I'd expect that a woman surviving would lead to more kids, but not actually 2 more, and similarly a missing man wouldn't just be replaced by the nearest available sperm-producer. I dunno how to put a number to it.

One way to start estimating it would be to correlate local sex ratios with local birth rates and try to control for as many things as possible. Unfortunately, this is probably very hard to do...

In an ancestral environment close to equilibrium (what you imply by saying that each person has 1 kid on average), the situation is even more egalitarian.

I'm actually most interested in the answer for modern poor countries, which are neither stable in population nor Malthusian. Basically, I'm wondering how interventions that save lives of one gender (but not the other) today will affect the population size 20 to 30 years in the future. Non-replacement fertility doesn't qualitatively change things: the question just becomes whether a life saved increases the population by more or less than "next generation's size / current generation's size". Replacement fertility is just the special case where the ratio is 1; I used that number in my question only for simplicity.