This is wrong because subsidies aren't just externalized costs. They are defined as monetary gifts. You're extending the definition by analogy and arguing that a company externalizing its costs is the same thing as a subsidy, but a subsidy implies direct and deliberate support.

This is important for a couple of reasons: The main one is that it's confusing to redefine subsidy when you can just say "cost" and have people understand you. People reading this survey might see the $5.2T number and assume that that cost is in addition to whatever the cost of climate change is and will have to read the paper to understand otherwise. This is unnecessarily confusing even if one were to grant the logic of it.

In addition, when people discuss subsidies, they are often most interested in government policy. The purpose of a subsidy often is to increase a certain kind of business, so we might worry about unnecessarily funding the fossil fuels and thus encouraging climate change in a manner above and beyond simply allowing them to be used the way we've always done, but that's not quite what's going on here.

One thing to keep in mind is that it's a lot easier to measure a genuine government subsidy than an externality. So the distinction matters in that regard as well. Any measure of the cost of climate change is to at least some degree speculation, whereas any attempt to measure the direct amount of money given to fossil fuel companies can probably be much more exact.