In mathematics, a local martingale is a type of stochastic process, satisfying the localized version of the martingale property. Every martingale is a local martingale; every bounded local martingale is a martingale; in particular, every local martingale that is bounded from below is a supermartingale, and every local martingale that is bounded from above is a submartingale; however, a local martingale is not in general a martingale, because its expectation can be distorted by large values of small probability. In particular, a driftless diffusion process is a local martingale, but not necessarily a martingale.
Local martingales are essential in stochastic analysis (see Itô calculus, semimartingale, and Girsanov theorem).
Definition
[edit]Let be a probability space; let be a filtration of ; let be an -adapted stochastic process on the set . Then is called an -local martingale if there exists a sequence of -stopping times such that
- the are almost surely increasing: ;
- the diverge almost surely: ;
- the stopped process is an -martingale for every .
Examples
[edit]Example 1
[edit]![](https://upload.wikimedia.org/wikipedia/commons/thumb/4/41/Local_martingale.svg/220px-Local_martingale.svg.png)
Let Wt be the Wiener process and T = min{ t : Wt = −1 } the time of first hit of −1. The stopped process Wmin{ t, T } is a martingale. Its expectation is 0 at all times; nevertheless, its limit (as t → ∞) is equal to −1 almost surely (a kind of gambler's ruin). A time change leads to a process
The process is continuous almost surely; nevertheless, its expectation is discontinuous,
This process is not a martingale. However, it is a local martingale. A localizing sequence may be chosen as if there is such t, otherwise . This sequence diverges almost surely, since for all k large enough (namely, for all k that exceed the maximal value of the process X). The process stopped at τk is a martingale.[details 1]
Example 2
[edit]Let Wt be the Wiener process and ƒ a measurable function such that Then the following process is a martingale:
where
The Dirac delta function (strictly speaking, not a function), being used in place of leads to a process defined informally as and formally as
where
The process is continuous almost surely (since almost surely), nevertheless, its expectation is discontinuous,
This process is not a martingale. However, it is a local martingale. A localizing sequence may be chosen as
Example 3
[edit]Let be the complex-valued Wiener process, and
The process is continuous almost surely (since does not hit 1, almost surely), and is a local martingale, since the function is harmonic (on the complex plane without the point 1). A localizing sequence may be chosen as Nevertheless, the expectation of this process is non-constant; moreover,
- as
which can be deduced from the fact that the mean value of over the circle tends to infinity as . (In fact, it is equal to for r ≥ 1 but to 0 for r ≤ 1).
Martingales via local martingales
[edit]Let be a local martingale. In order to prove that it is a martingale it is sufficient to prove that in L1 (as ) for every t, that is, here is the stopped process. The given relation implies that almost surely. The dominated convergence theorem ensures the convergence in L1 provided that
- for every t.
Thus, Condition (*) is sufficient for a local martingale being a martingale. A stronger condition
- for every t
is also sufficient.
Caution. The weaker condition
- for every t
is not sufficient. Moreover, the condition
is still not sufficient; for a counterexample see Example 3 above.
A special case:
where is the Wiener process, and is twice continuously differentiable. The process is a local martingale if and only if f satisfies the PDE
However, this PDE itself does not ensure that is a martingale. In order to apply (**) the following condition on f is sufficient: for every and t there exists such that
for all and
Technical details
[edit]- ^ For the times before 1 it is a martingale since a stopped Brownian motion is. After the instant 1 it is constant. It remains to check it at the instant 1. By the bounded convergence theorem the expectation at 1 is the limit of the expectation at (n-1)/n (as n tends to infinity), and the latter does not depend on n. The same argument applies to the conditional expectation.[vague]
References
[edit]- Øksendal, Bernt K. (2003). Stochastic Differential Equations: An Introduction with Applications (Sixth ed.). Berlin: Springer. ISBN 3-540-04758-1.