Narrowing the Rural Digital Divide with Consumer Vouchers

Rural broadband became an issue of national conversation in January 2018 when President Trump signed an executive order instructing agencies to prioritize pro-deployment policies in rural areas. Further, recently in the wake of nationwide school and business closures in 2020, millions of Americans have found themselves more reliant than ever on home broadband service in order to engage in remote work or distance learning. This has made measures to close the rural “digital divide”—the gap between wealthier, urban households and poorer, rural households in broadband service quality—more urgent for policymakers and industry members. Congressional leaders and the Trump administration, for instance, have identified rural broadband support as a priority of proposed coronavirus relief legislation.

Congress has debated and legislated rural telecommunications funding before. The Telecommunications Act of 1996 added a statutory “universal service” objective that the Federal Communications Commission (FCC) ensure that telecom services in rural areas (1) “are reasonably comparable to those services provided in urban areas” and (2) “are available at rates that are reasonably comparable to rates charged for similar services in urban areas.” Congress required interstate telecommunications providers to set aside a portion of their revenues to subsidize, according to FCC guidance, telecommunications in rural areas consistent with these statutory objectives.

Since 1996, the FCC, regardless of which party has led it, has determined that rural areas don’t have reasonably comparable services to urban areas. In 2018, for instance, 98 percent of urban households had access to fixed, high-speed broadband (typically via coaxial cable or fiber-optic lines), but only 78 percent of rural households had access to such services. The rural digital divide persists, even though efforts to close it have received bipartisan support and substantial public outlays. Over two decades, the federal government has granted more than $100 billion in subsidies to telecommunications and broadband providers in high-cost areas under the Universal Service Fund’s (USF’s) Connect America Fund program, also known as High-Cost Support.

We argue that the structure of rural broadband support programs merits change. The digital divide, while shrinking, requires a new approach to minimize wasteful overhead spending, complex eligibility requirements for providers, and substantial inequities in fund disbursements between similarly situated states and regions. Congress has delegated funding decisions to the FCC, which has delegated disbursement decisions to the private Universal Service Administrative Company, which has relied on a proprietary model developed by another private firm to define major program funding mechanisms. These convoluted subsidy programs, which have built up over 25 years, defy accountability to the public and government auditors.

As an alternative to directly subsidizing telecom providers in rural areas through these complex mechanisms, we propose that the FCC convert most High-Cost Support into a broadband voucher program for rural households. Essentially, the rural telecom subsidy programs should resemble Lifeline, the long-running program which credits $9.25 toward every eligible low-income user’s monthly smartphone or phone bill. By converting High-Cost Support into a voucher program, tens of millions of rural households could receive a substantial monthly rebate on their broadband and wireless bills. Using today’s High-Cost Support budget to illustrate the principle, we show that every rural household in the United States could receive between $5 and $45 per month ($60 to $540 annually) depending on how High-Cost Support funding is currently disbursed among the states. If participation by eligible households resembles participation rates in Lifeline, these rebates could be higher.

This rural broadband voucher idea has circulated for years. Rural broadband vouchers have proved popular in the United Kingdom and are gaining attention in the United States. The idea was recently endorsed, for instance, by Pennsylvania’s Communications and Technology Committee. Former FCC Commissioners Mignon Clyburn and Robert McDowell, a Democrat and a Republican, respectively, in May 2020 called for a temporary $100 voucher program for households struggling to pay bills because of the COVID-19 pandemic. Telecommunications experts Blair Levin and Larry Downes endorsed a similar temporary voucher program that same month. Finally, a novel broadband policy was recently announced in Arkansas that resembles a broadband voucher: in July 2020, Governor Asa Hutchinson created a statewide program that would use $10 million in federal funding from the Coronavirus Aid, Relief, and Economic Security (CARES) Act to provide wireless hotspots (and internet access) to up to 20,000 homes with children for two years.

Rather than propose a new law and new funding, we apply this idea using today’s High-Cost Support budget and current law as constraints. First, we summarize the state of the rural-urban broadband deployment gap. Second, we highlight deficiencies with the current subsidy programs, particularly their complex structure, perverse incentives, and inequitable disbursements. Finally, we illustrate a simple consumer voucher scheme using today’s High-Cost Support subsidies. Though a voucher program would represent an abrupt policy change, it would accomplish statutory goals while being less costly to administer and more transparent than today’s broadband subsidy programs.

The State of the Digital Divide

Rural broadband deployment has improved considerably in the past few years. The FCC tracks the number of people in rural and urban areas who lack options for fixed, high-speed broadband (25 megabits per second or greater). In 2013, according to the FCC, most Americans living in rural areas in the 50 states—nearly 32 million people—had no fixed, high-speed options. By 2018, those numbers had shrunk to just 22 percent of the total rural population in the 50 states—over 14 million people—with no fixed, high-speed options. As the FCC acknowledges, the agency does not measure how much of this coverage expansion is caused by the disbursement of High-Cost Support funding; much of the broadband expansion into rural areas in recent years has been from unsubsidized cable operators.

Nevertheless, high-speed buildout in rural areas lags behind the buildout occurring in urban areas. In 2013, for instance, only 8 percent of urban residents in the 50 states and the District of Columbia—just over 20 million people—had no fixed, high-speed options. At the end of 2018, only 2 percent of urban residents—about 4 million people—had no fixed, high-speed options. That gap—22 percent in rural areas having no options versus 2 percent in urban areas having no options (or about 10 million more rural people having no options, in absolute terms)—is the rural digital divide.

In total, about 6 percent of US households lack fixed, high-speed options. However, bringing landline services (namely, fiber optics) to those final few unconnected homes raises costs exponentially, likely exceeding $30,000 per subscriber. According to 2017 FCC estimates, providing fiber broadband to the last 2 percent of US households would cost $40 billion up front and require $2 billion of subsidies annually for operational costs. Since the Telecommunications Act of 1996 instructs the FCC to ensure that telecommunications rates and service options in rural areas are “reasonably comparable” to those of urban areas, the FCC will continue to disburse billions of dollars annually to narrow that urban-rural gap. Before spending those billions, the FCC should consider whether existing programs are up to the task of bringing broadband to the most rural households in a fiscally responsible way.

Deficiencies with High-Cost Support

While the COVID-19 pandemic has brought the digital divide into the limelight, a tremendous amount of public resources, over $100 billion, has been spent to connect rural households to telecommunications service over the past two decades. Under High-Cost Support, the FCC distributed about $4.8 billion dollars to providers in 2018 with the aim of stimulating rural broadband deployment, about 30 percent more than the $3.7 billion in High-Cost Support funding disbursed in 2001, adjusted to 2018 dollars. While drafters of the relevant statutes believed subsidies would eventually decrease, deployment costs rise exponentially as the percentage of households served increases, which makes the goal of reasonably comparable services and rates in rural areas extremely costly to achieve.

Inefficient Funding Design

When Congress passed the Telecommunications Act of 1996, Congress left the particulars of the USF to the FCC. The FCC chose to require telecommunications providers to pay into the fund on the basis of the amount of their long-distance phone revenues. As a result, USF programs currently suffer from a puzzling and counterproductive funding design: retail rates subsidized by long-distance telecommunications revenues.

This funding design extends what had occurred before the AT&T breakup: regulators reducing local phone rates by shifting long-distance revenues to local phone companies. As Milton Mueller points out, this cross-subsidization policy was more an accident of history driven by political realities than a coherent universal service policy. State lawmakers, who regulated intrastate phone rates that were highly visible to consumers, had more success in driving down local phone rates than the FCC, which regulated interstate phone rates, had in driving down long-distance phone rates. In the years preceding the AT&T breakup and the Telecommunications Act of 1996, local phone services were increasingly subsidized by long-distance revenues.

Excessive Complexity

High-Cost Support has numerous moving parts in its funding model. When the FCC introduced the Connect America Cost Model a few years ago, for instance, which represented a major change to the program, the commission noted that the model is “considerably more complex . . . and relies on more extensive data sets” than prior funding models. According to FCC documents, modelers participated in workshops for 10 months to incorporate “twenty-eight topics related to economic and engineering assumptions and input values” into the initial Connect America Cost Model. Constructing the model took an additional 16 months and 11 further iterations to complete. Furthermore, elements of this model are nonpublic and subject to protective orders, which makes independent review and validation difficult.

The FCC’s bespoke funding model incorporates highly specific inputs—occasionally modified at the level of individual rural telecommunications companies—which drove the more than two years of development, derivative models, and regulatory overhead described later. The result is a model that, per the FCC, includes “assumptions about literally hundreds of individual inputs.” This complexity in design drives some of the inexplicable disbursements and overhead waste.

Inequity in Disbursements

There is currently substantial variance in the distribution of High-Cost Support funding among the states. Given the other weaknesses of the program, this variance suggest a distorted funding system where providers serving a small number of households receive a disproportionate amount of subsidies. Some variance in benefits is to be expected given differences in population density and geography among the states. However, a handful of states see huge amounts of High-Cost Support funding flow in.

For instance, Alaska received over $2,100 in High-Cost Support funding per rural household in 2018. By contrast, Rhode Island received $0.76 per rural household in 2018. This disparity exists despite the fact that increasing the total amount of subsidy support per household has little effect on overall penetration rates.

These excessive disparities are likely caused by the existing subsidy model formulas, which incorporate the costs to provide landline service in low-population areas. Other inequities abound. Pennsylvania residents, for instance, have in recent years seen $800 million in federal USF fees (on net) leave the state despite Pennsylvania facing significant rural deployment challenges. In 2018, less than half of the estimated $384 million in USF fees collected from Pennsylvania residents was used to support Pennsylvania telecom providers participating in the four USF programs, including High-Cost Support.

States with similar numbers of rural households and similar geography have hard-to-explain discrepancies in High-Cost Support funding. For instance, Alabama has about 34 percent more rural households than Mississippi and similar geography, yet the providers serving rural Mississippi received almost two times as much total High-Cost Support funding as the providers serving rural Alabama in 2018. Similarly, New Hampshire has about 30 percent more rural households than neighboring Vermont, yet providers in Vermont received over 60 percent more total High-Cost Support funding for rural broadband. Table A1 in the appendix shows our estimate of rural USF support per rural household for all 50 states in 2018.

Excessive Overhead

These disparities do not necessarily mean that customers in states such as Alaska, with greater amounts of subsidies per rural household, are benefiting meaningfully more than those in states such as Rhode Island, with low subsidy-to-rural-household rates. Much of the subsidies never result in increased deployment for rural customers because the subsidies are spent on regulatory compliance and firm overhead. The design of High-Cost Support encourages significant waste and excess, despite regular efforts by the FCC to reform and restructure the program.

Scott Wallsten estimates that of all High-Cost Support subsidies between 1998 and 2008, 59 percent went to eligible providers’ “general and administrative expenses.” With expenditures under High-Cost Support totaling around $45 billion over those 10 years, that means about $26 billion was directed to overhead expenses, not to capital upgrades and network expansion. Even the estimates of spending on capital improvements should be viewed skeptically. Three studies of carriers receiving High-Cost Support funding in the 1990s and early in the following decade indicate that carriers were inflating their fixed costs in order to receive more subsidies. This “gold-plating” technique—using public funds to make eligible but unneeded network upgrades—is difficult to identify and estimate. Poor oversight and a lack of measurement for analysis of program effectiveness remain problems with contemporary USF funds.

Broadband Vouchers Put Consumers in Charge

An alternative mechanism for the federal government to encourage broadband adoption in high-cost rural areas is to offer broadband vouchers that households could use as an internet coupon. Subsidizing consumers, rather than providers, would better align the program with economic theory and the principles of public finance. Three factors are relevant for this proposed change:

  1. high-speed broadband is near universal, exceeding 90 percent coverage of US households;
  2. the costs of extending fiber broadband to the last small percentage of households rise exponentially; and
  3. wireless solutions such as fixed wireless, cellular, and satellite broadband are increasingly a substitute for wired broadband.

In practical terms, vouchers reduce the risk of disbursements being used for gold-plated services or being sunk in inflated operating costs. Vouchers also allow consumers to purchase cellular, fixed-wireless, and satellite broadband subscriptions, which provide service in rural areas at much lower costs than fiber-optic broadband but have been historically excluded from FCC programs to serve high-cost areas.

We illustrate how a tiered voucher system could be established, using the High-Cost Support disbursements across the United States in 2018 as a benchmark. Rural households in states that have received the most support historically (on a per-rural-household basis) receive the biggest coupons in our proposal. Historically, High-Cost Support subsidies have been based in part on “ruralness” and actual costs of buildout (i.e., areas receive more funding as the geographic distance between subscribers increases). Any rural telecommunications subsidy plan needs to recognize that rural households in a state such as Montana cost more to serve on average than rural households in Connecticut, and our tiered program does so by classifying today’s high-subsidy states in tier 1 and today’s low-subsidy states in tier 6. Also, there are statewide reliance interests that would be undermined by a sudden move toward strictly equal support among rural households across the states. The disparities we discussed earlier cannot be eliminated overnight, but in time some of the more puzzling disparities (like Alabama vs. Mississippi) should be smoothed.

Basing our calculations on the total High-Cost Support disbursements of $4.6 billion to the 50 states in 2018 and assuming 100 percent uptake among rural households in each state, we find that rural households in Alaska, Kansas, Montana, North Dakota, and South Dakota could be eligible for a monthly discount of up to $45, and rural residents in the remaining tiers of states would be eligible for $30, $19, $13, $9, and $5 monthly coupons. We arrive at these amounts by grouping like states together (based on historical disbursements) and stipulating that total voucher program costs would be equal to or slightly less than the current High-Cost Support budget. Table 1 shows the proposed tiers and states eligible for each tier.

A $5 monthly coupon may not seem like much for rural residents of the lowest-tier states (mostly small states in the Northeast), but the $39 million those states receive under our plan (again, if uptake equals 100 percent) represents a big increase over the current subsidy level to those states, which was about $4 million in 2018. The FCC’s Connect America Cost Model assumes average revenue per user of $75 monthly, which means that our most generous voucher—$45 monthly—reduces rural households’ expected out-of-pocket expense to $30 monthly.

Overall, the voucher system would be much more equitable among the states. A consumer voucher ensures that public funds are not spent disproportionately on a tiny percentage of rural households—a very real possibility as FCC subsidies under the current system increasingly go to the remaining households that are most expensive to serve.

Consider that currently under High-Cost Support, carriers in the most supported state (Alaska) receive over 2,500 times more in disbursements per rural household than carriers in the least supported state (Rhode Island). Under our proposed voucher system, Alaska’s rural households would receive only nine times more support than rural households in Rhode Island. Our proposed coupons compress those disparities (here, 9:1 rather than 2,500:1) in order to bring more equity between rural households and to give rural households incentives to use lower-cost, wireless broadband substitutes to fiber optics.

The voucher program treats all market participants equally. Subsidies are portable between competitors, and consumer preferences, not local or federal regulators, determine which providers succeed in the program. Another benefit of consumer vouchers is that they would make it administratively easier for consumers to add state USF vouchers or existing Lifeline vouchers, further reducing the out-of-pocket cost of service and increasing consumer demand for services. If participation is 50 percent or less for these vouchers, which is plausible given the participation rates in Lifeline, the program would have hundreds of millions of dollars unspent. Therefore, the FCC could either increase the subsidy to eligible households, refund subsidies to the states, or reduce USF fees. In time, the voucher program could simultaneously fulfill statutory goals and decrease public expense.


The gap in broadband access and quality between rural and urban households remains a pressing concern for state and federal policymakers, made even more urgent by the demands of social distancing during the COVID-19 pandemic. A lack of public investment or “political will” is not the reason that the digital divide persists. Up to $100 billion has been disbursed under various rural telecommunications programs over the past two decades, but a sizable share of those subsidies has gone to overhead or gold-plating costs that divert public funds from investment that benefits underserved and unserved rural households.

In its efforts to address the digital divide amid the COVID-19 pandemic, a future FCC ought to pursue an alternative plan that provides assistance to rural households that want a broadband connection, rather than disbursing public funds with the current complex programs. A monthly broadband voucher would align consumer preferences with subsidy disbursement and simplify the current administrative processes. Accordingly, we expect that this voucher program would see considerable uptake, and it would more efficiently and equitably disburse funds among the states and residents that fund these programs.