Suppose the following experiment. A town has some open space and can choose to develop it following the pattern of the existing town or attract a college campus. Is the campus worth it? On the plus side, it should attract more educate residents and bring amenities the town does not have to pay for. On the negative side, and that is a big negative, the town will get no taxes directly from this college (the recent attempt of Pittsburgh to tax tuition has failed, sort of). And college students tend to be rowdy.
Donald Vandegrift, Amanda Lockshiss and Michael Lahr perform the kind of hedonic regressions that have been used to value various amenities and add college dummies. It turns out size matters. A college raises house values by 11%, but larger enrollments reduce house value. And in any case, the presence of a college increases the tax base by 24%. While college towns will always complain about the non0-taxable real estate, they should be happy about the general equilibrium effect from colleges, which are strongly beneficial.