Model-Free Discretisation-Invariant Swap Contracts

TL;DR

This paper characterizes discretisation-invariant swap pay-offs that are free from monitoring and jump errors, providing exact pricing and hedging strategies, with empirical validation on S&P 500 data.

Contribution

It introduces a class of discretisation-invariant swaps with unbiased risk premia and exact hedging methods, expanding the toolkit for variance and higher-moment risk management.

Findings

Infinite variety of bias-resistant risk premia identified

Exact pricing and hedging achieved with vanilla options

Empirical validation on S&P 500 data confirms practical applicability

Abstract

Realised pay-offs for discretisation-invariant swaps are those which satisfy a restricted `aggregation property' of Neuberger [2012] for twice continuously differentiable deterministic functions of a multivariate martingale. They are initially characterised as solutions to a second-order system of PDEs, then those pay-offs based on martingale and log-martingale processes alone form a vector space. Hence there exist an infinite variety of other variance and higher-moment risk premia that are less prone to bias than standard variance swaps because their option replication portfolios have no discrete-monitoring or jump errors. Their fair values are also independent of the monitoring partition. A sub-class consists of pay-offs with fair values that are further free from numerical integration errors over option strikes. Here exact pricing and hedging is possible via dynamic trading strategies on a few vanilla puts and calls. An S&P 500 empirical study on higher-moment and other DI swaps concludes.