Say we kill all mosquitoes in the wild, and then discover that they were in fact protecting us from some Awful Thing. By the time we become aware of this, it may be too late to eliminate the problem by simply reintroducing caged mosquitoes to the wild, because some irreversible critical mass or adaptation may have happened in the meantime. (This is a Black Swan argument; it's a vaguely-defined scenario that seems very unlikely but potentially catastrophic.)
"The cost of these infections is immense, both economically and in terms of human suffering."
No doubt. But what we want is a cost-benefit analysis of removing mosquitoes from the ecosystem. Pointing to one large and obvious benefit does not help us calculate the costs.
I have described the immense harm caused by mosquitos. I have looked equally hard for mosquito benefits and found none that can't be provided by other species. To stand on the side of keeping mosquitos, you must find a benefit that outweighs trillions of dollars of lost economic output, millions of man-months of suffering, and several million deaths per year.
And the less likely your proposed benefit of mosquitos is, the better it must be to win in the expected value calculation.
As ab9 said, they could be outcompeting other pests that host far, far worse parasites. If mosquitoes were to disappear, these other pests would multiply to take over the niche and we would have something worse than malaria on our hands.
Of course, I'm not sure if there's anything like this. There certainly are other parasites, but I haven't heard of any where the hosts compete directly with mosquitoes. Then again, with something as huge as nature, there is always something waiting to be discovered. For example, we have been in contact with the host species for ebola for millions of years. Yet, ebola first appeared as late as 1976.
