Early Exercise and Dividend Capture: When Exercising Early Is Rational, and the Dividend Play

Early exercise of an American call is rational only just before the stock goes ex-dividend, and only when the dividend is worth more than the option’s remaining time value plus the interest saved by paying the strike later. Early exercise of an American put can be rational at any time, when the put is so deep in the money that the interest earned on the strike outweighs what is left of its insurance value. Everything else is usually money left on the table.

In short: Robert Merton proved in 1973 that, with no dividends, an American call is worth the same as a European call and should never be exercised early. Dividends are the exception: on the last day before the ex-dividend date, a call holder should exercise if the dividend exceeds the interest on the strike to expiry plus the value of the same-strike put. Studies by Pool, Stoll, and Whaley (2008) and by Hao, Kalay, and Mayhew (2010) found that a large share of holders fail to do this, and that market makers built a high-volume trade, the dividend play, to capture what those holders leave behind. The trade works because of how the Options Clearing Corporation assigns exercises.

The basic mechanics of exercise, assignment, and automatic exercise are in how exercise and assignment work. This article covers the economics and evidence behind the early-exercise decision. It describes mechanisms, not recommendations.

A common misconception: “exercise early to lock in the gain”

A trader holding a deep in-the-money call often thinks exercising “locks in” the profit. It does lock it in, but selling the call would usually lock in more. Exercise captures only the intrinsic value (stock price minus strike). Selling captures intrinsic value plus whatever time value is left. Exercise also pays the strike today instead of at expiry, and gives up the protection the call provides if the stock collapses. Those three costs are why the rule of thumb for calls is “sell, don’t exercise,” and why the exceptions are narrow.

Merton’s theorem: calls should not be exercised early

Merton’s “Theory of Rational Option Pricing” (Bell Journal of Economics and Management Science, 1973) built a set of restrictions on option prices from one weak assumption: investors prefer more to less. One of them is that, if no dividends are paid over the option’s life, a European call is worth at least S - K x e^(-rT): the stock price minus the present value of the strike. Since that is more than the S - K you get by exercising, a live call is always worth more than an exercised one. An American call is therefore never exercised early and is worth exactly what a European call is worth.

Merton then dealt with dividends. His conclusion: “dividends and unfavorable exercise price changes are the only rational reasons for premature exercising,” and such exercise “will never take place except at boundary points,” meaning just before a dividend is paid. He also gave a sufficient condition for never exercising: if the present value of the remaining dividends is less than the interest that could be earned on the strike over the same period, the call is worth more “alive” than “dead.” For puts, his Theorem 13 says that an American put will, with positive probability, be exercised prematurely, so no simple put-call relationship pins down its value. Later work by Richard Roll (1977), Robert Geske (1979), and Robert Whaley (1981) produced formulas for an American call with one known dividend.

The reason dividends matter is that US listed options are not adjusted for ordinary cash dividends. When the stock goes ex-dividend, its price falls by roughly the dividend and the strike does not move, so the call loses value overnight. A call holder who exercises before the ex-date receives the dividend; one who does not, does not.

The exact test for a call before a dividend

Suppose there is one dividend D left before expiry, and the stock goes ex-dividend tomorrow. Today, the last cum-dividend day, the holder compares:

Exercise today: pay K, receive a share worth S that includes the dividend. Value S - K.
Hold: after tomorrow the stock is worth S - D and no dividends remain, so the American call is worth a European call. By put-call parity, that is (S - D) - K x e^(-rτ) + P, where P is the same-strike European put and τ the time left.

Subtracting, exercise is better when:

D > K x (1 - e^(-rτ)) + P

The right side has two parts: the interest saved by paying the strike later, and the put, which is the insurance a call holder keeps by not owning the stock. If the dividend beats both, exercise.

Worked example: Maple Widgets before a $1.00 dividend

Maple Widgets Inc. trades at $60 on the last cum-dividend day. It goes ex-dividend tomorrow for $1.00. The calls expire in 30 days (29 days after the ex-date). Implied volatility is 25% and the interest rate 4%.

Strike	Exercise value	Hold value (call on $59, 29 days)	Interest on K	Put value	Test threshold	Decision
$50	$10.00	$9.17	$0.16	$0.01	$0.17	Exercise
$55	$5.00	$4.48	$0.17	$0.31	$0.48	Exercise
$57	$3.00	$2.95	$0.18	$0.76	$0.95	Exercise (barely)
$58	$2.00	$2.30	$0.18	$1.12	$1.30	Hold

Per contract of 100 shares, the holder of the $50 call who forgets to exercise loses $83.08 ($10.00 - $9.17, times 100). For the $57 call, the cost is only $5.49, small enough that commissions can wipe out the gain, which is one reason some non-exercise is rational. Allowing for the one extra day of price movement before the ex-date barely changes this: holding the $57 call through tomorrow is worth about $2.97, still below its $3.00 exercise value, while the $58 call’s hold value rises slightly to about $2.33, so the decisions stand.

The test is most likely to say “exercise” when the call is deep in the money (small put value), close to expiry (small interest and time value), and the dividend is large.

The timing detail

Exercise must happen early enough to own the shares on the dividend’s record date. Since the US moved to one-business-day (T+1) settlement in May 2024, the ex-dividend date and the record date are generally the same day, so an exercise notice submitted on the last business day before the ex-date settles in time. Brokers set their own cut-off times for early exercise instructions, often earlier than the exchange deadlines, and the holder must have the cash or margin to take the shares.

When puts should be exercised early

A put holder who exercises receives the strike in cash today. Waiting delays that cash, so the interest on the strike pushes toward early exercise, while the put’s remaining insurance value pushes against it. When the stock is low enough, the interest wins.

Take a $50 put on Maple with Maple at $30, six months to expiry, 25% volatility, 4% rates:

Exercise value: $20.00
European put value: $19.02 (it is worth less than its exercise value, because the $50 is received six months late)
American put value (binomial): $20.00, meaning exercise now is optimal
Interest on $50 over six months: about $0.99, against a call (the put’s remaining “upside”) worth $0.006

For these inputs, the binomial tree puts the early-exercise boundary at about $38.85: below that price, the American put is worth exactly its intrinsic value and should be exercised. At $40, the American put is worth $10.03 against $10.00 intrinsic, so holding is still marginally better.

Dividends work the other way for puts: the stock price drop at the ex-date helps the put holder, so a pending dividend delays put exercise.

The evidence on puts mirrors the evidence on calls. Kathryn Barraclough and Robert Whaley, “Early Exercise of Put Options on Stocks” (Journal of Finance, 2012), examined exercise behavior from January 1996 to September 2008 and found more than 3.96 million puts that should have been exercised early that were not, over 3.7% of all outstanding puts.

Frictions can change the answer

The rules above assume frictionless markets. Christian Skov Jensen and Lasse Heje Pedersen, “Early Option Exercise: Never Say Never” (Journal of Financial Economics, 2016), show that early exercise can be optimal when it reduces short-sale costs, transaction costs, or funding costs. A trader who cannot borrow stock cheaply, or who faces wide option spreads, can rationally exercise a call that a frictionless model says to sell. “Never exercise early” is a frictionless result.

The evidence: many holders do not exercise

Veronika Krepely Pool, Hans Stoll, and Robert Whaley, “Failure to Exercise Call Options: An Anomaly and a Trading Game” (Journal of Financial Markets, 2008), studied calls on stocks paying quarterly dividends of at least a penny from January 1996 to mid-2006. Their findings:

More than half of outstanding long positions went unexercised on days when exercise before the ex-date was optimal.
The failure cost call holders more than $491 million over the period.
Market makers captured most of the money left on the table using a “dividend spread” strategy, and that strategy produced abnormally high volume and exercise activity around ex-dividend days.

Jia Hao, Avner Kalay, and Stewart Mayhew, “Ex-dividend Arbitrage in Option Markets” (Review of Financial Studies, 2010), reached a similar estimate for 1996 to 2006: more than 40% of calls that should have been exercised were not. Their example is Altria, which paid a $0.68 dividend with an ex-date of March 11, 2004. Option volume on March 10 was roughly thirty times the usual level; per OCC data cited in the paper, 97% of it was calls and 89% was market-maker volume. They also found that the fraction left unexercised falls as the reward rises: in their data, 86.88% of the open interest stayed unexercised when the gain was one cent a share, and 39.34% when it was 72 cents.

How the dividend play works

The assignment rules it relies on

When a holder exercises, OCC assigns the exercise to a clearing member account that is short the same series. OCC’s procedures are not written into its rules themselves but are treated as a stated policy under OCC Rule 803, per an OCC filing published by the SEC in December 2003 (Release No. 34-48908), which said OCC “uses a random assignment procedure for most classes of options.” The clearing firm then allocates the assignment to a customer: under FINRA Rule 2360(b)(23)(C), on a “first in-first out” basis, an automated random basis approved by FINRA, or a manual random basis specified by FINRA, and the firm must tell customers which method it uses.

Hao, Kalay, and Mayhew describe the critical sequencing in OCC’s end-of-day processing as it worked during their study: for market-maker accounts, purchases were processed first, then exercises, then sales, so “a party can buy, exercise, and sell the same series on the same day,” and sales that close out positions were processed after assignment. A position opened and exercised the same day therefore counts toward the pool of shorts that OCC assigns against.

Worked example: diluting the assignment pool

Say the Maple $50 calls from the earlier example have 10,000 contracts of open interest on the last cum-dividend day. Rational holders of 60% (6,000 contracts) exercise; 4,000 contracts go unexercised. Each unassigned short gains $83.08 when the stock goes ex-dividend, because the call it is short falls from $10.00 to $9.17 of value. Without anyone else involved, the original writers expect to keep that windfall on 4,000 contracts, about $332,000.

Now two market makers each buy a large number of calls from each other and immediately exercise all of their longs. Both are left long and short the same amount; the longs are exercised, and the shorts join the assignment pool. With random assignment proportional to shorts outstanding:

Contracts traded by each arbitrageur	Share of shorts assigned	Original writers left unassigned	Arbitrageur shorts left unassigned	Arbitrageurs’ expected share of the windfall
0	60.0%	4,000	0	0%
20,000	92.0%	800	3,200	80.0%
100,000	98.1%	190	3,810	95.2%
500,000	99.6%	40	3,960	99.0%

The arithmetic for 100,000 each: shorts total 10,000 + 200,000 = 210,000; exercises total 6,000 + 200,000 = 206,000; so 98.1% of every short is assigned. The arbitrageurs’ unassigned shorts (about 3,810 contracts) earn about $316,000 of the $332,000, while the original writers’ expected windfall shrinks to about $16,000. The trade needs no capital and carries little risk because the two sides are offsetting, but it only pays when per-contract costs are tiny, which is why it is a market-maker trade. Hao, Kalay, and Mayhew found that exchange fee caps introduced in the period made the trade more attractive and that the resulting volume could distort reported volume, put-call ratios, and exchange market share. The unexercised holders lose the same amount either way; the play only changes who collects it.

What writers of calls should know

Assignment risk peaks the day before the ex-date. Covered-call writers and anyone short in-the-money calls on dividend stocks should expect assignment when the test above says “exercise.” If assigned, the writer has sold the stock before the ex-date and does not receive the dividend; if the writer was short the call without owning the stock, they become short the stock over the record date and owe the dividend. See assignment risk and early exercise.
Spreads can break. A short in-the-money call in a vertical spread can be assigned while the long call, further out of the money, is not worth exercising. The trader wakes up short stock, owing the dividend.
European-style index options are immune. SPX and other European index options cannot be exercised early, so none of this applies to them. The differences are in American vs European options.

What can go wrong

Exercising too early. Exercising a call when the test says hold gives away the put value and the interest; with Maple’s $58 call, that is $0.30 per share.
Missing the deadline. Broker cut-offs for early exercise can be earlier than the exchange deadline, and a missed exercise on a deep in-the-money call before a large dividend costs the full difference between the dividend and the remaining time value.
Surprise or special dividends. The test depends on knowing the dividend. A cut, an increase, or a special dividend changes the answer, and OCC’s treatment of non-ordinary dividends can differ from ordinary ones, so check the OCC adjustment memo for any unusual distribution.
Assuming random means rare. Random assignment means the share of shorts assigned depends on how many longs exercise, including arbitrageurs’ freshly created longs. On dividend-play days, a writer can be assigned with near certainty.
Ignoring frictions. Commissions, borrow costs, and margin can make the theoretically correct choice wrong for a particular account, in either direction.

The short version

Merton (1973) showed that without dividends an American call should never be exercised early, because selling captures time value that exercise destroys, and paying the strike later earns interest. A dividend is the one rational exception for calls: exercise on the last cum-dividend day when the dividend exceeds the interest on the strike plus the same-strike put value. Puts are different: interest on the strike makes early exercise optimal once a put is deep enough in the money, about $38.85 for a six-month $50 put in our example. Research by Pool, Stoll, and Whaley (2008), Hao, Kalay, and Mayhew (2010), and Barraclough and Whaley (2012) shows many holders get these decisions wrong, and the dividend play, which floods OCC’s random assignment with offsetting positions, lets market makers collect most of what call holders leave behind.