A Finite-Dimensional Approximation for Pricing Moving Average Options

Abstract

We propose a method for pricing American options whose pay-off depends on the moving average of the underlying asset price. The method uses a finite dimensional approximation of the infinite-dimensional dynamics of the moving average process based on a truncated Laguerre series expansion. The resulting problem is a finite-dimensional optimal stopping problem, which we propose to solve with a least squares Monte Carlo approach. We analyze the theoretical convergence rate of our method and present numerical results in the Black-Scholes framework.MSC 2010: 91G20, 33C45

A Pricing Formula for Delayed Claims: Appreciating the Past to Value the Future

“…with σ 0 and ϕ i as in (5). The following, well known fact (see [12]) is crucial for the rest of the paper.…”

Analysis of a multiple year gas sales agreement with make-up, carry-forward and indexation

“…In order to perform a detailed analysis of GSAs in terms of values, decisions, make-up usage, carry-forward usage, etc., we follow the assumption in [13] that the evaluation of the index follows a meanreverting Markov process. For the moving average pricing, we refer interested readers to [8] (a Least Square Monte Carlo approach), [11] (a binomial tree based approach), [26] (a willow tree approach) and [5] (a finite-dimensional approximation approach). that do not have make-up and carry-forward banks and those that do.…”

Finite-Dimensional Representations for Controlled Diffusions with Delay

“…where b, σ,α 1 ,β 1 are given functions and u = (u t ) t≥0 is a control process. Equations of this type appear in a variety of domains such as economics [23,24] and finance [1,4,5,19,25], as well as in physical sciences [29]. In general this equation is infinite-dimensional, which means that it can be formulated as evolution equation in an infinite-dimensional space of the form R × H 1 , where H 1 is a Hilbert space, for the process X t = (S t , (S t+ξ ) ξ≤0 ), but cannot be represented via a finite-dimensional controlled Markov process.…”

“…The actual convergence rate as n → ∞ will depend on the regularity of the functions α, β and γ. For example, from [4,Lemma A.4] it follows that if these functions are constant in the neighborhood of zero, have compact support and finite variation (this is the case e.g., for uniformly weighted moving averages) and w ≡ 1 then α − α n 2 w + β − β n 2 w + γ − γ n 2 w ≤ Cn −3/2 for some constant C and n sufficiently large. For C ∞ functions, on the other hand, the convergence rates are faster than polynomial.…”