the money earning nothing for the first year, etc.

    In fact their results show that analytically DCA is worse in all cases. It doesn’t matter what the investment horizon or the amount of money is. It applies equally well to investing $4000 into your IRA all on Jan. 1st vs. over the course of the year. The difference over 1 year of returns won’t be very significant but it will be if you repeat this every year.

  18. Thanks for posting up the paper, Jonathan. I think for 2007 I will invest a lump sum in my Roth IRA.

  19. RE: Andy

    That’s the market timing I was speaking of.

    My Roth has International Value and Small Cap Value, both of which have performed incredibly. I’m not optimistic that these two funds will continue such a pace. So instead of making my $5k contribution now, it’s in the money market and planned to move over gradually to the equity funds.

    But that’s just this year. Last year I was pretty darn optimistic and bought all my Roth shares in January. In June, I admit I thought I had goofed, but by the end of the year I was quite happy.

    Yeah, 2/3 of the time the lump sum will win. But if I’m really worried at the start of a particular year about my sinking feeling that my fund is going down, and DCA (or actually VA) curtails my worry and helps me invest, then it’s for the best.

  20. My take on reading everybodies comments:

    1) From the study: “DCA strategy is also substantially inferior to the Buy and Hold Strategy” [and the Rebalancing Strategy] LS wins. Read the article.

    2) Looking at whether investing every January or 4 times per month is irrelevant over the long term. It’s still DCA.

    3) The point is: invest as much as you can as soon as you can. Don’t wait, don’t time the market, don’t guess on your gut. If your gut/instinct/market timing ability was that good you wouldn’t be reading this.

    4) W.H your x+y/2 example is not correct since markets trend upward. Test your example over any period and you will see that (t1 +t2)/2 will most often be lower than simply investing all at once since probability says that t2 will be higher than t1. That’s the whole reason why LS beats DCA (t1 + t2)/2 IS DCA.

    Hey – but do what helps you sleep at night ! ;)

    -Wes

  21. I’d like to see more push for people to just START and less talk about these intricacies — though I guess there’s room for both. Most people just won’t start…

  22. Wes – I think the real question is “did you read my comments?”

    I don’t use lump sum investing. I always split up my buys into 4 separate trades over the course of the week.

    Maybe my comment was confusing?

  23. Jesse, you are so so right! Time is MOST important (actually, the lump sum being better just re-enforces that). But hopefully, if we’re reading PF blogs, we’ve all started…

    I think this comment was made already, but it needs to be stressed. DCA isn’t about improving your gains… its about reducing risk_

    If you LUMP SUM you have some chance of timing it right and some chance of timing it poorly. *statistically* that should be a wash.

    If you DCA, you might time it right and you might time it wrong. Again, it should wash out statistically. So, the only difference is that the money is in there for LESS time. This is why statistically, LS always wins.

    However, I keep saying “statistically” on purpose. What I mean by statistically is “rolling the dice a million times.” Well, we don’t do that. We lump sum invest once a year (maybe 40 times in our life). The number of times you time the market well probably isnt equal to the number of times you time it poorly. You might do better… you might to worse. The only thing for sure is that it probably won’t wash simply because you aren’t doing it enough times. Hence, you’re RISK is higher.

  24. If you contribute monthly or bimonthly to a 401K are you not dollar cost averaging? I would submit that about 95% of those who contribute $ to a 401 program do it monthly or bi monthly. Further more I would summise that the benefit of investing in this fashion (DCA) long term far out weighs the negative sum of not investing at all. Can one save $ long term by $ cost averaging money into a savings / money market vehicle and then spreading that money out in equities in lump sums yes. However, you are still using DCA just on a longer horizon. All investing utilizes DCA, doesn’t it…….?

  25. Mike -
    If you’re poor uncle Leo passes away and leaves you $100k, what would you do with it? That is the question.

    Is it better to invest it all at once in the market, or invest it a little at a time while keeping the non-invested portion in a risk-free asset?

    The conclusion of the paper is that you should make a lump-sum investment.

    IMHO, this can be extropolated to say that invest as much as you can as soon as you can. Short term gyrations in the market do not matter in the long run.

    -Wes

  26. For people who are more concerned with simply starting their savings instead of talking about this stuff, please refer to my Rough Guide To Investing posts.

  27. I agree 100% with what Wes says. However, I also agree 100% with what Miller says.

    I think people are disagreeing about apples and oranges on this thread. Or tangerines and grapefruits. :)

    The article, and many studies before it, says lump sum wins over DCA most of the time. A lot of people, myself included, have noted, “yes, but without the peace of mind of DCA-ing, a lot of investing wouldn’t have occured in the first place.” That’s all.

    If you’ve got an iron stomach for risk, of course it won’t make sense to you. You can just look at the findings of studies like this one and feel comfortable lump-summing no matter what.

    But if at any particular moment in your investing career you’re hesitant to continue investing per your asset allocation because, oh I don’t know, “Real Estate looks like a huge bubble right now, and I’m nervous about allocating any of Uncle Leo’s money to the REIT fund,” DCA-ing can help those with stomach’s made of balsa wood and duct tape actually feel comfortable sticking to their AAP’s despite themselves. This issue is more of a “sidebar” on the psychology of investing, not a “rebuttal” against the findings of the study.

  28. JimmyDaGeek says:

    As usual, the academic paper used some apparently arbitrary 20-year period of stock data and another arbitrary 10-year rolling period to make their calculations. I wonder how much bias is built into the data? None of that was explored in the paper. 99% of the paper was above my head with terms like coefficient of risk aversion and utility that had a specific academic meaning. The paper also used some arbitrary measures to prove their point. What is not mentioned is that if I were to have put my chunk in at the market top and 10 years later, not have gotten anywhere, I would be royally pissed, no matter what a paper said. Many practical studies have shown that DCA does work, in the long run. Comparisons have been made between putting your money in at the top of a period vs the bottom of a period. While there is a difference, it is not as great as you might think.

  29. In my opinion, and owing to current market conditions, dollar cost averaging or buying shares as they fall away because they seem “cheap”, is not smart investing for the simple reason that most investors do not know where the bottom is and to buy without knowing is speculating.

  30. Andy,

    I read a lot of “I just wanna wait until the market settles down and we know where the bottom is before I start investing again” posts on various blogs and boards.

    I agree completely that investors don’t know where the bottom is. However, I personally see that as an argument that ceasing investing one would otherwise do until we know where the bottom was is speculating.

    If you graph a total market fund during the last bear market in 2000-03, you see several dead cat bounces over two years before it finally started climbing steadily. By the time the market by consensus “settles down” and the consensus “bottom” is known to have occured, prices will be up considerably.

    A bear market like this is precisely where I feel dollar cost averaging has value to investors. A DCA investor may not know where the bottom is, but until we hit it and up to some time after we’ve bounced back, he or she will have paid the average market price for their shares over that period. In contrast, someone who waited to jump back in until it was “clear” a bottom was reached and the upturn was clearly not just another “dead cat bounce” will have paid more for their shares.

    Again, which investor is speculating?

Trackbacks

  1. [...] The Frugal Duchess likes Ugly Betty, a frugal character on an ABC Family show. MyMoneyBlog says dollar cost averaging is a poor way to reduce risk. [...]

  2. Links for 01/08/2007 - Investor Trip says:

    [...] Dollar Cost Averaging: A Poor way to Reduce Risk? [...]

  3. [...] Over the week, I read two posts on whether to DCA or lump-sum (LS) investing. Jonathan at MyMoneyBlog cited a research paper that DCA is not only a poor way to reduce risk, but also inferior in long-term return. Super Saver contributed to the full limit in IRA in January and pointed me to an article which also says LS beats DCA. All these articles/posts contradicted to what I consider a good investment strategy. So is DCA really bad for long-term investment? If so, how bad is it? [...]

  4. [...] Dollar Cost Averaging: A Poor Way To Reduce Risk?, My Money Blog [...]

Speak Your Mind

*