Designing an efficient reference rate: Lessons from SOFIA

Abstract

We evaluate the informational efficiency of the Secured Overnight Funding Index Australia (SOFIA™), currently in its beta phase, and how design choices may impact the benchmark. We use a state-space model to separate time-varying noise in the benchmark from the underlying efficient rate and investigate the determinants of noise in the daily time series and at the transaction level. Related-party transactions, high market concentration, and low transaction volumes are associated with higher noise. Our modelling suggests that considering alternative methods for trimming the transactions used in calculating the benchmark, compared to the current approach, may enhance its informational efficiency and robustness. These results provide evidence on how to optimise the benchmark’s design. Indeed, the Australian Stock Exchange has already adjusted its SOFIA methodology to reflect most of these findings, in preparation for the potential transition of SOFIA from its beta phase to a live benchmark.

Keywords: interest rate benchmarks, market microstructure, reference rate, SOFIA

JEL Classification: G11, G12, G14