About a decade ago I looked into moving some COBOL components off-mainframe (either as COBOL-on-Linux or a rewrite into Java, which itself is really COBOL Mk II at this point), and your point about the operating environment is one of the key elements, but not all of it; there's also the fact that the first big shift to automation, via mainframe assembler and COBOL, is when companies sacked a lot of the folks who knew how and why the pre-automation processes worked - that knowledge exists in the mainframe code and the heads of the people who work(ed) on it, and nowhere else. A rewrite or a replatform is very, very hard and risky as a result; the system is now defined by how the mainframe runs the processes, to a very large degree.

The third is that COBOL is only the tip of the iceberg. As soon as I spent time learning about the code I was being asked to look at, you get into decades of evolving programming practises. Modern COBOL is multithreaded, probably uses DB2 and relational datamodels. COBOL from thirty years ago is probably single-threaded, only runs right on high-clocked single-execution models, cuts down to hand-written s390 assembler regularly, and uses VSAM files with non-relational data. Older code still will be sharing data simply by banging it into memory regions for other code to read out of, because that's how you got performance back in the day.

Trying to identify how you'd pull a function out of that and move it off is somewhere between extremely difficult and impossible. It's usually so complicated and expensive it's easier to try and hire people who want to apprentice as mainframe programmers and keep the current codebase running.