1: \begin{abstract}

2: In this paper we consider the stability and convergence of numerical

3: discretizations of the Black--Scholes partial differential equation

4: (PDE) when complemented with the popular linear boundary condition.

5: This condition states that the second derivative of the option value

6: vanishes when the underlying asset price gets large and is often

7: applied in the actual numerical solution of PDEs in finance.

8: To our knowledge, the only theoretical stability result in the

9: literature up to now pertinent to the linear boundary condition

10: has been obtained by Windcliff, Forsyth \& Vetzal \cite{WFV04}

11: who showed that for a common discretization a necessary

12: eigenvalue condition for stability holds.

13: In this paper, we shall present sufficient conditions for stability

14: and convergence when the linear boundary condition is employed.

15: We deal with finite difference discretizations in the spatial

16: (asset) variable and a subsequent implicit discretization in time.

17: As a main result we prove that even though the maximum norm of

18: $e^{tM}$ ($t\ge 0$) can grow with the dimension of the semidiscrete

19: matrix $M$, this generally does not impair the convergence behavior

20: of the numerical discretizations.

21: Our theoretical results are illustrated by ample numerical experiments.

22: \end{abstract}