1. you're trying to imply that because food budget is greater than water budget, that raising water prices on farmers will affect households more. This does not make any sense. Only a fraction of the price of food paid by a household is attributable to the price of water, so you can't really conclude that just because households spend more money on food than on water, that increasing water prices on farmers will cost them more.
2. even if that were the case, it's a necessary evil to incentivize lower water use crops. Suppose there were only two types of food, apples and oranges. They both cost around the same right now, but oranges require 10x more water. Raising water prices will raise orange prices, but that's arguably a good thing because it forces consumers to choose more water efficient foods. I don't see how the alternative (ie. continue subsidizing oranges even though they're more costly to society) is any better. Not to mention, you can redistribute the excess earnings back to citizens to make this revenue neutral.