Every few years, a politician stands at a podium and declares that government should justify every single dollar it spends—from scratch. No more rubber-stamping last year's budget with a small increase. Every agency, every program, every line item must prove its worth as if it were brand new. It sounds revolutionary. It sounds logical. It's called zero-based budgeting.
And it has failed virtually every time it's been tried. From Jimmy Carter's White House to state capitals and city halls across the country, the idea keeps getting adopted with great fanfare—and quietly abandoned a few years later. The question isn't whether zero-based budgeting is a good idea in theory. It's why a reform that makes so much sense on paper collapses so completely in practice.
The Beautiful Logic of Starting from Zero
Most government budgets work on what's called an incremental basis. Last year you spent $10 million? Great—let's talk about whether you get $10.2 million or $10.5 million this year. The existing $10 million is barely questioned. It's the baseline, and the real negotiation happens only at the margins. Zero-based budgeting rejects this entirely. It says every agency must build its budget from the ground up each year, justifying every expense as though the agency were being created for the first time.
The appeal is obvious. Incremental budgeting lets outdated programs coast on inertia. A workforce training program designed for the 1980s economy might still receive funding simply because it received funding last year. Zero-based budgeting promises to expose these zombies—programs that persist not because they work, but because nobody ever asks whether they should exist.
In theory, it forces hard choices. Agencies would rank their activities by priority, and lawmakers could draw a funding line wherever the budget allows. Programs below the line get cut. Programs above it survive on merit. It's rational, transparent, and fair. It's also, as generations of budget reformers have discovered, almost completely unworkable in the real world.
TakeawayThe most logical-sounding reform isn't always the most practical one. Simplicity in concept can mask enormous complexity in execution.
Drowning in Paper: The Information Problem
Here's a number to sit with: the U.S. federal government runs roughly 1,300 agencies and sub-agencies delivering thousands of individual programs. Now imagine every one of them writing a detailed justification for every dollar they spend—every staff position, every office lease, every computer purchase—from scratch. Every single year. The paperwork alone is staggering. When Jimmy Carter brought zero-based budgeting to the federal government in 1977, agencies produced mountains of "decision packages" that no one had the time or expertise to meaningfully review.
And that's the fatal flaw. Zero-based budgeting doesn't just require agencies to produce more information—it requires decision-makers to absorb it. A state legislator reviewing a transportation budget doesn't have the technical knowledge to evaluate from zero whether a particular bridge inspection program is necessary. So what happens? Staff summaries get written. Shortcuts get taken. And the process quietly reverts to something that looks a lot like incremental budgeting wearing a zero-based costume.
Budget analysts call this the information asymmetry problem. The people who know whether a program works are the people running it—the very people whose funding depends on a positive review. Legislators and budget offices simply can't independently verify thousands of justifications each cycle. The tool designed to increase accountability ends up increasing paperwork while accountability stays roughly where it was.
TakeawayA decision-making process is only as good as someone's ability to actually process the decisions. More information doesn't automatically produce better choices—sometimes it just produces more noise.
The Constituencies That Won't Let Go
Even if zero-based budgeting could solve the information problem, it would still crash into the hardest wall in government: politics. Every government program, no matter how small or outdated, has people who depend on it—employees who run it, citizens who use it, contractors who profit from it, and legislators whose districts benefit from it. These groups form what political scientists call constituencies, and they fight fiercely against cuts.
Consider a hypothetical rural hospital subsidy program that a zero-based review identifies as inefficient. The data might show that the money could save more lives if redirected to urban emergency services. But the senator from that rural state doesn't care about system-wide optimization—she cares about keeping that hospital open because her constituents need it and her re-election depends on it. She'll fight the cut in committee, trade votes with colleagues, and rally public support. Multiply this scenario by hundreds of programs and you see the problem.
This is why zero-based budgeting's promise of cutting programs "on merit" misunderstands how democratic budgeting actually works. Budgets aren't technical documents—they're political documents that reflect compromises between competing interests. A process that pretends otherwise doesn't reform politics. It just adds an expensive analytical step before the same political negotiations happen anyway.
TakeawayGovernment budgets are not spreadsheets waiting to be optimized—they're maps of political power. Any reform that ignores this reality is solving the wrong problem.
Zero-based budgeting fails not because the idea is wrong, but because it assumes government operates like a rational machine. In reality, budgets are shaped by limited human attention, asymmetric information, and deeply entrenched political interests. No accounting method changes those facts.
That doesn't mean we should accept wasteful spending. But meaningful budget reform probably looks less like a dramatic overhaul and more like targeted program reviews, sunset clauses, and better performance data. Sometimes the boring, incremental approach to fixing incremental budgeting is the one that actually works.