How not to waste $45 billion in broadband subsidies
In the middle of the pandemic, the FCC used a reverse auction, or competitive bidding, process to save taxpayers about $7 billion on projected expenses of $16 billion for broadband service to unserved areas — nearly a 50 percent savings! But because of concerns over some outcomes of the reverse auction, the current infrastructure bill may abandon competitive bidding and instead allocate the $45 billion in broadband subsidies in ways that virtually guarantee higher costs and fewer people connected.
Rather than ditching a tool proven to deliver cost-effective results, we should take a page from the FCC’s successful spectrum license auctions and continue to use this powerful tool while learning from the mistakes to ensure that new auctions do not suffer from the same deficiencies.
The FCC’s spectrum auctions today are the global gold standard for allocating spectrum licenses. But they were not born flawless. In the FCC’s second spectrum auction in 1994, more than 100 winning bidders defaulted on their bids. In a subsequent auction, the “PCS C-Band auction,” NextWave bid $4 billion for licenses and declared bankruptcy before paying. Those policy failures led to finger-pointing and recriminations, but nobody seriously considered abandoning spectrum auctions as a way of allocating licenses because auctions were faster, more efficient, and fairer than the alternatives.
Instead, the FCC and others around the world carefully examined what went wrong and fixed those flaws. Since then, the Commission has gone on to run nearly 100 spectrum auctions and raise $200 billion for the Treasury, and the Nobel Prize committee awarded the 2020 prize in economics to Paul Milgrom and Robert Wilson in part for their work in helping to design the auctions.
Concerns about the recent Rural Development Opportunities Fund (RDOF) reverse auction are legitimate. Concluding that such auctions are fatally flawed as a result is not — because the concerns are unrelated to using an auction. The mistakes that worry people include offering subsidies in areas that actually already had broadband (hat tip to FreePress.net for identifying many such areas) and poor bidder screening. Any subsidy program, not just reverse auctions, must solve those problems. Abandoning reverse auctions does nothing to solve them and introduces a host of other problems.
Most importantly, history demonstrates that the business-as-usual alternative of using a grant review process to distribute subsidies will lead to waste and leave a larger number of currently un- or under-served people without service than would be the case with reverse auctions.
In 2009, the American Recovery and Reinvestment Act allocated $7 billion for broadband. A group of 71 economists, including Nobel and future Nobel prize winners (and the two of us), suggested competitive bidding as the best tool to allocate subsidies, just like a homeowner might get competitive bids to repair a roof.
But the government did not choose that approach. Instead, it asked for grant submissions and hand-reviewed each submission. The result was an incoherent set of criteria applied inconsistently across proposals, with no rigorous way of comparing one proposal to another. One recent study found that the grant review method was barely better than simple random selection and that a reverse auction would have resulted in at least twice as many new connections for the same amount of money.
As we readily admit, auctions are not perfect. The details matter, and when the details are wrong, the overall auction can go wrong too.
Ensuring that the next reverse auctions do not suffer from RDOF’s problems requires two fixes.
First, we need comprehensive, complete, and accurate knowledge of which areas do not have service. Achieving this objective requires careful analysis and mapping, which hopefully the FCC process will provide soon, and the bill relies upon for its planned distribution of subsidy money.
Second, it is crucial to ensure that the bidders can provide the service they promise. One way to solve this problem, as we argued in comments to the FCC, is for the companies to bear the risk of their proposal since they have the most knowledge and control of performance. Auctions should not exclude new technologies, but taxpayers should not bear all the risk of uncertain outcomes. If, for example, Elon Musk and SpaceX promise to deliver a specific service in exchange for a payment, require them to not only provide the service before the government pays them, but also to face a penalty if they fail to deliver on their promises. Once again, this is a procurement issue, not an auction issue.
RDOF failures show us how to build more successful auctions. The 2009 Recovery Act experience should convince us to never use arbitrary grant selection mechanisms again.
The bottom line is not to throw the auction out with the bathwater.
Gregory L. Rosston is the Gordon Cain Senior Fellow at the Stanford Institute for Economic Policy Research and Director of the Public Policy program at Stanford University. He previously served as Deputy Chief Economist at the Federal Communications Commission. He provided advice to a bidder during the RDOF auction. Follow him on Twitter @grosston