If agriculture were to be included in Australia’s carbon price scheme, a key decision for government would be how to estimate greenhouse gas emissions. We explore the consequences of three different methods for measuring on-farm emissions: national accounting methods, an amended version of those methods and use of best-available local data. Estimated emissions under the three methods can vary widely; for example, on a case study farm in Western Australia, local data indicated 44 per cent lower emissions than did the national accounts method. If on-farm emissions are subject to an emissions price, the impact on farm profit is large and varies considerably with different measurement methods. For instance, if a price of $23/t of CO2-e applies then farm profit falls by 14.4–30.8 per cent depending on the measurement method. Thus, the choice of measurement method can have large distributional consequences. On the other hand, inaccurate measurement results in relatively minor deadweight losses. On-farm sequestration through reafforestation may lessen the impact of an emissions price on farm businesses, although it will require a high carbon price to be viable, especially if sequestration rates are underestimated or low.