Multilateral trade reforms, such as may eventually emerge from the WTO's Doha Development Agenda (DDA), tend to be phased in over a decade or so after agreement is reached. Given the DDA's slow progress, that implementation may not be completed before the end of the next decade. Ex‐ante analysis of the DDA's possible effects thus requires first modelling the world economy to 2030 and, in that process, projecting what trade‐related policies might be by then without a DDA. Typically, modellers assume the counterfactual policy regime to be a ‘business‐as‐usual’ projection assuming the status quo. Yet we know developing country governments tend to switch from taxing to assisting farmers in the course of economic development. This paper shows the difference made by including political economy‐determined agricultural protection growth endogenously in the baseline projection. We reveal that difference by projecting the world economy to 2030 using the Global Trade Analysis Project (GTAP) model with those two alternative policy regimes and then simulating a move to global free trade (the maximum benefit from a multilateral trade reform) in each of those two cases. The welfare effects of removing the counterfactual price distortions in 2030 are shown to be much larger in the case where agricultural protection grows endogenously than in the case assuming no policy changes over the projection period. This suggests the traditional way of estimating effects of a multilateral agricultural trade agreement may considerably understate the potential welfare gains.