90,887 Front Doors

The short version

A business that wants to sell to the federal government goes to one website. Registration is in one system. Opportunities are posted in one place. There is a single front door, and every federal buyer is behind it.

A business that wants to sell to state and local government faces 90,887 separate purchasing entities — and no equivalent front door at all.

That number is not an estimate. The U.S. Census Bureau counts governments every five years, and the 2022 Census of Governments found:

More than half of those — 52,101 — are special-purpose governments: school districts, water authorities, transit agencies, fire districts, utility districts, port authorities. These are real buyers with real budgets, and a vendor would never think to check most of them. The Census also excludes 1,313 dependent school systems from this count, so the number of actual purchasing bodies is arguably higher.

This paper is about what that structure costs — and it is not primarily a cost to vendors. It is a cost to the governments doing the buying.

The comparison worth drawing

In fiscal year 2024, federal contract obligations totaled approximately $755 billion, according to the Government Accountability Office.

State and local government purchasing is substantially larger. We are deliberately not putting a single dollar figure next to it, and the reason is instructive: no clean number exists. The commonly cited figures either measure total government expenditure — which includes salaries, Medicaid transfers and debt service, none of which are procurement — or come from vendor marketing materials without published methodology. We looked for an authoritative procurement-specific figure and could not find one.

That absence is itself part of the story. We can say precisely what the federal government buys because it flows through one system that produces one dataset. Nobody can say what American cities, counties and districts buy in aggregate, because there is no system that would know.

So the honest comparison is not dollar-to-dollar. It is architectural:

$755 billion flows through one front door. A larger sum flows through 90,887 of them, and no one is counting.

What this costs the buyer

The intuitive reading is that fragmentation is a vendor problem — an inconvenience for businesses that have to check many places. That reading is incomplete.

Fragmentation raises what governments pay, through a direct mechanism: the harder your solicitation is to find, the fewer bids you receive, and thin bidder pools cost money.

Research at Yale, Columbia and UC Berkeley, funded by the U.S. Department of Transportation and the National Bureau of Economic Research, found that deliberate outreach to widen the bidder pool is associated with 17.6% lower project costs — and that 70% of states rarely do it. Public procurement typically draws two to three bids per solicitation.

When those researchers asked officials to name what drives costs up, thin competition came back repeatedly — and unprompted.

Every additional front door is a place a qualified firm might not know to knock. Multiply that across 90,887 entities and the aggregate effect is not inconvenience. It is a structurally thin market, everywhere, all the time.

(That study covers state highway resurfacing, so the dollar magnitudes should not be transplanted onto a municipal budget. The mechanism does not depend on asphalt.)

Why no front door emerged

Federal procurement has a single system because federal law created one. State and local procurement has no equivalent because no law requires it, and the obligation that does exist points somewhere else entirely.

Publication requirements are set at the state level, and the typical standard is publication in a "newspaper of general circulation." These provisions are genuinely old. Florida's traces to 1877. Texas's current version dates to 1987 and has not been substantively amended since 1993. They were written when a newspaper was how information reached a community, and most have never been revisited.

Some states have modernized. Virginia (2013) makes website posting mandatory and newspaper publication optional. Florida (2021–22) made website publication co-equal and now requires that a bid advertisement on a government website "include a method to accept electronic bids." Most states have not followed.

No state requires its cities to post to a statewide portal. Several states run one for their own agencies — Texas operates a searchable electronic bid board for state purchases while Texas cities remain on a 1987 newspaper standard. The gap is not oversight. It is simply that nobody legislated across it.

What filled the vacuum

Where no public front door exists, a private one gets built — and it charges admission.

Commercial aggregators collect notices from thousands of separate sources and sell access to the compiled result. Published rates from one major service: a free tier covering a single agency, $5 per document downloaded outside a subscription area, up to $2,699 per year for national coverage.

Three features of this layer follow directly from the fragmentation:

Aggregation is valuable precisely because it is hard. Ninety thousand sources is a genuine engineering problem, and solving it creates something worth money. That is a legitimate business.

But the value depends on the fragmentation persisting. An aggregator's product is the compilation. Terms of use in this sector routinely prohibit "data scraping" and use "for purposes of commercial data aggregation." One network markets exclusivity directly to vendors: solicitations that go "exclusively on BidNet Direct — giving suppliers first access to bids they won't find anywhere else."

And nothing is portable. Of nine widely used commercial eProcurement and bid-distribution platforms, not one publishes an open API or feed of open solicitations. A city that wants its notices to travel further has no supported technical path to send them anywhere.

The result is a market where the intermediaries have a commercial interest in the front door never being built.

The one city that tried the alternative

The Open Contracting Data Standard is an international format for publishing procurement information in a structured, machine-readable, freely reusable way. Adoption in the United States is close to nonexistent.

Portland, Oregon is the first and only U.S. city publishing to the standard. Of 134 datasets in the international registry, exactly one is American — roughly 0.75% of global publishers, one city out of approximately 90,000 U.S. local governments.

Portland's experience explains the adoption rate. Producing the feed required assembling data from four or more separate internal systems. That is the work in front of any city that wants its procurement data to be openly usable — and most cities do not have the staff to do it once, let alone maintain it.

The barrier is not willingness. It is that every city is being asked to solve the same problem independently, with no shared infrastructure and no requirement that would justify the budget.

What follows

Fragmentation is not going to be legislated away. Ninety thousand governments will not adopt a common system, and no authority exists to make them.

But the practical problem is narrower than the structural one. A city does not need every other city to change anything. It needs its own notices to be findable by firms that do not already know it exists — which is a distribution problem, not a governance problem, and it can be solved one city at a time.

Three things make that tractable:

Additional publication is expressly authorized. California's code permits an agency to "give such other notice as it deems proper." Ohio's permits notice "distributed by electronic means." Virginia's contemplates "other appropriate websites." No authority we found suggests a city takes on risk by publishing more widely — and a city that posts only to a third party has a defective solicitation.

The marginal cost is near zero. Publishing a notice a second place is not a procurement reform, a system migration or a budget line. It is a copy of information the city has already produced and is already required to release.

The benefit is measurable. More firms aware of a solicitation is more bidders, and the research on what that does to price is consistent in direction even where it varies in magnitude.

Where we fit

Mindy publishes municipal, county and district solicitations on a free public map. Cities post at no cost. Businesses search at no cost. No subscription, no per-document charge, no login between a public notice and the public that paid for it.

We are building the front door that the structure never produced — one city at a time, and publishing what we learn as we go.

Sources

Entity counts U.S. Census Bureau, 2022 Census of Governments — Organization Component, released August 24, 2023. 90,887 total; 3,031 counties; 19,491 municipalities; 16,214 towns/townships; 12,546 independent school districts; 39,555 special districts. Excludes 1,313 dependent school systems.

Federal contract spending U.S. Government Accountability Office, A Snapshot of Government-Wide Contracting for FY 2024 (June 24, 2025). Approximately $755.1 billion.

On the absence of a state/local procurement figure. Commonly cited totals measure government expenditure rather than procurement, or originate in vendor marketing without published methodology. We could not locate an authoritative procurement-specific aggregate and have not published one.

Procurement cost and competition Liscow, Z., Nober, W., & Slattery, C., Procurement and Infrastructure Costs (July 2024), funded by U.S. DOT and NBER. Outreach finding p. 25, Appendix Table E.1 (Romano-Wolf adjusted p = 0.01); 70% figure p. 17, Appendix Figure C.2; official quotation p. 17. Scope: state highway resurfacing, projects begun 2018–2019. Open Contracting Partnership (April 2025) — two-to-three bids observation; cross-jurisdictional, not a U.S. municipal statistic.

Publication requirements and authority Fla. Stat. § 50.011 and § 50.0311(9); Tex. Loc. Gov't Code § 252.041; Va. Code § 2.2-4302.1(2); Cal. Pub. Contract Code § 22037; Ohio Rev. Code § 731.14.

Open contracting adoption Open Contracting Partnership Data Registry; City of Portland OCDS publication policy, covering April 2015–March 2026.

Access layer DemandStar / Euna OpenBids published supplier pricing and terms of use (accessed 2026); SOVRA / BidNet Direct supplier marketing (2026); API and feed availability assessed across nine commercial platforms.

GovCon Giants · Mindy — getmindy.ai