Bead and the choice of technology
Bowling Green, Kentucky (April 11, 2023) — In the bipartisan Infrastructure Investment and Jobs Act (IIJA) passed in November 2021, U.S. Congress set a course to universal broadband access for the first time. That means that every serviceable location in America would able to order high-speed internet at a reasonably fast speed, such as 100 Mbps download/20 Mbps upload, or at least 25/3.
But what technology or technologies will be used for all this broadband deployment? Well, it remains to be seen and depends on a lot of factors, including decisions by private ISPs and state broadband offices. This post, like others in the series, aims to give state broadband offices good advice as they make their BEAD (Broadband Equity, Access, and Deployment) program plans.
Fiber prioritization in the BEAD program
When it comes to broadband technology, the National Telecommunications and Information Administration (NTIA), the agency authorized by Congress to administer the program with the help of states, chose to provide unusual clarity about what it prefers. The IIJA statute included the term, “priority broadband project,” strongly suggesting — yet not clearly defining — that it refers to end-to-end fiber. In contrast, the NTIA says bluntly that “priority broadband project” means end-to-end fiber.
Past federal and state programs have generally prioritized projects by speed tiers and other performance metrics, while maintaining some claim to “technology neutrality.” For example, the FCC’s Rural Digital Opportunity Fund (RDOF) prioritized projects promising gigabit speeds, but it didn’t specify what technologies should be used to deliver them.
This probably contributed to the RDOF’s auction’s anomalous outcomes, in which large awards were made to fixed wireless providers with promises of gigabit speeds that most experts don’t think fixed wireless technology can deliver at scale. Since then, some of those awards have been disqualified, and in general, doubt remains about whether and how the ambitious deployment promises of RDOF winners will be fulfilled.
The NTIA’s decision to abandon such unreasonable technological agnosticism is a welcome change.
There is a sound engineering basis for having the BEAD program prioritize, not hard-to-check promises of eventual high performance, but the physical technology (end-to-end fiber), which alone is well understood to be capable of delivering the fastest broadband speeds.
But that’s not to say BEAD can only fund fiber. On the contrary, the NTIA requires states to define an Extremely High Cost Per Location Threshold that regulates the degree of fiber prioritization. Further guidance may narrow states’ options, but as far as we know at present, the NTIA might allow states to define a Threshold of $20,000, $10,000, $5,000, or $2,000, though not $0 — at least some fiber prioritization will be required.
But it’s up to state broadband offices to make sophisticated technical determinations, or even just casual judgment calls, about how much their state can afford to spend on end-to-end fiber projects if they want to have enough money to get to the hardest-to-reach areas using the cheapest technology.
To apply the Threshold, states will need to evaluate each incoming proposed broadband expansion project using end-to-end fiber in terms of (a.) the subsidy cost, and (b.) the number of BEAD-eligible locations that will get 100/20 broadband service as a result of an expansion project. Dividing (a.) by (b.) yields the subsidy cost per location connected. Then:
- If the subsidy cost per location is less than the Threshold, no non-fiber projects will be considered. The state will choose among fiber projects using a rubric.
- If the subsidy cost per location exceeds the Threshold, the state will consider non-fiber projects and may fund deployment in the area using a technology cheaper than fiber.
- Which technology outcome should the state favor?
- How should the state achieve its preferred technology outcome by setting the BEAD program rules?
The higher the Threshold, the more fiber prioritization. So, states should think hard about how far up or down to turn this policy dial.
‘Reliable’ broadband technologies other than fiber
Three other technologies, defined as “reliable broadband technologies” in the BEAD Notice of Funding Opportunity (NOFO), are next in line for consideration if no fiber projects below the cost threshold are available. They are DSL, cable, and fixed wireless using licensed spectrum.
Though defined as “reliable,” DSL probably can’t achieve the BEAD speed requirements of 100 Mbps download/20 Mbps upload at scale. The key factor determining speed is distance from the DSLAM, and the way speed deteriorates with distance makes fast DSL uneconomic in very rural areas. DSL providers are more likely to be overbuilt than to win BEAD money to extend their networks. Their best bet is probably to overbuild themselves using a better technology such as fiber.
Cable, also known as “coax” because of the coaxial copper wires it uses to carry signals, or sometimes “hybrid fiber-coax” since a coax last mile links up to a fiber backhaul for better long-haul capacity and speed, can achieve 100/20 speeds, and sometimes much exceed it on the download (less so on the upload) side.
Many areas will be excluded from BEAD eligibility because they are adequately served by cable. Cable companies continue to expand the footprint of their coax networks in some areas. Nonetheless, broadband expansion using cable technology will probably play a small role in the national push to close the Digital Divide.
Where the legacy cable TV infrastructure is available, cable broadband at speeds above 100/20 has generally been deployed already. Most unserved and underserved areas lie beyond the reach of cable TV networks that were built out decades ago, and large-scale network expansions are usually carried out with end-to-end fiber-optic cable rather than any kind of copper.
So in practice, the main alternative to end-to-end fiber for new deployments at the network edge is fixed wireless. Fixed wireless can usually be deployed faster than end-to-end fiber, and with much less upfront capital expenditure. It also has the advantage of blanketing an area with coverage instead of delivering it only to specific locations such as homes and businesses. That’s of limited importance, however, since most people don’t want to live off the grid, and for those who do, low-earth orbit satellite broadband from Starlink is now usually a viable option.
Fixed wireless requires line of sight, more or less, and is subject to signal blockage by slopes and even foliage. Some people living in wooded areas report that coverage deteriorates in summer when the trees have leaves, relative to winter when they are bare. In treeless areas, foliage is a non-issue, which can favor fixed wireless. The NTIA’s new requirement that fixed wireless ISPs use licensed spectrum in order to be considered “reliable” will raise ongoing costs somewhat, while protecting consumers against certain service interruptions.
All in all, there’s a pretty strong consensus that fiber-to-the-premises, with its far greater speed and data capacity, is the superior technology solution for the connectivity and economic development of a region, if the necessary capex can be mobilized somehow. But fixed wireless is a much more cost-effective solution in many rural areas, at least in the short run.
In the remotest, hardest-to-serve locations, it may be hard to get any “reliable” broadband technology to deploy, and the NOFO suggests that states might fall back on still less-favored technologies like satellite and fixed wireless using unlicensed spectrum. But that’s a topic for another post.
The subsidy costs of fiber vs. fixed wireless
In post 1, post 2, and post 3 of this series, I developed an abstract model of a BEAD grant project that is useful for understanding choice of technology in the BEAD program. Figure 1 represents a case that’s likely to occur in 2024 as states roll out their BEAD grant programs and seek to find projects to support.
The area in question has both a fiber provider and a fixed wireless provider willing to deploy. The fiber project has higher capex needs than the fixed wireless project. In this case, as will likely be true often, the fiber project has lower opex. It can also look forward to higher revenue from customers because the higher quality of service increases their willingness to pay. Part of this might be due to settlement in the area by high-income teleworkers, who need very fast and reliable connectivity.
In the example shown in Figure 1, the willingness to provide private matching investment is substantially higher on the part of the fiber provider than the fixed wireless provider. This reflects both higher anticipated revenue and lower anticipated opex. Nonetheless, the difference in private willingness to invest is far smaller than the difference in capex, so the required subsidy is much greater for fiber than fixed wireless.
Two questions arise:
To the first question, while the fiber project is better, it’s so much more expensive that fixed wireless is probably the better choice. The extra money spent on the fiber project is somewhat wasted, in the sense that the area residents who benefit from it would rather have gotten the money than the fancy fiber-to-the-premises broadband.
Of course, the BEAD program can’t just hand out the grants as cash to the citizenry. It has to spend the money on broadband projects. And so, if the state thinks it’s on track to get everyone broadband service anyway, it might see fit to upgrade an area to super-fast fiber because the state is so flush, even if its residents don’t value the upgrade.
Still, there are probably better options. For example, the state could negotiate lower prices for consumers or higher wages for workers, while still using fixed wireless technology to serve the area shown. Digital skills training programs are another allowable BEAD spend after the state can show that it’s on track to achieve universal broadband service.
If the state wants a fixed wireless solution for this area, it can get it simply by not setting the Extremely High Cost Per Location Threshold too high. If the Threshold comes in above the required subsidy, then the fiber provider will bid under the Threshold and get prioritized over any non-fiber competitor. But set the Threshold lower, and the fiber provider will have to breach it, putting it in competition with the fixed wireless project, which will probably win on cost unless the rubric strongly favors fiber.
Fiber for sustainability?
In Figure 1, fiber is a luxury that’s not worth the money. Tweak the bar lengths — e.g., lower capex for fiber, higher opex for fixed wireless, a bigger difference in consumer willingness to pay, etc. — and you can easily get a case where the fiber project is worth it. But there’s a more counterintuitive case where fiber wins, even though it’s not worth it, because the front-loaded nature of BEAD subsidies excludes fixed wireless projects. Though cheaper, these projects need more subsidies for opex rather than capex, which the BEAD program can’t accommodate.
To see how that works, consider Figure 2. As in Figure 1, there are feasible projects to serve an area with fiber and fixed wireless. Again, the fiber project involves far more capex, resulting in lower opex and higher customer revenue because of the higher-quality broadband product that fiber makes possible, but the advantages of fiber fall far short of justifying the high capex, even when consumer surplus is taken into account. The difference is that the fixed wireless project is not commercially sustainable.
What happens next in the scenario in Figure 2 depends on how the state enforces rules for restricting the use of grants funds to capex costs. Likely, the fixed wireless ISP won't bid, and the fiber project will win unopposed. But an optimistic fixed wireless ISP might apply anyway, win on cost, operate for a while, and then shut down because revenue doesn’t cover opex. State broadband offices could make this less likely by strictly monitoring subgrantees’ expenditures and making sure they’re limited to bona fide capex.
To conclude on a very nerdy note, Figure 2 represents a situation of “dynamic inefficiency,” where a too capital-intensive solution is used because the institutional basis for sustaining a cheaper, less capital-intensive solution is lacking. States could fix this by instituting a system of universal service subsidies for broadband, modeled on those that have long supported rural landline telephony. But state broadband offices probably have their hands full without designing and advocating that sort of legislation right now!
About the Author: Nathan Smith is the Broadband Policy Specialist. Dr. Smith monitors federal broadband policy, writes public comments for federal agencies that request advice on broadband policy implementation, and helps with business development and proposals.