4
$\begingroup$

Let $\mu$ be a probability measure on $\mathbb R$ and set

$$c(K):=\int_{\mathbb R}(x-K)^+d\mu(x).$$

Assume that one has a sequence of probability measures $(\mu_n)_{n\ge 1}$ s.t.

$$\int_{\mathbb R}\left(x-\frac{i}{2^n}\right)^+d\mu_n(x)=c\left(\frac{i}{2^n}\right) \mbox{ for all } -n2^n\le i\le n2^n.$$

It is easy to see that the sequence $(\mu_n)_{n\ge 1}$ admits a weakly convergent subsequence. Without loss of generality, we may assume that it is weakly convergent. Let $\mu_0$ be its limit. My question is that can we show that $\mu_0=\mu$? I believe that the answer is no, but I can't find a counterexample. Thanks for the reply!

$\endgroup$
4
$\begingroup$

Let $c_n(k)$ be as $c(k)$ except for the measure $\mu_n$, $E_n$ be expectation under $\mu_n$ etc. . Since $| (x-a)^+ = (x-b)^+ | < |(a-b)| $, $c_n(k)$ is continuous. It is clearly also monotone. So it is easy to show that $c_n \rightarrow c$, even uniformly. Let $f$ be twice differentiable with compact support. $ f(x) = \int f^{\prime \prime} (k) (x-k)^+ dk$ (by integrating by parts twice, so $E_n(f) = \int f^{\prime \prime} (k) c_n(k) dk$. It follows that $E_n(f) \rightarrow E(f)$ for such $f$. By choosing such a function that is 1 on $(-N,N)$ where $\mu(-N,N) \approx. 1$ it follows that most measures have most of their support in that interval. As any bounded continuous function can be uniformly approximated by a twice differentiable function with compact support over $(-N,N)$ it follows that the measures converge weakly.

$\endgroup$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.