I have had a little bit of professional experience with the types of "impact models" (Input/Output analysis, economic impact modeling, whatever you want to call it) that people love to use to justify things like sports stadiums, olympics, investments into various similar things...
Actually bought and used the software/data that most people use and tried to replicate and tweak some analysis ourselves.
Our overall conclusion was that they are pretty much 100% bullshit. The guy who came up with them got a nobel prize, but either his work has been perverted by people trying to justify projects, or his work was simply wrong (I'd like to think it is the former...I think his work was originally done at the level of trade between nation-states, not the level of building a stadium in some city).
Two big things stuck out to me:
First, is that the people who do these impact analyses basically never have their work checked. You hire a consultant to make a number, you pitch your product, build the thing, and then forget all about the guy who ran the model. Nobody goes back 5 years later and says "did these numbers actually work?". IIRC, one of our better pieces of evidence was some researcher who actually did go back and check the models for the Beijing olympics--of course finding that it didn't come true at all and actually worked backwards (mostly because non-olympic tourists avoided beijing like the plague for most of the rest of the year).
Second, these models literally cannot create a negative number. They are basically just tables of multipliers. If you spend 100 dollars, that will mean 75 dollars in the hands of local workers, which in turn puts 50 dollars in the hands of local shops and restaurants, which in turn puts $30 dollars in the hands of local distributors, who pay their workers $20, and the cycle repeats. Now your $100 investment has generated almost $200 in revenue for your town. None of the multipliers are negative. A sports stadium can never mean that on game day, more people actually avoid the neighborhood and the community loses money.
They have some value as a predictor of where money might go in a functioning economy...but for actually justifying a major NEW project (and thus disrupting the local economy that the multipliers are based on), I am highly skeptical.