Population density and broadband economics
The economics of wireline broadband deployment are fairly straightforward, yet acknowledgment of this reality is usually painfully absent in broadband policy discussions. Much of the cost to rollout or upgrade a broadband network is driven by the miles of outside plant required, while the revenue opportunity depends on the number of homes passed. Simply put, densely populated areas are more profitable than rural areas and there is a cut off point below which broadband services are not financially viable without an external economic lifeline, such as a government subsidy.
Data released yesterday in a report from speedmatters.org illustrates this well. They compiled data on actual download and upload speeds from each state based upon 413,000 speed tests performed over the past year. We took this data and combined it with geography, population, and income data from the US Census to perform some quick and dirty analysis.
The scatter plot below shows population density (in people/square mile) vs. average broadband download speeds (in kbps) for all 50 states:
Higher broadband download speeds are delivered in states with higher population density
We also looked at median household income and “income density”, but these relationships proved to be not as interesting (income density = household income * # of households / square miles).
As you can see from the chart, the top 6 most densely populated states have much better broadband performance than rest of the country. New Jersey, Rhode Island, Connecticut, Massachusetts, Maryland, and Delaware all have population densities over 400 people per mile and average 8.7 Mbps in download speed as a group. The average download speed for the bottom six states, however, is only 3.3 Mbps. These least populated states all have under 20 people per mile: New Mexico, South Dakota, North Dakota, Montana, Wyoming, and Alaska.
From my first hand professional experience as both executive and advisor, including recent work on getting broadband stimulus funding, I can tell you that the business case for serving 20 people per mile with DSL or cable is pretty ugly!
In spite of this basic economic fact, public policy debate usually centers on how we are falling behind other countries, and how the problem is corporate greed, monopolistic cable or telco power, or some other nefarious culprit that can be legislated away. According to Speed Matters’ report, the key benchmark countries we are falling behind are:
- South Korea (1,260 people per mile)
- Japan (870 people per mile),
- Sweden (52 people per mile)
- The Netherlands (590 people per mile)
- Germany (600 people per mile)
Of these, only Sweden is a fair comparison to the US as a whole, which has 80 people per square mile. I submit to you that the difference in performance between the US and these leading broadband countries is primarily based on real economic costs, not corporate greed. It is much cheaper on a per household basis to deliver broadband to these markets than it is in the US.
[...] In our last post on broadband policy, I highlighted the six least densely populated states: New Mexico, South Dakota, North Dakota, Montana, Wyoming, and Alaska. These states all have fewer than 20 people per square mile and actual broadband performance that is demonstrably lower than the US average. Of these laggards, only Alaska makes the cut in the top 10 for total funding requests, but four out of the six are represented in the top 10 on a per capita basis – and a fifth, Montana, is ranked 11th. [...]
[...] We agree 100% with a couple of the key points that Scott makes: The rural portions of Alaska are in, in fact, the most remote and unserved in America (see our earlier post on population density and broadband economics) [...]