Change in Dividend Payout Ratio

Assuming the management of a company knows more about the prospects of the share price of its company that anybody else, it could be assumed that an increase in the dividend payout ratio leads to an outperformance of the share price.

I investigate this thesis for 97 companies of the S&P100 for the period of January 1980 to January 2009 (3 stocks were omitted as there was not enough data available). For various companies payout-ratio data is only available for a shorter period than 1980 - 2009, that's why the average length of the analysis is only around 18 years (i.e. since around 1990). Please note that for the entire analysis dividends are ignored. I focus uniquely on share prices.

The average yearly performance of a simple buy-and hold strategy would be 8.6% per stock on average. When we buy only stocks where the payout ratio has been increased and sell it again when it has been decreased (always on the next day's closing price), we get only an average return of 2.93% p.a. per stock (i.e. underperforming a buy-and-hold strategy). The distribution looks as follows: On the top we can see that around 20 stocks yielded almost 10% p.a. with a buy-and-hold strategy while the lower diagram shows the effect of applying the payout-ratio increase/decrease strategy: most of the stocks yield only around 2% but we were able to avoid some of the big losses (although the dataset is too small to draw any detailed conclusion about the distribution of the returns).

As usual, I focus my analysis only on 'disproving' the trading strategy. So in order to show that it is unlikely to work, I apply the reverse: We buy the stock only when the payout ratio is decreased and sell it when it is increased. If we do this, the histogram looks as follows:

The reverse payout-ratio distribution looks slightly different but interestingly the average return is almost unchanged, in fact it is even slightly higher: 2.94% (as opposed to 2.93%). There is no statistically significant difference between the two which makes it very unlikely this trading strategy has any value at all for the time period the 100 stocks were analysed.

Here is some additional information about the returns: When we do a stock-by-stock comparison whether the strategy outperforms a simple buy-and-hold we get the following distribution: In most cases it leads to underperformance.

The amount of trades per stock with the strategy are as follows: For 22 stocks we are buying 5 times and selling 5 times (i.e. 10 trades).

Winning trades with the payout-ratio increase/decrease strategy:

The following histogram shows that for most stocks we are holding a long position of slightly more than 20% of the time. For 20 stocks we are almost never long or not long at all. For none of the stocks we are long more than 55% of the time.

For most of the trades we are not yielding any significant return: 31 stocks yielded 0% return (among which are 20 which we didn't trade at all).

Conclusion: The most compelling case again the strategy is the fact that when we apply the reverse, the performance is almost identical. For the given sample it is safe to say that the strategy does not work.