Key takeaways
  • Dividend jumps stay visible in the model
  • Price and exercise timing solve in sequence
  • Cash and proportional payouts fit one setup
  • Grid-heavy solvers are not the only option

Owning an American option on a dividend-paying stock is a timing puzzle: exercise too early and you may leave money on the table; wait too long and a dividend can knock the share price down. This paper tackles that problem for assets such as stocks and ETFs that pay discrete and/or continuous dividends. The author turns the pricing task into an integral Volterra equation using the Generalized Integral Transform, instead of attacking the usual free-boundary partial differential equation head-on. In the example model, discrete cash and proportional dividends are written with Dirac delta functions, which lets the option price and the early exercise boundary be solved step by step. The paper says this semi-analytical route handles dividend jumps cleanly and avoids the grid-heavy backward induction used in binomial trees and finite-difference methods. Several examples show the method is highly accurate and computationally efficient, making it a practical alternative when standard continuous-dividend models break down.

Imagine buying an option on a stock that pays a dividend tomorrow. The ex-dividend date is the day the dividend leaves the share price. That drop can make exercising early look smart. American options let you exercise any time before expiry. So the timing choice never really stops. The usual smooth-dividend model blurs that jump away. This paper keeps the jump in the picture. It turns the pricing task into an integral equation. That lets the price and the early exercise boundary come out step by step. The strange part is the best move here is to respect the mess, not hide it.

Why a dividend can flip the answer

This model starts with geometric Brownian motion, a stock-price model with drift and random wiggle. It adds two kinds of payouts. Cash dividends pay a lump sum. Proportional dividends take a slice of the price. Dirac delta functions, tiny math spikes for instant jumps, mark both events. Several examples show that the method stays accurate. They also show that it stays efficient. Binomial trees build a branch-by-branch price grid. Finite-difference methods use small time and price steps. Both need big grids and backward steps. Those tools can struggle when the price drops in a jump. Here, the jump stays explicit instead of getting smoothed away.

How GIT rewrites the pricing puzzle

GIT means Generalized Integral Transform. It is a math move that reshapes the whole problem. Instead of attacking a partial differential equation, GIT turns the task into a Volterra equation. A partial differential equation is a rule that tracks change in time and price. A Volterra equation is an integral equation that looks back across earlier times. This model uses a GBM setting with discrete cash and proportional dividends. GBM means geometric Brownian motion, a stock model with drift and random wiggle. It writes those jumps with Dirac delta functions. Then the solver finds the option price first. It then finds the early exercise boundary, the stock level where early exercise beats waiting. That order matters. It turns one hard puzzle, where the best exercise point is part of the answer, into a sequence.

  1. The transform keeps discrete cash dividends explicit.
  2. It also keeps proportional dividends explicit.
  3. The solver finds the option price before the exercise boundary.

“Discrete dividends create abrupt price drops and affect the optimal exercise timing, making traditional continuous-dividend models unsuitable.”

From the abstract

Why this matters for pricing desks

For pricing desks, the payoff is cleaner timing. A dividend can make early exercise the best move just before the ex-dividend date. This method keeps that jump visible. Tree methods also force heavy backward induction. Backward induction means you solve from the end back to the start. Grid-based solvers can lose speed or accuracy when jumps break the smooth path. The GIT route gives a practical alternative when a stock or ETF pays cash or proportional dividends. Several examples back up that claim. The result is not just a tidier formula. It is a way to keep the key choice tied to the real price jump.

What still needs to hold up

The surprise here still matters. A dividend jump does not have to wreck the pricing path. It can become the thing the math tracks best. That opens the door to cleaner tests of American options on dividend-paying stocks and ETFs. The next hard check is whether the same step-by-step route holds under other dividend patterns. Several examples already show the route. It does not claim a full cure for every market setting. It does show a useful shift. The jump no longer forces the model to fight the problem on a giant grid.