Much the same as calculating the likely cost of an infrastructure project (think HS2), and then setting the assumptions about what savings will result in the future from the project, until the benefits outweigh the costs, at least on paper.
A process called "Cost/Benefit Analysis" that I learned and applied to "feasibility studies" for grossly elaborate and costly airport projects around the world, when it was well understood that the third-world client wanted the CBA to prove that the new airport, named after the President, should go ahead, financed, natch, by friendly powers extending their area of influence. My favourite variable was "Cost of business travellers' time" which could easily be increased as needed. It amused me hugely when I read that this particular variable has been increased by 1000% since the first estimate, so as to justify the waste of money that is HS2.