A good discussion came up in my previous post on my IRAs – Should I invest the $4,000 earmarked for my 2006 Roth IRA contribution all at once in January, or dollar-cost-average it over the entire year? Dollar Cost Averaging (DCA) involves investing a fixed amount at a regular interval. For my $4,000 example, I could invest $335 a month, every month, for 12 months. The idea is to buy more shares when the price is low, and and less shares when the prices are high. Sounds good, right? Maybe.
This question actually came up last year, but I didn’t research it very much. My own thoughts were that because the markets trend upwards overall, if you are investing for a long-term period you should get your money in as soon as possible. Sure, you might run into a huge drop, but you could just as easily (in fact more easily) miss a huge rise. But this is too hand-wavy, as scientists would say. I want numbers. So I found some.
Now, wouldn’t it be nice to have a comparison of DCA vs. lump-sum investing for the past 50+ years? We could compare investing $10,000 all at once in January of 19xx, versus using DCA equally over all 12 months of that year. Wouldn’t it be even nicer if we could take into account that any money not used be put in a high-yield interest bearing account?
Well MoneyChimp did just that all the way back to 1950. The result? I used 4.25% rate for bank interest, and over 60% of the time, lump-sum investing beat dollar cost averaging. This result of DCA losing out about 2/3rds of the time is supported by historical back-testing from 1926 in this article from the Financial Planning Association: ‘Lump Sum Beats Dollar-Cost Averaging’. (Just read the conclusion if you get bored.)
Of course, past performance does not guarantee future results. And DCA would smooth things out if your time frame is really short. I think everyone should consider the facts above and make their own decision. But I bet with the odds, and the odds are that I should invest it as a lump-sum.
However, if you don’t already have the money to max out your IRA, then by all means dollar-cost average!! In essence, you’re still getting your money in there as soon as your can.
The Best Credit Card Bonus Offers – October 2024
Big List of Free Stocks from Brokerage Apps
Best Interest Rates on Cash - October 2024
Free Credit Scores x 3 + Free Credit Monitoring
Best No Fee 0% APR Balance Transfer Offers
Little-Known Cellular Data Plans That Can Save Big Money
How To Haggle Your Cable or Direct TV Bill
Big List of Free Consumer Data Reports (Credit, Rent, Work)
I think you are better off making your IRA contribution in January. It should be for your 2005 contribution,not 2006.
It seems a mathematical truism that LS should beat or match but never underperform DCA.
If one has a pile of cash, of course you should invest it otherwise it’s deflating. However, investing from your paycheck over time, DCA is a great by-product.
Wes
Ok, so I spoke too soon, it’s not a truism. DCA will beat LS if there are enough negative returns early on in the first half of the investment period. I guess if you know you’re heading into a depression, use DCA.
Wes
Interesting articles. Of course, as you noted, lump sum investing is predicated on the the fact that one actually has a lump sum to invest. Since you do have a lump sum to invest why not invest in a fund with a lower expense ratio such as the Vanguard Total Market Index or an ETF? I looked at VTTHX and VTIVX (your funds) and noticed they each have a 1.20 % expense ratio! Compare that to .19% for the Vanguard total market index (VTSMX) or even .08% for some of the VIPERS ETFs. Granted, some of these funds are more expensive to buy; but, if you are earing 7% on a fund, minus the 1.20% expense rate, minus approximately 3% for inflation (keeping in mind the CPI does not include things like gas, groceries, or medical services) you are looking at a much less attractive net gain. Over time that expense ratio can really eat into gains, especially when people like us are moving money around just because Emigrant Direct or some other bank upped its interest rate by a quarter percent.
The all-in-January plan is really just dollar cost averaging with less granularity. In my mind, DCA is more of a risk-avoidance strategy that a strategy for maximizing returns. The more installments you break it into, the less risk, so I’m not terribly surprised that this hurts your return. And yes, I’ve always been of the same school of thought as you… If the market trend is upward then, all else being equal, you should get your money in the market ASAP, as it will go up more likely than down.
Or, following the Ben Stein and Phil DeMuth book “Yes, You Can Time the Market,” you could put the money in stocks (S&P 500) when they are cheap and in a high-yield interest bearing account (or bonds) when stocks are expensive. They use the 15-year moving average (time for one recession) relative to a few different indicators to determine the broad based value of stocks. A very interesting concept which they back up with tons of data.
http://yesyoucantimethemarket.com
I like the p/e and I am going to do this in the 401k starting in January.
It makes perfect sense that lump sum will beat DCA most of the time since you take more risk when investing the lump sum. It’s the same risk/reward concept as investing in an individual security vs. a mutual fund.
guy incognito – where are you getting your numbers?? VTTHX and VTIVX both have a low total expense ratio of 0.21%, including all underlying expenses. Try Morningstar.
nickel – that’s a good way of thinking about it. But the question is how much extra risk for how much extra return? What is most efficient?
trip – yes, I have those Ben Stein books in hand, and am reading one now 😉
guy incognito, the expense rations for VTTHX and VTIVX are not 1.20%, where are you getting that information?
Yes, I agree. DCA works better when the market is all but certain. I especially agree with the last sentence of the article, with DCA, I do feel good even when the market is down.
to the author: how much is your salary, you’re 27 and you save 4400 per month?! how is that possible?
When did I save 4400 a month? I think I would remember that 😉
I’m putting all my money in at once, cause I don’t want to think about it.
Forgive my laziness to read the details, but in the DCA versus LS study, what was the timing invovled? Meaning, did the LS occur at the same time the DCA *started* or mid-way or the end…? From the above comments, it sounds like it was the beginning (and hence, merely supports “get your money in ASAP”).
Ironically, if all my above assumptions are correct, it seems like you can get the best of both worlds in some sense. I don’t have thousands of dollars to invest at the beginning of every year, however, I do have a few hundred. So, I invest monthly, right? Well, that is (a) getting the money in ASAP and (b) spreading it out over time and hence diversifying. Best of both worlds???
Jonathan,
Little off topic, but have you heard about BofA’s “keep the change” offer? If you haven’t look it up and tell me what you think.
I am with Miller on this one. It sounds like you are assuming DCA over the following 12 months or lump sum now. In that case, yes, I would assume most of the time that lump sum now is better. When the question was first brought up yesterday in the other post, I assumed that you should have been DCA’ing over the past 12 months…not lump summing now.
The keep the change program was mentioned earlier here:
https://www.mymoneyblog.com/archives/2005/10/some_updates.html
fivecent is right. True lump sum would be to put money in one time and thats it.
If you make an annual max-out contribution to your IRA, you still are DCAing. You’re just DCAing on an annual instead of monthly basis.
I am going to lump sum my investment also.
I think DCA works if you are buying individual stocks, which have high volatility and can fall quite a bit. If you’re investing in an index fund, then it is less likely to fall a lot in a year to open up opportunities to buy it cheaper.
I modified moneychimp’s script to print total for all years on the page. I summed the difference with Excel. In order for DCA to match Lump Sum, your bank account must be offering ~7.99% rate.
One more vote for lump summing every year.
Contribute $4000 to a IRA MM fund in Jan. Then move $335 to a stock/balanced fund every month. That’s what I’m doing
a have always wondered about this…
isnt december and january historically the best time for the market? therefore putting a lump sum in late january kinda dumb? wouldnt putting it in in october, the historically worst month for the market be smarter?
thats not too mathmatical, but it makes sense to me.
But you miss out on 10 months of potential gain. Dont you?
or 10 months of loss.
If you’re contributing to an IRA annually, then you are already dollar-cost-averaging, the installments are a year apart and made over decades rather than say, monthly over the course of a year.
If you can, make your IRA contributions every year, as early as possible. January 2 would be just fine.
I’m a huge fan of DCA because it’s really the only way to go with 401k’s etc. If I come in to a lump sum of money, I’ll most definitely do a lot of soul searching and analysis of where the market is at and will most likely move “all in” when it feels right. Of course I’ll be putting it in a low expense index fund that mimicks the market.
Hazzard
DCA is a risk avoidance strategy. If you calculate the Sharpe ratios of both methods using reasonable assumptions, the ratio is lower for DCA vs lump-sum. DCA reduces risk, but it reduces return even more.
As has been said, there is no difference between dollar cost averaging monthly, or dollar cost averaging yearly. Every time you contribute on your set date, you could be buying at a higher or lower price. Unless you are a mack-daddy market timer, it works out all the same.
Stats articles that leave out the important bits of variance, p-values, true random samples, variable terms, risk, etc. make me laugh. I wouldn’t bet a lump sum on a 60% odds in Vegas. Better to look at the prevailing variables with high associations to the prevailing economic conditions. Maybe if someone had r^2s for those variables over time, I would stop laughing at this crap, but they don’t, do they?
What if they recreated this information and divide profit by risk. A number for risk-adjusted return would be a kick in the pants for anyone who actually took this advice.
Anyone reading who doesn’t understand statistics, the rule of thumb is that I have only found one article that used statistics correctly this year. Maybe start learning statistics by imagining your odds that anything with a “%” in the article actually means anything.
Any thoughts on this topic for this year? Everyone is predicting recession!