September 4th, 2014, 5:25 pm
I assume you refer to Variance Swap replication. In theory, you're right; the replication argument refers to euro-styles. More theoretically correct would be to first estimate what the euro-style otm option values would be (yes, probably using CRR), and then apply the VIX methodology to those.In practice, it likely makes little difference, but you could always try the `more correct' method if you're worried.