[lisper]

> how can any private ISP compete with that?

If the answer to your rhetorical question is supposed to be "they can't" then government-provided broadband is manifestly superior to anything the private sector can provide. Why would we as a society not want that?

[JoshTriplett]

It's not necessarily superior; it's subsidized and doesn't have to actually make money, or even break even. Ultimately its costs can be hidden away in taxes or debt, where they're still present but unseen. Municipal broadband also often gets fast-tracked or special-cased in regulation; governments don't do a good job of regulating themselves.

Private ISPs can't do any of those things; they actually have to pay for infrastructure, follow regulations in installation, etc.

Personally, I'd prefer to see either community-run pseudo-ISPs ("we laid some fiber and contracted for bandwidth"), or fiber made available as infrastructure but the bandwidth handled via the market (much easier to have healthy competition).

[vidarh]

Historically, most places with government owned telco's starting ISPs still had arms length regulation. If that was really the concern, looking at the dozens of countries with experience in regulating this would be rather simple.

That said, I agree that you'd get far just by providing the last mile. In the EU, deregulation basically required the owners of monopoly last mile infrastructure to separate them out into separate business units that are heavily regulated to provide equal service. Nothing prevents providers from choosing to build their own, but this ensures "anyone" can start an ISP.

E.g. in the UK, BT separated out OpenReach - not only are they restricted to offering same service and prices to everyone, their wholesale prices are published on their web pages for everyone to see.

There are still problems - e.g. BT is often accused of milking OpenReach rather than investing in the network. But those problems could have been solved by regulating profit-taking so that profits can only be taken as a proportion of investment.

OpenReach offer both local-loop-unbundling (IPSs put equipment in BT exchanges and handle backhaul themselves, and get a raw copper or fibre connection to the subscriber) and backhaul where ISPs connect to BT one or more places and get a raw IP connection to the subscriber.

It works ok, with the caveat above, and it also doesn't stop alternatives, like FTTP from other providers, in areas where it is economically viable to lay new networks.

[rayiner]

European unbundling rules mostly don't apply to cable or fiber, which are the dominant modes of broadband access in the U.S.: http://www.oecd.org/sti/broadband/2-7.pdf.

> There are still problems - e.g. BT is often accused of milking OpenReach rather than investing in the network.

The caveat is perhaps the opposite of what you suggest. Unbundling works okay in the U.K. because BT OpenReach has been allowed to be quite profitable. (Their profits as a percentage of revenue are between TWC and Comcast.) Unbundling in the U.S. failed, in contrast, because the FCC set wholesale rates so low there was basically no incentive to invest in DSL networks.

It's my theory that stuff that works in other countries fails in the U.S. because everyone is so ideological. In the U.K., the privatization of BT was preceded by tons of economic studies analyzing how to balance broadband availability and price against the need to make investing in broadband a profitable endeavor. In the U.S. it's entirely ideological. Pro-business free marketers on one side versus people who see broadband as a social justice issue, economics be damned. We lurch form one mode to the other based on who happens to be in power.

[vidarh]

It mostly does apply to fibre, where the fibre is installed as part of upgrading the dominant carriers network, which is to say: almost always.

Where other providers lay new networks, you're right, it doesn't apply, because when other providers lay new networks it is generally evidence there is an actual competitive situation. In the UK most of the fibre deployment is done by OpenReach, with Virgin Media second, and tiny ISPs focusing on high density urban deployments a distant third onwards.

> The caveat is perhaps the opposite of what you suggest. Unbundling works okay in the U.K. because BT OpenReach has been allowed to be quite profitable.

It works ok but could work much better. The problem today is that there is very limited reason for BT to upgrade. They've had to be dragged kicking and screaming into fibre upgrades, and have been given quite a lot of funding to incentivise it. If the UK had tied BT's ability to take out dividends from OpenReach to the amount invested in infrastructure improvements we could have done much better. In that case I'd even support allowing them to set higher profit margins. As long as it'd be matched with higher investments.

> In the U.S. it's entirely ideological. Pro-business free marketers on one side versus people who see broadband as a social justice issue, economics be damned.

But irony is that the European approach has been much more pro-business and free market than the US approach. The US approach is the one that has ended up with massive regulatory capture, but on behalf of the largest ISPs, where the European approach has centered on isolating and minimizing the un-competitive part of the business as much as possible exactly to ensure there's a free market competing over the rest.

[remir]

Perhaps something like a coop would be optimal? The coop would have to turn a profit, but then it has to be transparent in its management because in the end, it must serve the interest of its members.

[JoshTriplett]

Agreed completely; I'd love to see this.

[deusofnull]

So yr ideal example is kinda like public roads (fiber) and private trucking (bandwidth). I think that's a neat solution tbh. Especially if the bandwidth suppliers are properly taxed for use of the public fiber, tax revenue which then could be used to maintain and expand the fiber infrastructure.