Last week’s post “Risk and Return in Pictures” received some requests to chop up the data in different ways, so this is the second follow up post to try to do just that. I suggest taking a look at the original post by clicking here if you want a refresher, and then also take a look at Risk and Return in Pictures Part II by clicking here.
Dollar Cost Averaging
A reader indicated that it would be interesting to see the results of Dollar Cost Averaging versus Lump Sum investing when looking at these portfolios. What the readers want, the readers get… sometimes. This happens to be one of those times. :)
First, I only took three sample portfolios: All Equity, 50% Equity and All Fixed Income. This should be enough to tell us what we want to know. Let’s start by looking at the full period I had previously used, January 1988 to August 2008:
The total amount invested would’ve been $24,800 in all three cases and the all equity portfolio had 28.64% more value when compared to the all fixed income portfolio ($76,277 versus $59,296).
Next, let’s look at $100 monthly starting at the peak of the tech bubble:
In this case there were only 96 months, so the total investment would’ve been $9,600. The all equity portfolio had a 26.94% advantage over the all fixed income portfolio at the end ($15,414 versus $12,143).
And finally, let’s look at $100 monthly starting at the bottom of the market after the tech bubble burst:
This was the shortest time period and had only 71 months and $7,100 invested. The all equity advantage was 27.30% over the all fixed income portfolio ($10,576 versus $8,308).
In all three cases, the All Equity portfolio returned superior results when using dollar cost averaging versus the All Fixed Income portfolio. The conclusions are not as clear as you would think though. Over time as more money is invested the volatility differences between portfolios would increase – eventually you get to a point where the monthly fluctuations in the portfolio become larger than the amount of money you are adding monthly. As always, there is more to discuss and I will revisit these graphs in the future.
Sorry this is slightly off topic, but I’m still very interested in the similar concept of Dollar Value Averaging. I hope to see or find some more discussion of it.
The concept was new to me when I read about it in Bernstein’s four pillars book. All of Bernstein’s numbers and examples convinced me it can truly add value without additional risk by systematically buying more low and selling high then dollar cost averaging.
So I figured I’d see a lot of support and discussion about it in the personal financial blogosphere since everyone is very hands on and ready to put in the extra effort it takes to squeeze out extra returns. I was a bit more surprised to hear it’s not really advocated by Dimensional (at least their software models don’t support it).
So I’m left with doubt, is there something I missed, that this theory was later disproven? or is just way more hassle then it’s worth? Why don’t PF or Dimensional use this strategy?
Thoughts and insights are much appreciated!
Jordan
I’m making an educated guess here Jordan, but my guess would be fees. If you have to pay a fee for each transaction, 12 transactions a year could easily eat up any benefits of dollar cost averaging.
So it really depends on the fees for each contribution. If there aren’t any, then dollar cost averaging makes sense as a strategy, but if there are, you have to consider how much the fees are, and whether it is worthwhile or not. Even if the fees don’t make sense for monthly investments, you can also cost average over a longer time frame (say every 4 months). Just transfer the money to a high interest savings account monthly like you normally would, but then every 4 months you can use the accumulated cash to purchase funds/stocks/whatever. It cuts down on fees (assuming there are some), you are still putting away money every month and it is earning interest in the meantime, and you are still dollar-cost averaging, just over a longer time frame.
Hi Xenko,
I agree that the fees have to be considered, unless you were buying into a mutual fund I don’t think you’d do monthly contributions. But I am referrering to Dollar Value Averaging (DVA), not Dollar Cost Averaging (DCA).
Here are two explanations if you haven’t read Four Pillars:
http://www.financemind.com/investment/dollar-value-averaging.html
http://www.investopedia.com/articles/stocks/07/DCAvsVA.asp
Your suggestion of saving money in a high interest savings account is the same to invest with DVA as well. The difference would be the amount you invest each month fluctuates with the actual value of the portfolio compared with how much you want your portfolio to be. So you need a pool of uninvested cash to draw from when the market goes down. If the market goes up you don’t contribute as much or if it goes up very quickly you are actually supposed to sell and put the money back into cash pool.
If the market is down significantly like it has been this year you need to make bigger contributions, this is how the technique systematically buys more when stocks/funds/indexes are low and optionally sells high without having to time the market.
Jordan
@Jordan – I would peg it down to the hassle factor for the most part, plus when you are talking about longer time frames, you need to be more and more accurate with your estimated long term return to maximize your results – that’s tough to do. That an equity premium exists is one thing, that we know what it will be in the future is something entirely different. If you are value averaging in a taxable account, the taxes may eat up much of the extra gains as well (not to mention having the fixed income taxed at your marginal rate while on the sidelines).
Bah, I didn’t realize you were talking about DVA and not DCA. I’d agree with Preet that predicting the long term rate of return is an impossible task and that you could be potentially limiting your gains using DVA. Also, over shorter time frames, it is impossible to predict where the market will go (look at the last 12 months), and a 1 year DVA plan (like that used in the example from investopedia) wouldn’t work so well. Another issue with DVA is that in some months you have to contribute a large amount if the market is falling, but people might not have access to that much money at that point, whereas DCA fits into a person’s budget very easily.
DVA seems to assume that you know how much you will need, when you need it, and what interest rate you will get. You probably know the first two things, but you have no idea what the interest rate will be over that time frame, and that seems to be the crux of the issue. It would be interesting to see how DVA performed vs. DCA over the most recent 20 year period in an index fund, with DVA assuming different initial interest rates and see how that would compare with DCA.
Another thing I just thought of is that DVA in the examples appears to be just 1 stock/fund/bond and a cash position. In essence, isn’t this basically like a simple portfolio with 1 stock and 1 bond with a target allocation of say 90:10. If the stock is hot, you would have to sell to get it back to the 90:10 ratio, and if it is dropping, you would have to sell some of the bond and buy more stock? It seems to me like DVA is just a method of buying one stock at some theoretical “fixed rate” with a cash position that is constantly in flux to try and maintain that fixed rate.
Hi Preet,
This article has been very useful for me. Would you provide a link to the spreadsheet that you used in your dollar cost averaging calculation?
Thanks, and keep up the great work!
Brian
@ Brian – I knew someone was going to ask as soon as I pushed the delete button on the spreadsheet. But if you have the time and inclination you can just download the daily index values from a site like yahoo finance and then put it in excel and then create a simple formula to show how DCAing would work.