Launching Probabilistic Forecasts for Quarterly Trimmed Mean Inflation: AUSTRALIA AND NEW ZEALAND ECONOMICS ANALYST

Research evidence excerpt

Launching Probabilistic Forecasts for Quarterly Trimmed Mean Inflation: AUSTRALIA AND NEW ZEALAND ECONOMICS ANALYST

Economics Research

22 July 2026 | 3:43PM AEST

AUSTRALIA AND NEW ZEALAND ECONOMICS ANALYST

Launching Probabilistic Forecasts for Quarterly Trimmed Mean Inflation

n High inflation remains the key policy concern for the RBA. We launch Oscar To

+61(2)9320-1367 | oscar.to@gs.com

probabilistic forecasts for quarterly trimmed mean inflation, the RBA’s preferred Goldman Sachs Australia Pty Ltd

measure of underlying inflation, to help inform the near-term inflation outlook.

n Our probabilistic forecasts provide perspective on the balance of risks to our

modal forecast by taking a strictly empirical approach to quantifying the

likelihood of different inflation outcomes. We derive them using Monte Carlo

simulations, estimating a distribution for trimmed mean inflation based on

historical forecast errors for individual CPI components.

n Our modal forecast and probabilistic forecasts are designed to be

complementary. Our probabilistic forecasts do not capture qualitative

information and subjective judgements that shape our modal forecast. By

publishing both forecasts, we show the distribution of outcomes that a

mechanical, historically-calibrated view of forecast uncertainty would suggest,

and on top of that, our view of the most likely outcome taking into account all

other available information.

n There are two key benefits of our Monte Carlo approach. Firstly, the simulations

generate many possible sets of component-level forecasts to capture the

uncertainty in individual CPI components that a single set of forecasts misses.

Secondly, the probability distributions update throughout the quarter as partial