American Put Pricing with Backward Induction and Early Exercise
Summary
The document explains why a binomial-tree valuation of an American put compares the immediate exercise payoff with the continuation value at each node. Backward induction starts from the option’s terminal payoff and moves toward the present, calculating the discounted risk-neutral expectation of later node values. For a European option, this continuation calculation alone determines the value along the tree.
An American option can be exercised before expiry, so each node also has an immediate exercise value. The holder should take the greater of that payoff and the continuation value: exercising when the payoff is higher, or retaining the option otherwise. The answer frames this as optimal exercise rather than as a claim that the American option’s price process is a fair-game martingale; it notes that the process is a supermartingale and points to the Snell envelope. The explanation is conceptual and does not specify tree parameters, market inputs, or how early exercise behaves outside the illustrated put context.
Key ideas
- Backward induction values a European option by discounting risk-neutral expectations of later node values.
- An American option's value at a node is the greater of its immediate exercise payoff and continuation value.
- Early exercise is optimal when the immediate payoff exceeds the value of holding the option.
- The answer describes the American option price process as a supermartingale and relates valuation to the Snell envelope.
- The document gives no numerical tree example or market calibration details.
Tags
Full text
# On pricing american put options # On pricing american put options How come we pick the highest between the discounted weighted average (with risk neutral probabilities) and the early exercise value at each node of the binomial tree? I dont understand why, I can see why it "logical" to pick the greater but not why this would be the fair value. ## Answer by Kevin (score 2, accepted) https://quant.stackexchange.com/a/48649 Firstly, recall that American-style option may be exercised at every time point. If you model the stock as a tree, you need to check at every node, whether the investor would like to use his/her right of early exercise. So, how do you price derivatives in general? You build the tree for the stock price and then a second tree via so-called backward induction: You begin with the terminal payoff and work backwards through the tree by computing the ``discounted expectation'' of the future nodes. In the end, this gives you the price of a European-style option. For American-style options, you must also take early-exercise at every node into account. As you probably know, if the stock price has fallen sufficiently, you may wish to exercise your put option early. Thus, at every node, you firstly compute the so-called continuation value (discounted expectation of future nodes as above) and the immediate payoff which you would obtain when you exercised the option at this node. Which value do you choose? As Alex said, the maximum of both. Thus, you only exercise the option when the immediate payoff suceeds the value of continuing to hold your option. You wonder whether the price process of American options is perhaps not fair. Actually, it really is not fair. For European-style options, we can find a price process which is a martingale (model for fair games) but for American-style options, the price process is a super-martingale which is indeed, not a fair game. (you may want to read about Snell envelope)
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.