When a stock goes ex-dividend, the stock price must drop by the same dividend amount since the declared wealth of the company is distributed to shareholders. This logic makes perfect sense but does it really go that way in reality? What if the market ignores the event?
This article presents my recent study about a trading strategy that exploits the investor behavior of ignoring dividend events. It appears that stock price may drop but not to the full extent of the dividend payment.
Of course, the inevitable market forces still dictates the price of the stock based upon other factors that might be unrelated to the dividend event. However, when statistics shows consistent results, I don't think there is any coincident. This article explains how the study was performed and its surprising result.
The trading strategy is extremely simple. We shall buy the stock at closing price the day before ex-dividend date and then sell it immediately at the opening price on ex-dividend date.
Let's revisit what happens on ex-dividend date. If you buy the stock prior to the ex-dividend date, your trade will settle just in time to be eligible for the declared dividend. Anyone buying the stock starting on ex-dividend date will miss the upcoming dividend.
As we limit the holding period to just half a day (from closing to opening), we hope that no other news should move the stock in a big way. The profit/loss would be calculated as follows:
P/L = gain from dividend + gain/loss from stock movement - transaction cost
For example, XOM declared a dividend of $0.63 per share and it went ex-dividend on Feb 6, 2014. Using this strategy, we would purchase the stock on Feb 5 at closing price of 89.58 and sell it on Feb 6 at opening price of 89.51. Suppose you purchase 100 shares, then the cost basis would be $8,958. Assuming a $10 transaction fee, the profit turns out to be:
P/L = $63 - $7 - $10 = $46
I obtained 4 years of stock price and dividend data (2010-2013) for the current list of DJIA stocks. I ran through the trading strategy over all ex-dividend dates. The P/L is calculated for each transaction for each stock. The average P/L for each stock is finally plotted in a bar chart.
Will you make any money?
During the 4-year period, the mean return is calculated as 0.137% per transaction. That means we can make $13.70 per $10,000 investment. That doesn't sound like too much, does it?
Let's assume that your broker charges you $10 per trade. Since we must open & close the position for each transaction, the total fee would be $20. Obviously, we won't be making any money with just $10,000 investment.
As the trading fee is fixed, we just need to scale up our investment amount. So let's say you invest up to $50,000 per transaction. In that case, the average profit per transaction goes up to $68.50. Taking away the trading fee, you pocket $48.50 per transaction. Now, since you have 30 stocks to play with and 4 quarterly dividends each, you will end up with 120 transaction per year. So the total profit in a year would be $48.50 * 120 = $5,820. Given an investment amount of $50,000, the total return would be 11.64%.
S&P 500 Analysis
You may wonder if 30 stocks is enough to draw the above conclusion. Let's confirm our theory by expanding our analysis to S&P 500 stocks.
I ran the same trading strategy over all dividend-paying stocks in the S&P 500 index. There are 415 stocks satisfying that requirement.
The result is surprisingly similar to what we've seen from DJIA stocks. The result is best represented in a histogram:
Out of 415 stocks being analyzed, 309 of them produced positive return and 106 gave negative return. The average P/L is 0.163%, excluding transactional cost.
How about SPY?
It would be interesting to see if we can just apply the same principal to SPY. Here's a boxplot about the return of this same trading strategy. Only 4 trades per year. We've gotten positive return every year.
While dividends are considered a distribution of wealth, a stock's price is largely determined the bulls and bears in the market. We can argue that most investors do not really adjust down their valuation of the stock immediately on ex-dividend date. It appears that everyone just take the dividend for granted, at least for a day.
The above analysis shows that buying a stock the day prior to ex-dividend date and sell it the immediately after would be a profitable strategy. While the net return seems fairly small, it could be profitable if you execute the strategy consistently.
A friend of mine asked how it would perform if we extend the analysis to more than 4 years. There's no reason not to go that extra mile, so here's the result:
Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
Additional disclosure: The entire article only represents my original and personal opinion and is unrelated to my current employer. In no event should I be liable for any loss or damages of any nature resulting from, arising out of, or in connection with the use of the content.