Today we have a guest post from Sam Marullo, a senior fellow at the Institute for Progress and former counselor to the Secretary of Commerce.
Going around lately is Ezra Klein’s appearance on Jon Stewart’s podcast explaining the gap between the Biden Administration’s infrastructure goals and the results to date.
The clip closest to my heart is when Ezra explains the 14-step process to access rural broadband funding.
The Biden Administration set a bold goal: connect 100% of Americans to reliable, high-speed internet. High-speed internet is essential to everyday life, the argument goes. Remote work, telehealth, and virtual school are leaving behind areas without high-speed internet. And so the administration went to Congress and passed a bipartisan infrastructure law investing roughly $50 billion in rural broadband.
This is a good goal! It’s also a program which will overwhelmingly benefit red areas of the country, but both Democrats and Republicans could support. This is how government is supposed to work.
But Congress’s overcomplicated design undermined the law’s support. In the clip, Ezra reads off each of the fourteen steps while Jon Stewart makes dramatic expressions of disbelief. Every state is on the hook for no fewer than seven distinct planning documents, each requiring federal approval. There are reasons a planning grant application is different from a five-year action plan and an initial proposal, but none of them are good. Maps are important for broadband—you need to know where to build—but incredibly, Congress set up fifty-seven different mapping processes (one nationally and one in each participating state and territory).
Watch the clip—it’s well done.
For my part, I think the broadband program will be a clear success by 2030—unless changes disrupt the good work already underway. And I’m not suggesting scrapping the planning in place. Starting over would only cause delays.
But these criticisms are important. Government must be able to deliver more quickly.
Why does government act this way?
I think I can add to this conversation why government acts this way.
Congress structured the rural broadband program as a classic waterfall project. It moves through distinct phases in strict sequence. And Congress was unusually prescriptive—specifying each step in detail. You don’t start step two until step one is finished. And once you’ve moved on, there’s no going back.
This makes sense on a flow chart, but it’s hard in the real world. What this means is the program builds no broadband until all the planning is done. Why was Congress obsessed with maps? Because when you run a program linearly, the maps have to be perfect. If your house was mistakenly omitted from maps during planning, you aren’t getting broadband. The maps are locked. And that means failure because success is connecting 100% of Americans.
Congress designs programs this way for a few reasons. One reason is to capture the perceived efficiencies of planning. There’s a balance here—building broadband before doing any planning would obviously be wasteful. Still, there’s a harmful tendency to delay decisions until every detail is known. If some planning is helpful, surely more planning would be more helpful. But delay is costly. Delay leaves people without services longer and means plans are stale by the time they’re executed.
Another reason is the general American distrust of government. The federal government is built around fair notice and the opportunity to comment. Notice and comment constrains the government, giving individuals a chance to make their case. Congress created fifty-seven mapping comment periods for two reasons. Yes, comments improve accuracy, flagging missed houses and misstated service. More importantly, it wouldn’t be fair not to take comments.
But public comment has a cost. It takes time, and it turns fluid, informal decision making into a structured, formal process. Strong comment processes almost necessitate a waterfall approach. If you start construction before finalizing plans, when do you take comments? Before the plans are done? After you start building?
A third reason is because government faces immense pressure to avoid waste. Congress tries to immunize programs against attack by imposing rigid processes to limit discretion. Once implementation passes to the executive branch, the same pressures continue. The broadband program answers to no fewer than six oversight bodies1, some of which have partisan incentives to make bad faith attacks. It’s a cheap but effective attack to criticize a program for inefficiency or duplicative spending when there’s no master plan to point back to. Delay occasionally leads to unfavorable oversight reports, but waste often will.
How to do better
A better approach is to be agile. Iterate quickly. Don’t plan 100% before building anything. Instead, start small, learn fast, and adjust in real time. You’ll have fresher information, and real-world feedback is invaluable.
A more agile broadband program might have begun by immediately addressing known coverage gaps. Maps could be refined as initial construction progressed. As better data revealed new gaps, more funding could be deployed. To be fair, several other rural broadband programs followed exactly this approach. But Congress withheld the largest pool of funds until plans were finalized.
An agile approach can fail too. Commit too many resources upfront and you won’t have enough to finish. Some design decisions really do need to be made upfront. But the approach offers some real advantages. For one, more rural broadband would be built today.
Waterfall and agile aren’t my terms. I’m stealing them from project management, and particularly software development. In software development, the agile approach has largely won out. Famously, the disastrous rollout of healthcare.gov was partially the result of a waterfall approach.
Outside of software, things are more complicated. Some sectors, like construction, must use a waterfall approach. You can’t build walls without a foundation. Later steps (literally) rest on earlier ones. But where that’s not the case, agile approaches are gaining ground and bringing benefits like quicker timelines, higher productivity, lower risk, and fewer resources devoted to planning.
Beyond broadband
I’ve focused so far on rural broadband, but you can see the waterfall approach throughout government once you start looking for it.
The National Environmental Policy Act (NEPA) requires assessment of environmental impacts before the federal government takes action. Although well intentioned, NEPA can delay projects. None of a project may begin until all of the impact is assessed. Sound familiar?
Another example comes from federal research grants. Researchers apply for a grant—and then they wait, sometimes more than a year. Work can’t begin until the government has decided to fund the entire project. Some programs now use more flexible approaches, like multi-stage competitions with pilot phases. The usual process, though, remains the old-fashioned one.
Not all programs are like this. The CHIPS Program, designed to restore U.S. semiconductor manufacturing, got special permission from Congress to negotiate directly with companies. This rapid, flexible approach resulted in larger investments and succeeded in attracting all five of the world’s leading semiconductor companies—something no other nation has achieved. IRS Direct File, which allows individuals to file their taxes without a commercial preparer, smartly launched with a limited scope—simple returns in a few states—and expanded from there.
Government is stuck with the waterfall to some extent. After all, the process of passing a law and executing it is a waterfall. The law must be finished before implementation can start.
But how laws are written matters enormously. Congress should avoid the rigidity it wrote into the rural broadband program and instead give agencies the flexibility to adapt to changing circumstances. Even better, lawmakers should encourage agile approaches by explicitly directing the use of pilots and experimentation within big programs. On the executive side, leaders should embrace agile management principles and ask their teams to do the same.
I believe in a government that can do big things. Government has the capacity to be a tremendous force for good, if it can deliver. That means good management, and agile management where possible. That’s how we ensure the next round of programs delivers.
The program is overseen by the Commerce Department Inspector General, the Government Accountability Office, the Senate Commerce Committee, the House Energy and Commerce Committee, and both the House and Senate Appropriations Subcommittees on Commerce, Justice, Science, and Related Agencies.
The program is unlikely to be a success in the future, either. It missed the boat.
I live in a rural exurb. Previously, people who needed broadband Internet, me included, bought property here near main roads, near schools, or near other government offices. There was a real need in the rest of the county. Note, was, not is.
At this point, people out here who need broadband buy wherever, and if it doesn’t have fiber, they get Starlink.
I understand the logic of not supporting what was then a monopoly, but the result of that is, literally no one was connected, a lot of money was wasted, and rural people now already have a solution without any help from the federal government.
Now Starlink has a competitor rolling out, and it won’t be a monopoly any more, either.
I feel like this article doesn’t address some basic questions. Why does this program make any sense? You can get high speed broadband in any rural area through Starlink.