I don't want to spend too much time on it now because I know the book (and thus I) will get into this in more depth later, but the basic idea is that anything that raises the death rate improves the standard of living because it keeps the population smaller. Disease happens to be the only thing that gets worse as civilizations become more advanced. I suppose you could argue crime and war, but war kills mostly through disease, kills mostly men and thus doesn't affect the birth rate much, and neither have such an obvious tipping point the way disease does.

Note that the timeframe here is ~1300-1800 AD, although he occasionally talks about earlier.

One thing that bothered me about this hypothesis was that I thought China, Japan, ancient Greece, and ancient Rome had a fairly good standards of living. I haven't actually verified the citations, but assuming they're valid, Farewell To Alms shows that the average per capita income of ancient Chinese peasants, as measured by the number of calories they could afford, was actually lower than the equivalent European peasant. One possibility is that the larger a society gets, the wider the income variation it sees, so China and Japan, which were large unified kingdoms, could support a more sophisticated and complex upper class, and I was simply more aware of Medieval European peasantry than I was Asian or Classical European peasantry.