Americans interact with their government through interfaces that would embarrass a 2008 startup. The gap isn't technical. It's political.
What the Gap Actually Looks Like
The private sector has spent twenty years making complex transactions frictionless. You can deposit a check by taking a photo of it. You can dispute a credit card charge in thirty seconds. You can track a package across three continents in real time.
Then you try to access a government service.
The IRS's primary taxpayer-facing system, IRS.gov, processes over 150 million individual returns annually. For most of its existence it ran on COBOL — a programming language developed in 1959. The agency spent years trying to modernize and consistently failed, not because the engineers couldn't build better systems, but because Congress repeatedly underfunded the effort, and the political constituencies that benefit from tax complexity actively lobbied against simplification.
The Veterans Affairs digital infrastructure serves 9 million enrolled veterans. A 2021 audit found that VA systems included software running on operating systems no longer supported by their manufacturers — meaning systems that couldn't receive security patches, serving people who had risked their lives for the country.
Healthcare.gov launched in 2013 and crashed immediately, not because the government hired incompetent people, but because the procurement process that governs how government agencies hire technology vendors is so byzantine that the best engineering firms in the country routinely decline to bid. The process was designed to ensure accountability. It produces paralysis.
These are not isolated failures. They are the output of a system that has not been structurally updated to account for the fact that software is now infrastructure.
The Procurement Problem
The root cause of the civic technology gap is not talent, funding, or political will in the abstract. It is procurement.
The Federal Acquisition Regulation — the rulebook governing how the government buys things, including technology — runs to over 2,000 pages. It was designed primarily to prevent corruption and ensure competitive bidding for physical goods: tanks, office furniture, construction contracts. Applied to software development, it produces predictable dysfunction.
A government agency that wants to build a new digital service must write a requirements document describing, in detail, what the software will do — before any software has been built. It must then issue a competitive bid, wait months for responses, evaluate proposals, award a contract, and begin work. By the time development starts, the requirements are often already outdated.
Private sector technology development works in the opposite direction. You build something small, ship it to users, watch what happens, and iterate. The entire discipline of modern software engineering is organized around the impossibility of knowing in advance what the right thing to build is. Government procurement is organized around the assumption that you can know — and penalizes you when reality disagrees.
The result is a graveyard of expensive, late, and broken government technology projects. The Congressional Budget Office estimates that federal IT spending runs between $80 and $90 billion annually. A substantial fraction of that is spent maintaining legacy systems that should have been replaced years ago, because replacing them requires navigating a procurement process that makes replacement harder than maintenance.
The Political Economy of Bad Interfaces
The civic technology gap persists not because no one has noticed it but because the incentives that would drive improvement don't exist in government the way they do in the private sector.
A private company with a bad checkout process loses customers to competitors. The DMV does not lose customers to competitors. A bank that takes three weeks to process a loan application loses to a fintech that does it in three minutes. The Social Security Administration does not lose to a competitor that processes disability claims faster.
Without competitive pressure, the incentive to invest in user experience is almost entirely absent. What replaces it? Political will — which is episodic, inconsistent, and easily redirected. The Obama administration created the U.S. Digital Service in 2014 and 18F in the same year, both explicitly modeled on private sector technology culture and tasked with fixing government software. Both produced genuinely good work. Both have operated at the margins of a system that consistently resists what they're trying to do.
There is also a subtler political economy at work. Bad civic technology is not politically neutral. It falls hardest on the people who interact with government most: benefits recipients, small business owners navigating regulatory compliance, immigrants processing visa applications, veterans filing disability claims. These are not, in general, the people with the most political power. The people with the most political power interact with government primarily through lawyers and lobbyists — and those interfaces work fine.
"Bad civic technology is not politically neutral. It falls hardest on the people who interact with government most."
What Good Civic Technology Looks Like
It exists. It's just unevenly distributed.
The IRS Free File program — when functioning — allows millions of Americans to file taxes at no cost through software that actually works. The VA's My HealtheVet platform, despite its legacy infrastructure problems, has steadily improved and now allows veterans to access records, schedule appointments, and communicate with providers online. USAJOBS, the federal jobs portal, was rebuilt in the early 2010s into something that functions like a modern job board.
At the state and local level, the variance is enormous. Some states have built genuinely excellent digital infrastructure for motor vehicle services, business registration, and benefits access. Others are still faxing documents between agencies. The difference correlates strongly with state IT investment levels, the presence of a dedicated digital services team, and — crucially — political leadership that treats digital infrastructure as infrastructure rather than overhead.
The model that works is consistent across every successful implementation: small, cross-functional teams with real authority, working in short cycles, shipping to real users, iterating on feedback. It is, in other words, how good software gets built everywhere. The question is whether government can consistently organize itself to do it.
The $11 Billion Irony
Here is the number that should land hardest in this context: in 2024, Americans spent $11.1 billion on political advertising. That figure — the one this site tracks — is the cost of persuading voters during a single election cycle.
The entire annual budget of the U.S. Digital Service is approximately $200 million. The federal government spends less on the teams trying to fix its technology than the 2024 election spent on political ads in Pennsylvania alone.
"The federal government spends less on fixing its technology than the 2024 election spent on political ads in Pennsylvania alone."
The gap between what we spend to win power and what we spend to make power functional is not a technical problem. It is a political one. And it will not be solved by better engineering until it is first solved by better priorities.
The code that would fix the VA's claims processing system exists. The engineers who could write it are available. The money — relative to what we spend on other things — is not the obstacle. What's missing is the political will to treat the experience of interacting with government as something that matters.
Until that changes, Americans will continue filing their taxes on software that crashes in February, renewing their licenses through portals designed during the Bush administration, and trying to find their representative's actual policy positions on a page last updated in 2019.
The gap isn't technical. It never was.
SOURCES: Congressional Budget Office (federal IT spending), Government Accountability Office (VA systems audit), U.S. Digital Service annual reports, Office of Management and Budget IT dashboard, Federal Acquisition Regulation (48 CFR).
This is the fourth installment of "The Data," a series examining the numbers behind American political spending and governance. Previously: "Why the Next American Economic Boom Will Be Built on Energy."