There are things one can buy in the market that fluctuate in value 20% or more a day. How would one figure out if dollar cost averaging could make one money, based on instruments with large fluctuations like this? Or perhaps more succinctly, would it and how much? Note, here the interval of dollar cost averaging might be several hours, so that over a 3 day span of time, one might dollar cost average 10 times, over something that varies approx 20% in a day.
As you mentioned in the title, what you're asking about comes down to volatility. DCA when purchasing stock is one way of dealing with volatility, but it's only profitable if the financial instrument can be sold higher than your sunk costs. Issues to be concerned with:
- The expected value of this financial instrument when you plan to sell it.
- How much you can stand to see the aggregate value in this instrument fluctuate
- Overtrading (i.e. you loose your expected profits from transaction fees)
- The extent that you need to depend on this instrument to defray margin-calls
Let's suppose you're buying a stock listed on the NYSE called
FOO (this is a completely fake example). Over the last six days, the average value of this stock was exactly $1.00Note 1. Over six trading days you put $100 per day into this stockNote 2:
11 Jan 2012: BUY 86 shares of FOO at $1.15 (Total: $98.90 + $5 Trading Fees) 10 Jan 2012: BUY 121 shares of FOO at $0.82 (Total: $99.22 + $5 Trading Fees) 09 Jan 2012: BUY 83 shares of FOO at $1.20 (Total: $99.60 + $5 Trading Fees) 06 Jan 2012: BUY 125 shares of FOO at $0.80 (Total: $100.00 + $5 Trading Fees) 05 Jan 2012: BUY 84 shares of FOO at $1.18 (Total: $99.12 + $5 Trading Fees) 04 Jan 2012: BUY 117 shares of FOO at $0.85 (Total: $99.45 + $5 Trading Fees)
At market close on January 11th, you have 616 shares of FOO. You paid $596.29 for it, so your average cost (before fees) is:
$596.29 / 616 = $0.97 per share
Let's look at this including your trading fees:
($596.29 + $30) / 616 = $1.01 per share.
When the market opens on January 12th, the quote on FOO could be anything. Patents, customer wins, wars, politics, lawsuits, press coverage, etc... could cause the value of FOO to fluctuate. So, let's just roll with the assumption that past performance is consistent:
Selling FOO at $0.80 nets:
(616 * $0.80 - $5) - ($596.29 + $30) = $123.49 Loss
Selling FOO at $1.20 nets:
(616 * $1.20 - $5) - ($596.29 + $30) = $107.90 Profit
Every day that you keep trading FOO, those numbers get bigger (assuming FOO is a constant value). Also remember, even if FOO never changes its average value and volatility, your recoverable profits shrink with each transaction because you pay $5 in fees for every one.
Speaking from experience, it is very easy to paper trade. It is a lot harder when you're looking at the ticker all day when FOO has been $0.80 - $0.90 for the past four days (and you're $300 under water on a $1000 portfolio). Now your mind starts playing nasty games with you.
If you decide to try this, let me give you some free advice:
- Only trade with money you don't care about loosing
- Do not trade on margin
- Look into calculating support and resistance points for this instrument; then buy near support and sell near resistance. DCA is much better-suited for long-term investing.
- Trade with standing stop-loss orderNote 3
Unless you have some research (such as support / resistance information) or data on why FOO is a good buy at this price, let's be honest: you're gambling with DCA, not trading.
- (1.15 + 0.82 + 1.20 + 0.80 + 1.18 + 0.85)/6 = 1.00
- Numbers don't add up to $100 each day because you can't buy a fraction of a share.
- Calculating the stop-loss is outside the scope of this answer
if you know when and by how much something will fluctuate, you can always make money. Buy it when it's cheaper and sell it when it's more expensive.
If you just know that it fluctuated a lot recently, then you don't know what it will do next. Most securities that go to zero or go much higher bounce all over the place for a while first. But you don't know when they'll move decisively lower or higher.
So how could you figure out if you'll make money - you can't know.
DCA will on average make you better off, unless the extra commissions are too high relative to your purchase sizes. But it will in retrospect make you worse off in many particular cases. This is true of many investment disciplines, such as rebalancing. They are all based on averages.
If the volatility is random then on average you can buy more shares when the price is lower using DCA. But when the lowest price turns out to have been on a certain day, you'd have been better off with a single lump sum put in on that day. No way to know in advance.
Degree of volatility shouldn't matter; any fluctuation is enough for DCA or rebalancing to get you ahead, though it's true they get you ahead farther if the fluctuations are larger, since there's then more difference between DCA and a lump purchase.
I think the real reason to do DCA and rebalancing is risk control. They reduce the risk of putting a whole lump sum in on exactly the wrong day. And they can help keep a portfolio growing even if the market is stagnant.
Dollar cost averaging is beneficial if you don't have the money to make large investments but are able to add to your holding over time. If you can buy the same monetary amount at regular intervals over time, your average cost per share will be lower than the stock's average value over that time. This won't necessarily get you the best price, but it will get you, on the whole, a good price and will enable you to increase your holdings over time.
If you're doing frequent trading on a highly volatile stock, you don't want to use this method. A better strategy is to buy the dips: Know the range, and place limit orders toward the bottom of the range. Then place limit orders to sell toward the high end of the range. If you do it right, you might be able to build up enough money to buy and sell increasing numbers of shares over time. But like any frequent trader, you'll have to deal with transaction fees; you'll need to be sure the fees don't eat all your profit.