I have been reading articles on Seeking Alpha for many months now, especially ones concerning dividend growth investing (DGI) and I have learned a great deal from all the different contributors. I thought it was time I tried to contribute something myself in return.
I have a 401K retirement account which I use to invest in dividend growth stocks. My plan is to hold 20-30 good DGI stocks for the next 20+ years, and purchase more shares of the individual stocks quarterly when I get my pension contribution from work. My question is would it be better to do a simple dollar cost averaging technique (DCA) of buying new shares as soon as I receive the quarterly contribution, or would it be better to hold the extra money in my account and wait for dips in the individual stocks before purchasing them. I call this "timing," not because I would try to sell at the highs and buy back in at the lows, but simply because I would try to wait for the best time to purchase the additional shares.
For this particular test I chose to look at the dividend yield. I wanted to see whether if by following the yield, and waiting for times when the yield was higher than normal for that particular stock, and therefore the stock could be considered undervalued, would I be able to get a better return than if I did a strict DCA. I chose 10 stocks to study, Aflac (NYSE:AFL), Air products (NYSE:APD), Boeing (NYSE:BA), Deere (NYSE:DE), Emmerson Electronics (NYSE:EMR), Johnson and Johnson (NYSE:JNJ), McDonald's (NYSE:MCD), Procter and Gamble (NYSE:PG), Target (NYSE:TGT), and United Technologies (NYSE:UTX). I chose these stocks because I currently own all of them in my portfolio, and because they all have a long history of paying dividends. Using the historical price data from Yahoo.com I took the adjusted share prices going back to Jan 1, 1990, and the dividends paid, and put them into a spreadsheet. For the sake of ease I used the weekly closing prices rather than the daily prices. I then calculated the dividend yield at the end of each week, and calculated a running 13 week yield average. I then compared the weekly closing yield to the running average yield, and calculated what percentage it was above or below the average. This is the percentage above average yield (PAAY).
For the DCA part of the test, starting in 1990, I purchased $500 of each stock every thirteen weeks at the closing price, no matter what the price was. For the timing part of the test I would only purchase shares if either of the following two criteria were met: following the weekly change in yield, if there was an increase in the PAAY 3 weeks in a row, culminating in a yield that was at least 5% above the average yield, I would purchase the stock, OR if the PAAY increased above 10% at the end of any week I would purchase the stock. These purchases would only occur if there was a quarterly contribution sitting in my account waiting to be used. If no money was in the account at the time of the buy signal no extra funds were added to take advantage of the buy signal. If neither of these criteria was met I would hold the cash in the account until a buy signal occurred.
At the end of the test period I calculated the value of the portfolio for both the DCA and the PAAY tests.
The results are as follows:
As you can see for only one stock, DE, did the attempt to time my purchases based on yield lead to an increase in the value of the portfolio, and this increase was minimal. In all other cases the DCA method worked better. When you look at the actual data the reason is obvious. There are periods of as much as 3 years when the PAAY buy criteria are not met because the stock is in a long term up trend. During this time using DCA I would be making purchases all the way up, but using the PAAY method I would only make purchases when the stock was significantly higher. For example, from 2/93 to 4/96 Boeing rose from 11.61 to 28.68 with basically no down periods. During that time the PAAY method did not create a single buy signal, so when the buy signal finally occurred $6500 worth of stock was bought at 28.68 rather than buying $500 worth at lower prices all the way up.
My conclusion from this is that not only is DCA easier to do, by setting up an automatic purchase plan every 13 weeks, but it also had superior results than trying to time my purchases based on yield, hoping to get the stock when it is undervalued.
My next project is to compare DCA to a timing method based on technical indicators. I plan on using a low in the MACD histogram as my buy indicator.