A cost-benefit analysis, or a risk-benefit analysis, is only as good as its data and its process. Data-cropping will absolutely ruin results. Failure to account for the financial cost of lost lives, for example, is one such error. Insurance companies, government disaster departments, and other such groups have presented a price-per-head. Medical industry usually thinks in terms of "a year of healthy life." These can be very useful because environmental degradation both ruins health and ends lives.

Frex, skyrocketing heat is changing acute kidney damage, which used to be fixable with fluids and maybe a few days in the hospital, into chronic kidney failure that won't go away. That's expensive. As another example, a clearcut forest generates a profit only once from one product, logs. Sustainable forestry takes more effort, but provides more jobs over the long run, and produces many products from logs and nuts to hunting, fishing, and photography. The latter is more profitable, but you have to know what you're doing to capitalize on all those different income streams.

A lot of times, people look at profits of mistreating the Earth, and the costs of respecting it; but they forget about the costs of mistreating it and the profits of respecting it. Slovenly accounting makes for slovenly science makes for bad business decisions.

However, if you're good at math and science, then you can crunch those numbers and use them to convince people or shred your opponents' arguments. It's not hard if they're doing a half-assed job in the first place. Some predictions are quite solid, others are flimsy.