Dividends and Stock Prices

I've been writing about finance on a regular basis since 2000, yet it took me until this week to get clear in my own mind the significance of the questions: do stock prices fall in value in response to a  forthcoming dividend payment? and its related question: if so, why? There is a lot of material about which I am still very naive, I concede.

If I had been asked, I might have remembered some long-distant lesson about two guys whose names each begin with the letter "M," and the notion that dividend policy, in an efficient market, is neutral as to the value of a stock. So I would have denied that any move at all could be predicted with any degree of confidence.

That may still be the "right answer," but I now believe I understand that there is a controversy here, and why.  Figuring it out involved wrestling with vocabulary and chronology. My understanding is that the usual process is this: a company will say that it will pay dividends this quarter, and it will set a "record date" in the near future, and a "payment date" about a week after that. The payments will go out to everyone who owns the company's stock -- who is a "holder of record," as of the record date.  Hence the term.

But to make things more complicated, two days before the record date comes what is called the ex-dividend date. This exists because it can take a couple of days for a stock transaction to settle: for the necessary paperwork to get done between the time somebody shouts "sold" on a trading floor on your behalf and the time you are in deed a owner of record. Thus, before the ex-dividend date the stock was trading "with the dividend," -- part of what you were purchasing in buying it was the expectationof that dividend. On and after that date, the stock is trading "ex" the dividend. 

Intuitively, then, one would expect stocks to increase in value at the time of the announcement and drop in value again on the ex-dividend date. As of the announcement, the stock carries with it the promise of a near-immediate cash rebate, whereas after the ex-dividend day, the stock no longer carries the expectation of a cash payment that it had carried the day before.  Why, then, wouldn't it be worth a bit more after the one development and a but less after the other?

But the money doesn't come out of nowhere. The market at the time of the announcement understands that by these cash payments the company will be depriving itself of that amount of cash, and losing the opportunity to re-invest it in something productive. Further (and this was the key to the Miller-Modigliani argument to which I alluded above) the market is indifferent between an increase in the value of the stock by one dollar on the one hand and the pay-off of $1 as a dividend on the other. So these announcements don't seem to produce any increase in value.

There is an arbitrage argument for the irrelevance of the ex-dividend date, too. After the declaration date, everyone in the market knows when the dividend will be paid, and when the ex-dividend date arrives. If this situation were sufficient to create a price drop, then a lot of speculators would rush in a short sell the stock in the days leading up to the ex-dividend day, betting on that price drop. Their short sales would cause the price to fall earlier than that date, perhaps as soon as the day after the announcement. The date itself, then, would be an irrelevance.

The situation is complicated by the issue of taxation. Dividends are taxed more than are capital gains, a fact that may make some investors and traders less willing to buy a stock that has announced a dividend in that run-up to the ex-dividend day than they would otherwise be, and might thus reduce the extent of the drop, if any, on that day.

Theories notwithstanding, there is evidence that there is a decline ceteris paribus on or around the ex-dividend date.

Is the decline equal to the full value of the dividend to be paid, perhaps with some modification for tax considerations?  That is another question, and not one I yet want to try to tackle.