Skip to content
All library documents

Early Exercise of Calls: Dividends, Put Value, and Optimal Stopping

Article Quant Q&A · Author: user2521987

Summary

The document clarifies why a dividend threshold alone is not enough to conclude that early exercise of an American call may be rational. It uses put-call parity for European options with dividends to relate the call’s continuation value to the stock price, strike, present value of dividends, interest, and the corresponding put value. Since the put has positive value, meeting the stated dividend threshold still leaves the European call worth more than its immediate exercise payoff in the argument presented.

For early exercise to be a candidate, the call’s value must be no greater than its intrinsic value. Rewriting that condition through parity shows that the put value must be considered alongside dividends and the financing benefit of delaying payment of the strike. The numerical example satisfies this necessary comparison, but the response cautions that comparing exercise now with exercise only at maturity does not prove that exercising now is optimal. The full decision requires considering all possible exercise times as an optimal stopping problem.

Key ideas

  • A dividend threshold by itself does not establish that early exercise of a call is rational.
  • Put-call parity shows that the European put’s value affects the comparison between continuation and immediate exercise.
  • Early exercise is a candidate when the call’s continuation value does not exceed its immediate exercise payoff.
  • Comparing only exercise now with exercise at maturity does not identify the optimal exercise time.

Tags

Full text
# Understanding the necessary and sufficient conditions for rational early exercise of a call option


# Understanding the necessary and sufficient conditions for rational early exercise of a call option












I am self-studying for an actuarial exam, and I encountered the following in my text:

The author states that if $PV_{t, T}\text{(Divs)} < K(1 - e^{-r(T - t)})$, early exercise is not rational.

That made me wonder if the converse is true: If $PV_{t, T}\text{(Divs)} \geq K(1 - e^{-r(T - t)})$, then early exercise may be rational.

But in an exercise, the author seems to suggest that there is more to it than that:

The author seems to suggest that we have to also consider the value of the put. I don't understand why we have to consider the value of the put in this example, yet we did not consider the value of the put in the statement underlined in red.

I'm looking for clarification on this.

## Answer by Gordon (score 2, accepted)

https://quant.stackexchange.com/a/21756

If $PV_{t, T}(\text{Divs}) \ge K\big(1-e^{-r(T-t)}\big)$, since $P_{Eur}(S_t, K, T-t) >0$, the identity \begin{align*} C_{Eur}(S_t, K, T-t) = P_{Eur}(S_t, K, T-t) + (S_t-K) -PV_{t, T}(\text{Divs}) +K\big(1-e^{-r(T-t)}\big), \end{align*} implies that \begin{align*} C_{Eur}(S_t, K, T-t) > (S_t-K). \end{align*} That is, it is not rationale to exercise the option at time $t$ in this case. Note that the conclusion also depends on the statement that "The put must be worth at least zero", which you should also highlight.

$$ $$ Note that, it may be rationale to exercise the option at $t$, instead of the maturity $T$, only if \begin{align*} C_{Eur}(S_t, K, T-t) \le (S_t-K). \end{align*} From the above identity, this is equivalent to \begin{align*} P_{Eur}(S_t, K, T-t) -PV_{t, T}(\text{Divs}) +K\big(1-e^{-r(T-t)}\big) \le 0. \end{align*} That is, the put option price is indeed taking into consideration. For the Quiz you provided, since \begin{align*} P_{Eur}(S_t, K, T-t) -PV_{t, T}(\text{Divs}) +K\big(1-e^{-r(T-t)}\big) &= 0.82 -2.9851 +1.4049 \\ &< 0, \end{align*} it may be rationale to exercise this option early.

$$$$ Here, we say that it may be rationale to exercise early. However, this does not mean it is optimal to exercise at time $t$, as we only considered two possible exercise times $t$ and $T$. In general, we need to take all possible stopping times, $\tau$, ranged from $t$ to $T$ into consideration. That is, we need to solve an optimal stopping problem.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.