What Is Dollar-Cost Averaging And Why Is It Bunk?

Dollar cost averaging (DCA) is a method of investing whereby you spend a fixed amount on a stock per month, regardless of price.

According to its proponents, which include School House Rock, this reduces your risk because you buy less of a stock when it’s high, and more when it’s low.

DCA was debunked, using scary equations, in,”A Note On The Subotimality of Dollar-Cost Averaging As An Investment Policy” published in the Journal of Financial and Quantitative Analysis in 1979 *. Comparing DCA to lump-sum, George Constantindies wrote, “Both investors face the same prospects irrespective of the composition of their endowment, and any claims of gambles on temporarily overpriced or underpriced prices are simply fallacious.”

Statistically, DCA underperforms lump-sum investing (aka, putting a buncha money in at once), as this calculator shows. An exercise by MSN Money found that DCA even underperforms putting in money at random intervals. Their rate of return was 9.8% for DCA, 10.5% by happenstance, and 11.7% with a lump-sum.

However, there is a method that fares even worse than DCA: not investing at all. — BEN POPKEN


* From A Note On The Subotimality of Dollar-Cost Averaging As An Investment Policy:

Where, then, does the intuitive rational of DCA fail? Its rationale is that the investor replaces one major gamble on a temporary shift of prices by a number of smaller gambles and thus diversifies risk. The fault of this argument is misrepresentation of the state of the world, before a decision is made. DCA implies than an investor with all his endowment in asset A is in some way different from an investor with all his endowment in asset B, but otherwise identical. DCA ignores the simple fact that the latter investor may costlessly convert his endowment from asset A to asset B before he considers the optimal investment decision. Both investors face the same prospects irrespective of the composition of their endowment, and any claims of gambles on temporarily overpriced or underpriced prices are simply fallacious. We do not claim that the investor should not incorporate in his optimal decision his beliefs on whether assets are overpriced or underpriced. What we do claim is that these beliefs lead to the same optimal portfolio irrespective of the composition of initial wealth.

Comments

Edit Your Comment

  1. Lewis says:

    Wow, thank you for posting this – I had long embraced DCA as I thought it was a widely-accepted strategy (as market strategies go) to mitigate downside risk.

    But I guess the numbers speak for themselves!

  2. skittlbrau says:

    I think a benefit of dollar cost averaging is in perceptions – it doesn’t hurt as bad to have $50 or $100 leave your checking account every month as opposed to seeing $3000 or $4000 leave at once.

    DCA makes investing easier for some people, and that is really the first step to building wealth.

  3. Skiffer says:

    @LewisNYC: Don’t put too much faith in the MSN numbers – I don’t think they had enough sample trials, and the uncertainty in the results is too big. If they had enough samples, then the DCA returns should be equal to the “happenstance” random returns – because in the long run, DCA at a set interval should approach random times.

    @baa: DCA is “lazy” investing (not necesarily in a bad way). It does mitigate risk, and works great for monthly 401K contributions, etc. But if you do already have the money available and sitting there – never use DCA. If you’re investing that money, you should be researching the investment enough that you can identify a local low point to buy at. DCA will probly also incur more cost in trading commissions. It’s your money, don’t invest without researching, and in doing so, you can figure out when the right time to buy is that will save you a couple percent…

  4. douglips says:

    Of course if you invest a lump sum it does better than dollar cost averaging. The main reason why DCA is a good idea for most people is that most people do not have a lump sum.

    I keep trying to get my employer to write me a check for my full salary on January 1 of each year, but he never goes for it.

  5. Juancho says:

    @baa: and @Skiffer: Right. It’s easier to “pay yourself first” with money out of your paycheck to do DCA.

    Similar example? I don’t withold anything from my paycheck. People tell me I’m crazy, letting the IRS have interest on my money. These are the same people who have to scramble when they owe every April.

    Look, I admit I’m not rigorous or organized enough to handle the extra cash that might come in my paycheck every two weeks if I withheld. I don’t think the average American is.

  6. TSS says:

    This is nitpicky, but the money Schoolhouse Rocks came out after the originals. The originals were in the 1970′s, the money ones were in the 1990′s.

  7. Timewalker says:

    But, Suze Orman told me I should rely on Dollar Cost Advertising….. My entire faith in investing has been shattered. I’m gonna go buy some lottery tickets.

  8. skittlbrau says:

    @Skiffer:

    Right – if you have the money, by all means do something with it.

    I’m just saying that sometimes its easier to part with $100 than think you have to come up with $1000 or more to invest.

    Lazy sure beats not doing it at all.

  9. Ass_Cobra says:

    There was a simple piece of paper hanging over Paul Tudor Jones’ desk that summed all of this up, it read “Losers Average Losers” which is basically a short way of saying, buying a stock programatically without considering what your other investment opportunities is is a sure way to lose money.

    Link the the picture and brief explanation below:

    http://www.turtletrader.com/losers-average-losers.html

  10. zolielo says:

    DCA is a joke. Programmable, trigger buying, can work but the econometrics is far more advanced. Plus everyone with “a system” that works keeps it proprietary.

    I gave a tip about GARCH and ARCH based buying strategies which do lead to moderate gains. However, such data, analysis, and interpretation is beyond most.

    For most I would just suggest value investing using simple fundamental analysis to determine underpriced stock.

  11. notallcompaniesarebad says:

    Don’t confuse regular investments (i.e. investing a set amount at various times) with DCA. The reason DCA doesn’t work is that the alternative is to keep funds in a suboptimal investment while you dribble them into your main investment. Regular investing (say, investing $100 a week) is often the result of the way we make money: over time. We don’t get our yearly pay on Jan 1, we get it in little chunks over time. Investing it as you earn it is not DCA (although it looks like it). It is actually many lump sum investments. If you save up to get a big wad of cash, and then invest that cash in regular intervals, that is DCA. But you can’t lump sum invest money you don’t have.

  12. leejames says:

    I wonder how much those percentages change over time? If I had to guess, the DCA ones would probably be slightly less volatile overall than the the happenstance or lump sum ones. Sure, lump sum was 2 ticks higher when they did the study, but would it be the same if they repeated it now?

    I’m just wondering, because is it really so bad to tell people to consistently save?

  13. Chicago7 says:

    If they are going to compare with lump sum investing, shouldn’t they invest the lump sum in the MIDDLE of the year? Of course, you are going to earn more if you invest $3000 at the start of the year versus investing $750 quarterly. That’s the miracle of compound interest, not a better investment scheme.

    If they wanted to really check DCA against lump sum, they should check it over a long period of time to smooth out that effect.

  14. notallcompaniesarebad says:

    @lee-reamsnyder: “is it really so bad to tell people to consistently save?”

    That’s the thing people need to get over. This is not a critique of savings, it is a critique of waiting to invest in order to do it in little chunks. If you can afford to put away $100 a week for a year, it looks very similar to the person who had $5200 at the beginning of the year and invested it in 52 weekly installments. Looks similar, but it isn’t: you didn’t have the choice of investing at the beginning of the year, as where the other guy did.

  15. notallcompaniesarebad says:

    @Chicago7: “If they are going to compare with lump sum investing, shouldn’t they invest the lump sum in the MIDDLE of the year? Of course, you are going to earn more if you invest $3000 at the start of the year versus investing $750 quarterly. That’s the miracle of compound interest, not a better investment scheme.”

    No. The alternative is withholding your investment and dribbling it in over time. The “findings” here are not anything more breathtaking than your last sentence in the quote above. To DCA money before you would have otherwise been able to invest it in a lump sum is an incorrect comparison.

  16. notallcompaniesarebad says:

    @douglips: “The main reason why DCA is a good idea for most people is that most people do not have a lump sum.

    I keep trying to get my employer to write me a check for my full salary on January 1 of each year, but he never goes for it.”

    You’re right (I didn’t even read your comment by the time i had referred to a january 1 lump sum salary!). Too many people confuse DCA as defined in these analyses with the apparent DCA that comes from earning a steady paycheck, paying (relatively) steady life expenses and investing the difference.

  17. Chicago7 says:

    @notallcompaniesarebad:

    Well, if the market went DOWN, you would lose more money in lump sum deposited at the start of the year.

  18. bedofnails says:

    I think this has changed somewhat with the advent of the 6% internet savings account. I think the liquidity and high return (4,5,6%) of these accounts mitigates the incremental gains made by DCA.

    Unless somehow one were to invest in incredible value picks with each DCA “dump”, I would find it hard to believe that an individual placing $100 a week into an HSBC online savings, then funding a Sharebuilder account once a year with $5200, repeat, repeat, etc – would lag an individual jamming $400 a month into a brokerage account.

    Sharebuilder has a $12 a month plan that allows 6 free trades, but you can turn this on and off as you please. So right away, the DCA guy is down 3% each month ($12/$400), while the other guy is already up 6% as his money relaxes in a high yield savings, meanwhile the investor with $5200 would only pay $12 a year.

    Years ago, this was probably the same principle with CD’s; but to achieve those sorts of rates, you would loose a lot of liquidity, thus preventing the above.

  19. notallcompaniesarebad says:

    @Chicago7: Depending on the timing and the nature of the downward trend, yes. But if you believe in a form of efficient markets, you don’t know it’s going to go down in advance (otherwise, why invest anything?), and therefore this analysis is a red herring.

  20. notallcompaniesarebad says:

    Okay, I ran some numbers thusly:
    Stock returns based on dividend adjusted S&P 500 figures since 1950, sourced from Yahoo finance. The hypothetical investor has $1200 to spend, and he can either invest it all at once, or in installments equal to $100 plus the interest earned over the prior month (all figures calculated by month). The returns are calculated through the beginning of May on an IRR basis. There are no transaction fees.

    End result: Lump sump beats DCA by a few basis points per year at interest rates below 6%. I would imagine that fees would eat up a good chunk of the DCA’s returns, similar to what bedofnails says above.

  21. notallcompaniesarebad says:

    @bedofnails: “Unless somehow one were to invest in incredible value picks with each DCA “dump”, I would find it hard to believe that an individual placing $100 a week into an HSBC online savings, then funding a Sharebuilder account once a year with $5200, repeat, repeat, etc – would lag an individual jamming $400 a month into a brokerage account.”
    One of the ways around this would be to use a no-load fund. That’s what I do, and thus fees are not a concern for me.

  22. swalve says:

    notall- you’re right in that dollar cost averaging is different from planned, regular contributions.

    If I buy 100 shares at 10, and then six months later it goes down to 5, I might buy another 100 shares to lower my average cost to $7.50. That, in essence is dollar cost averaging. Maybe six months later the stock goes back to 10 and you cash out. The first way, you lose due to transaction costs. The second way, you win or at least break even.

    It’s about having a good stock that you believe will ultimately perform, and buying it when it’s a bargain, and *whenever* it’s a bargain. Keep your average cost as low as possible keeps your profit as high as possible. If you’re a retail store, you don’t buy all your VCRs in January and then hope to sell them. No, you buy your supply when it’s low and hope to sell it to your customers when its higher. Stocks are no different.

  23. SCO says:

    Dollar-cost averaging is a psychological trick that helps keep people from being whipsawed in and out of the stock market. Since jumping in and out is the reason most investors never see returns even close to the major stock averages, DCA can be a valuable tool. Emotions are the investor’s worst enemy, but we all fall victim to them to some degree.

  24. tz says:

    How has DCA worked with the NASDAQ since 2000 when it was 5000+?

    How well did DCA work with ENE, WCOM, pets.com and many others?