Hi everybody, I am working on the implementation of the pricing of american options on discrete-dividend-paying stocks. Specifically applying Monte carlo methods.some time ago , I have already worked on the Longstaff Schwartz method; but it is applied to american options on non-paying dividends stocks.If someone could give me some advice about a improved Longstaff Schwartz method or another Monte carlo method for this case.Anyway, any suggestion about any other efficient method for pricing american options on discrete-dividend-paying stocks is welcome!Thank you

For discrete-dividend paying stocks, this is easy: Just introduce a drop of the size of the implied dividend payment D in stock price S: S_(t+) = S_(t-) - DI hope, this helps!Darou

Due to tax effects, the price drop on the ex-dividend date may be somewhat smaller than the full amount of the dividend. A small amount of empirical work should identify how strong that effect might be.

I got some articles giving a model that just decrease the dividend from the value of the stock. (As Darou says)But all the pricing formula derivation for american options that is used in Longstaff-Schwartz algorithm considers that the stock price is a geometric brownian motion without those discontinuities due to dividends.So, If my intuition is not wrong, I think it should be another pricing formula ad-hoc for this kind of dynamics with fixed jumps, or at least an algorithm that gives a sequence of approximated values that converge to the real price. What do you think?Thank you for the comments.

if you know how much dividend you are going to get in discrete times, you can annualized that dividend and use in GBM equation as Darou suggested