- It has been suggested that dividends permanently reduce share price in a relative sense, while others have suggested that the effect is only temporary and disappears after a few days.
- A previous pilot study showed that a large annual dividend paid out by Cheung Kong Holdings had a significant effect on the 5-day price change of the stock.
- In the present study, the 5-day price change of Exxon Mobil after 50 ex-dates was investigated.
- Preliminary statistical analysis suggested that the share price of XOM did not recover to the level it was before the dividend was paid.
Is the ex-date dividend reduction permanent, or temporary? This issue has stirred fervent debate amongst Seeking Alpha authors and commenters, with logic and emotion seemingly blurred together in passionate discourse (refer to my previous article for a flavor of the discussion).
As a newer investor, I certainly do not have the credentials to authoritatively put the question to rest one way or another. But my training in science encourages me to understand the relationships between factors affecting a particular outcome, to simplify with deconvoluting assumptions wherever possible, and to reach conclusions that, though necessarily limited in scope, can help to develop my own personal investment philosophy and inform my own investment decisions.
Therefore, when well-respected SA authors and commenters offered up this testable hypothesis, my interest was stirred:
Chuck Carnevale (article):
As you review the ex-dividend graphs, notice how sometimes the stock price rises almost immediately after the ex-dividend date, and sometimes it falls. In other words, post ex-dividend date pricing is more random, than affected by the ex-dividend date.
David van Knapp:
Because of this price adjustment, shareholders who owned the stock before the ex-dividend date, and who hold on straight through, see the price of their stock momentarily lowered by the amount of the dividend. In practice, as soon as trading opens on the ex-dividend date, market participants establish the actual price of the stock. The exchange's price adjustment is usually hard to detect after an hour or two of trading. There certainly is no long-term impairment to the price of the stock.
OK, look up the exdividend dates for JNJ, XOM and CVX and chart their stock prices on exdividend date and for the next 5 days afterward. Tells us if the price stays down or if it comes back up. Balls in your court bud.
Seeing is believing.
People can look up stock prices in Yahoo historical stock prices and check out the closing stock price prior to the ex-dividend date and look up the stock price on the opening price on the ex dividend day and the prices on following days. The prices are literally all over the place. Sometimes the opening price on the ex-dividend day is down less than the dividend amount and sometimes more than the dividend amount.
In my previous article, a pilot study was conducted to investigate the 5-day price behavior of Cheung Kong Holdings (OTCPK:CHEUY) after the dividend ex-date for 15 past ex-dates. This Hong Kong-based company was chosen because it pays a "large dividend" of 2% once a year, making the effect of the ex-date reduction more obvious. The results showed that the stock price of Cheung Kong Holdings, on average, did not recover the dividend amount five days after the ex-date, and that the difference was significant.
Therefore, this article sets out to answer the research question: does Exxon Mobil's (NYSE:XOM) share price recover 5 days after the ex-date dividend date?
Disclosure: I am not "pro-dividend" or "anti-dividend." Hopefully, the results of this study will be useful to commenters engaged in the dividend debates, as well to the wider Seeking Alpha audience, and shed light on the apparent contradiction between the two camps (temporary vs. permanent ex-date reduction). My own personal "buy-the-dip high-yield" investment philosophy is described here and here.
- Exxon Mobil was chosen for these reasons: i) it is a favorite DGI stock, ii) it pays a relatively smaller dividend (currently 2.7% per year or 0.67% per quarter), making it interesting to determine whether or not the ex-date reduction has a significant effect on the stock price, and iii) it was one of the stocks mentioned by the commenter above.
- The 5-day % price change was chosen as the benchmark as this was deemed to give sufficient time for the share price to "recover," but short enough that any stock-specific news would not have a large effect. Percentage change (rather than price change) was used as this was thought to allow for a fairer comparison between changes at different times
- Data was collected from Yahoo Finance historical prices. The share price of XOM before the ex-date recorded, as well as the share price five trading days later. For example, if the ex-date occurred on a Monday, then the previous day's close price was recorded (i.e., last Friday's), as well as the close price on the Friday of the week of the ex-date.
- This was done for 50 XOM ex-dates, spanning over 10 years. For comparison, the 5-day % price change for SPY was also recorded for the same time period (i.e., the time period covering XOM's ex-date).
In the previous pilot study on Cheung Kong holdings, which pays out a 2% large dividend once a year thus making the effect of the ex-date reduction more obvious, 15 ex-dates were used. However, I chose to use 50 ex-dates for Exxon Mobil because the dividend paid out each quarter is much smaller (0.67%). Collecting more data points should add more power to the statistical tests.
Exxon Mobil the company
Exxon Mobil is the largest publicly traded multinational oil and gas company in the world, and is also high-quality stock that is a favorite of both DGI and non-DGI investors alike. The figure below shows how XOM's dividend has increased steadily over the past ten years, though its share price has seen its fair share of fluctuations. In the process, XOM's dividend yield ranged from a low of about 1.4% to a high of about 3% over the past ten years.
Moreover, XOM has had a 10-year total return far outpacing that of the broader index (NYSEARCA:SPY) by about 6% per annum.
Results and analysis
The first table shows a simple count of the number of times (out of 50) that the share price of XOM was able to recover fully to the level at which it was at before the dividend was paid after 5 days. The same data is presented for the broader index over the same time period covering the ex-date of XOM.
|Yes (%)||17 (34%)||30 (60%)|
|No (%)||33 (66%)||20 (40%)|
The data shows that in the majority (66%) of cases, XOM was not able to recover to its pre-ex-date price after 5 days. On the other hand, the benchmark index appeared to show an upward bias in the period surrounding XOM's ex-date. To test whether these percentages were statistically significant, either the binomial test or the chi-squared test can be performed. For either test, the null hypothesis was that a 1:1 ratio of "yes" and "no" counts should be expected. The data for the statistical tests are presented below.
The p-value tells us, in layman's terms, the likelihood that the observed data are due to chance. The p-values calculated for XOM are below the standard threshold for significance (0.05). Simply put, this means that we are over 95% certain that there is a significant deviation between the observed percentage (34% chance of recovering to the pre ex-date price) and the expected percentage (50% chance, assuming that the post ex-date pricing is random). In other words, there is an inherent bias to the downside for the 5-day % price change of XOM after the ex-date that cannot be sufficiently explained by chance. For SPY, the p-values are above 0.05. This means that the 30:20 counts for SPY are not statistically different from an expected 1:1 ratio arising from the random nature of stock price changes on a short time scale.
A further analysis was carried out: of the 17 (out of 50) times that XOM was able to recover its share price to the pre ex-date price, how many times was it "helped" by a positive 5-day movement in the underlying index?
|XOM share price recovers||XOM share price doesn't recover|
|SPY 5-day positive change||15||15|
|SPY 5-day negative change||2||18|
The table above shows that of the 17 times that XOM recovered to its pre ex-date price after 5 days, it was aided by a positive movement in the underlying index in 15 of those times. In only two times (out of 17) was XOM able to recover its share price 5 days after the ex-date when the movement in the underlying index was down.
The next table shows data for the 5-day % price change for XOM for its past 50 ex-dates. The % change was calculated with the share price of XOM on the day before the ex-date as the denominator. Comparative data for SPY are also shown.
|5-day % price change||XOM||SPY|
The results show that the 5-day % price change of XOM after the dividend was paid out was around negative 0.90%, which is slightly higher than its historical 5-year quarterly dividend of 0.60%. This suggests that holding XOM through the ex-date might even be a negative (on the order of 0.3%). However, preliminary statistical analysis showed that this difference was not significant. Nevertheless, given that the benchmark index declined by an average of only 0.10% during the same time interval, the 0.9% drop in XOM is unlikely to be attributable to movements of the larger stock market as a whole.
The correlation between XOM and SPY in the 5 days after the ex-date of XOM is shown in the figure below.
There is a weak positive correlation, which is expected because XOM is an important constituent of the US market, but also a slight negative intercept, suggesting that XOM tends to move lower than the corresponding price change of SPY in the 5-days following the ex-date of XOM.
Due to the idiosyncratic fluctuations of individual stocks, the standard deviation of 5-day % price changes for XOM was rather high (2.85%). Therefore, a t-test was conducted to investigate whether the decrease of -0.90% was statistically significant compared to the expected value of 0% (assuming that share prices are random over the short term). Comparative data for SPY are also displayed.
As the p-value for the t-test is again below 0.05, the 5-day % price change decline in XOM (-0.90%) can be considered to be significantly different to the expected value of 0%. On the other hand, there was no significant difference between the -0.10% decline of SPY over the same period and the expected value of 0%. This indicates that the percentage drop in XOM is significant, i.e. on average, XOM does not recover its share price after 5 days.
The figure below illustrates the distribution of the % ex-date reduction in the stock prices of XOM and SPY graphically.
The figure above indicates a slight downward bias for XOM in the 5 days following its ex-date compared to the broader market as a whole.
Discussion and evaluation
This analysis was performed solely to answer the research question: does XOM's share price recover 5 days after the ex-dividend date? The results of this study suggest that the answer is no. Hence, we may conclude that the ex-date dividend reduction is not nullified by market fluctuations over the course of a 5-day time span for XOM. This therefore disagrees with the assertions presented the introduction.
As discussed previously, this conclusion makes sense because if the share price of XOM post ex-date was truly random, then the "dividend capture" strategy should be highly successful. Quoting from my previous article:
On an intuitive level, this makes sense because why would someone pay the same price for a stock a few days after the ex-date if you weren't entitled to receiving the cash?
However, there are limitations to such a simple study as this one. Is this reduction "permanent" (in the relative sense, at least)? In the absence of a wormhole that allows us to peer into an alternate universe where XOM has never paid a dividend, I cannot say definitively. Hence, this study does not address the questions:
- Would XOM have done better if it had not paid a dividend at all?
- What would the current share price of XOM be if it had never paid a dividend these past years?
- How many days would it take for XOM's share price to recover?
Additionally, there was an interesting issue that was posed by astute readers in the previous article.
Your analysis was very interesting but it doesn't take into account the possible rise in the stock price just before the ex-date which frequently happens. It would be interesting to see the results if you used the closing price about a week or ten days before the ex-date instead of the previous days'.
Thanks for your well-written and thoughtful article. I agree with a few others who have suggested you look at a time period of a week before the ex-dividend date also since the price tends to climb during this period.
Unfortunately, I had not yet had the time to fully explore this issue yet, given the much larger body of data to process, so this may have to wait for a further article. I hope this article has been useful, and I welcome all comments, suggestions and criticisms.
Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.