Wednesday, April 23, 2008

Portland Wi-Fi and the Private Provision of Public Goods

On Monday The Oregonian reported on the potential demise of the free public Wi-Fi network in Portland. This is not terribly surprising, as economists have long recognized the public goods problem. I have blogged about the public goods problem before, most notably, perhaps, about the Sellwood Bridge, so I'll be succinct: it is hard to make money from public goods and free markets will provide too little of them relative to the socially optimal amount. The fact that MetroFi's business model was flawed suggests that they forgot to ask an economist about this problem. With free access, it is very hard to figure out a way to make the thing pay for itself. Of course, you could say the same thing about Google and yet they are thriving, but this success rests on the very nature of the web search - the learn a lot about you based on what you search for and can target advertising surprisingly effectively (it turns out). In economics, as a rule, information is gold.

It seems to me that there are two important questions to ask at this juncture. One, is a private, fee-based network the way to go?, and two, is the public benefit of a publicly provided free network larger than the cost?

The first question: is public provision of this public good a good idea or should we just rely on markets? What is interesting about Wi-Fi is that it can be made in to a private good very easily and thus a market solution is possible. The problem with doing this is that it requires a large fixed cost investment in infrastructure (for a city-wide network) and is therefore likely to be only profitable for one company. This is what economists call 'natural monopolies.' Monopolists in general charge more than the competitive market price and provide less of the good. This is why telephone, power and cable monopolies were regulated - to counteract their market power. Market power also rests on the fact that you have a good that has no or few good substitutes and Wi-Fi may have a big one: internet service from cell phone providers. I am not a techno geek and do not know how long it will take for this to become as good as Wi-Fi, but my basic knowledge is that when the digital TV conversion is done the next generation of wireless communication systems will be ready to go to fill the space once occupied by analog TV signals. So it may not even be necessary to regulate even if the network went private. But it may not be profitable at all. If MetroFi has to charge a fee and few people want it, potentially because of competing wireless telecom based services, this business might not be profitable even as a fee based service. And even if it can be profitable, do we want really want a private network?

So this leads to the second question: is public Wi-Fi beneficial enough for Portland to provide it itself? It is possible; it could spur efficiency and productivity, lead to new investment and improve the connectedness and "web-literacy" of Portlanders (and especially kids) - a good skill for the 21st century. But the marginal benefit to free Wi-Fi versus private wireless phone networks is a question - but as many Portlanders would not be able to afford the wireless service, I think this marginal benefit woud be large. There is another factor, which is that the cost to cable and DSL customers should decrease with free Wi-Fi as the demand for these services should fall. I am inclined to believe, therefore, that the overall benefit to Portlanders would be well worth the cost, especially since Portland can amortize the investment cost very cheaply using municipal bonds. In other words, the city of Portland can probably borrow money at much lower rates than MetroFi to finance the building of the network.

So just because MetroFi's management and Portland's city councilors failed to study enough economics should not doom the system, Portland should take the lead and provide Wi-Fi for its citizens.


Dann Cutter said...

I wonder, however, if the argument of public good significantly weights the idea of the Free Rider problem (and gads, considering the source, and the likelihood of one day having you as my instructor again, I hope I get this right).

As I understand it, the free rider problem consists of two issues in this concern. The first, that excessive use of a public resource will exist. The second, that participants will not necessarily shoulder their fair cost of the system.

The first issue regarding excessive usage, from the standpoint of technology, is a real problem. Network access is not an unlimited quantity. Much like a highway, there are only so many lanes and spaces of capacity before unreasonable delays can creep into the system. And again similar to crowded highway systems, there is a real cost, albeit small, for every user with the inclusion of each new connection. Since the service is free, and there is no prohibition to have connections for any and every purpose whether it induces increased productivity or not, people will overuse it. Thus, usage will grow to fill the allowable capacity, eliminating any efficiency in the system (I5 at rush hour). Why?

That is the second part... the wireless network will serve only a small portion of Portland. The claim of 100% does not take into account the realities of wireless networking (many homes and apartments will need a $100 booster). However, Portland is not a closed system, and many users from outside the system and with no municipal tax burden will take advantage of the free system while not contributing to its cost. And while some residents will continue to use DSL/Cable, many will forgo this service causing an artificial alteration to the customer/equipment ratio of market providers possible putting an undue financial burden on a once sustainable business model. By removing market revenue and allowing tax free users, we create a significant usage without corresponding tax based paid access.

Finally, government will tend to not maintain the system efficiently, as it is not driven by market restraints. Thus, the burden to citizens paying taxes will be likely disproportionate to the actual cost of the service. Considering we do have market alternatives which can provide localized demand service (frequently free as an incentive to proximity - Starbucks for example), it seems to make little sense to provide a public provision, running rampantly over cost, without a much closer analysis of the generation of a public good.


Patrick Emerson said...


That is a very good point (and you are correct in your understanding of the free rider problem). I was thinking of Wi-Fi as essentially non-exhaustable meaning my useage does not leave any less for you (think radio). But if it is true (and obviously it is) that bandwidth is used up by each person, then this is something serious to take into account. I wonder how much bandwidth costs? Do you know? In other words, what is the marginal cost of individual usage? To the extent that people will not pay their marginal cost, it will be subject to the free-rider problem.

Good point and a good dent in my argument.

Michael Weinberg said...

I'm the president of the Personal Telco Project, Portland's Community Wireless Networking non-profit. I think this discussion is very interesting, and would love to talk personally with you, Patrick, about your ideas and provide some insight from my perspective.

To answer the question regarding bandwidth costs:

For the purpose of this comment, I'm using "bandwidth" to mean the purchased capacity to access the Internet, and not the technical capacity of the network.

Bandwidth is finite, as any one user's usage will impact the available bandwidth for other users; however, technical considerations/limitations and not cost limitations are far more likely to be an issue.

In general, with a wireless system, particularly a mesh (which Metro-Fi uses for backhaul), the available bandwidth to any single connection is going to be considerably lower than the overall bandwidth at the Point of Presence. So one user's usage should not adversely impact the purchased bandwidth. Similarly, it would take a concerted effort on the part of an individual to adversely effect the technical performance of the network. So the possibility of one or even many "free-riders" hampering others' use of the network isn't really an issue. This assumes, of course, that your free-riders are just using the system, and not attempting to sabotage it.

Internet access is cheap, especially when purchased on the scale of a municipal network (I would guess well under $10K a month). Even a Cable or DSL connection is cheap, when you consider the amount of traffic that it is capable of transmitting in relation to the amount of traffic needed for typical communications (BitTorrent is arguably a different story, but I won't get into that).

I am confident that you are correct in identifying capital costs as the major financial hurdle. The actual bandwidth costs should be small enough to have little impact on the financial viability of the system.

Mike said...

Listen to you guys talk about "free riders" like you're back in economics 101. Metro-Fi is reporting that the network is profitable based on current coverage and speaking from experience it works. Portland gov't has always had unrealistic expectations. My feeling is any money that doesn't end up in the hands of Comcast leaves the public better off. If you want to dicuss an economics issue, lets talk about Comcasts monopoly and the terrible service and high prices they charge.