We should definitely not expect the “true fraction of beef consumption” to be proportional to impact. Steaks are consumed basically the same way as they were before subsidies (though in much larger quantity); they don’t respond much to the subsidy to take advantage of it. Fast food isn’t restricted to being prepared or sourced in a particular traditional way and therefore will change itself to best exploit subsidy. Estimating that effect as a 2.5x multiplier seems like a perfectly good conservative approximation, so you should just stick with 1%.
We should definitely not expect the “true fraction of beef consumption” to be proportional to impact. Steaks are consumed basically the same way as they were before subsidies (though in much larger quantity); they don’t respond much to the subsidy to take advantage of it. Fast food isn’t restricted to being prepared or sourced in a particular traditional way and therefore will change itself to best exploit subsidy. Estimating that effect as a 2.5x multiplier seems like a perfectly good conservative approximation, so you should just stick with 1%.