[rayiner]

Stokab ultimately achieved 90% coverage by building out in a demand-driven way (though it took 16 years, compared to the 3-5 year timeline typically imposed in the U.S.). Which is how everything else works: at first only rich people had smart phones, now everyone does.

Universal service is an admirable goal, but the implementation is economically nonsensical. It is, in essence, a tax on telecom service. In the case of the universal service fund, it's an explicit tax. In the case of build-out requirements, it's an implicit tax: instead of taxing an ISP in cash, you're compelling it to provide a service it would not otherwise provide.

But it makes no sense to impose industry-specific taxes on telecom service. It's not like gasoline or alcohol, which warrant taxation to discourage the negative externalities they create. You tax industries that you want to discourage, not ones you want to encourage.

The downside of this economic distortion is that it suppresses competition and entrenches incumbents. There is no "minimal viable product" if you're starting an ISP in Baltimore. There is no sniping away at an incumbents' highest-margin markets. You end up not being allowed to use the tactics that startups in other industries use to take on incumbents.

Contrast how Sweden approaches broadband deployment in rural areas. It simply gives rural residents a tax credit to subsidize construction of a fiber line. It's the same way we approach other forms of welfare. We don't require Whole Foods to build stores in poor neighborhoods. We give direct aid in the form of SNAP and WIC benefits.