This paper uses the growth accounting approach to estimate total factor productivity in the Australian agriculture industry between 1949 and 2012. To shed light on an unresolved debate on quantifying the roles of capital and labor, we compare the “ex‐ante” and “ex‐post” approaches to the estimation of returns to capital and labor, and find the former performs better than the latter in the context of the agricultural production account. We also demonstrate how the measurement of agricultural productivity may be improved by accounting for heterogeneity in output and input quality. Finally, our estimates are distinct from existing statistics in both the time length and industry coverage and provide new information about the long‐term trend of agriculture productivity in Australia.