M Mindy for Government
The Mindy Institute · Evidence library

Supplier Discovery: the case file

The marshaled, verified body of evidence behind the thesis — under-competed public solicitations cost governments money, firms mostly don’t bid because they never heard, and a large share is a reachable market, not a missing one. Every source links out to the original document.

Every citation is real and independently checked. New sources — additive to our white papers. · getmindy.ai
✓ Confirmed finding read directly off the source ▲ Lead real source; exact figure to confirm before printing
⚠ The one rule that governs this file — under-SERVED, not under-SUPPLIED

Our market is categories where a real competitive supplier base already exists but the solicitation isn’t reaching it — janitorial, landscaping, IT services, construction trades, professional services, commodity goods. Firms exist; they just never heard.

It is NOT sole-source markets — jet engines, satellites, missile systems — where only one vendor can do the work. The supplier-concentration sources below are used only as contrast; the size-of-prize model nets true sole-source out.

The on-thesis core

Competitive commercial categories — paving, janitorial, landscaping, roofing, staffing — draw too few bids despite ample capable suppliers. A reach problem, not a supply problem.

The direct answer to "isn’t this just sole-source?" — No. These are ordinary, low-barrier trades with many firms, still under-competed because the notice doesn’t reach them.

South Carolina Legislative Audit Council, "A Review of Competition for the Department of Transportation’s Road Paving Contracts" (Dec. 2001)
✓ Confirmedtext extracted from the primary PDF and confirmed firsthand

Of 496 paving projects, 49% drew two or fewer bidders and 12% drew only one. Single-bid projects came in 5% above the cost estimate; competed ones 4% below — a ~9-point swing.

The flagship on-thesis source. A routine, low-barrier trade with many capable contractors, yet half its solicitations were under-competed, and the uncompeted ones cost measurably more — a real market that isn’t being reached, not one that doesn’t exist.

Scope: One state, one trade (paving), 2001. A state audit of competition (not an AASHTO/FHWA product).

Urban Institute, "Do Minority-Owned Businesses Get a Fair Share of Government Contracts?" (1997)
✓ Confirmed

A meta-analysis of 95 disparity studies finds capable "ready, willing and able" firms are under-utilized across every ordinary category — construction, goods, professional and non-professional services.

Cross-jurisdiction proof that capable-firm availability outstrips utilization across ordinary categories, not one niche.

Scope: 1997; disparity framing; broad industry buckets, not fine trades or bid counts.

BBC Research & Consulting, "State of Oregon Disparity Study" (2023)
✓ Confirmed

Disparity indices (100 = parity): construction 41, professional services 31, non-professional services / goods 21 — with landscaping and janitorial named among the under-utilized commercial trades.

A recent statewide study putting disparity numbers on the exact reachable trades (landscaping, janitorial) that are our sweet spot.

Scope: Oregon statewide, 2023; disparity/availability framing, not raw bid counts. (Re-extract a specific per-trade index before printing it.)

Also in file as leads (confirm figures before printing): Vancouver WA 2024 (names janitorial/landscaping/roofing/staffing), Alaska DOT&PF 2021 (59.55% of capable firms "seldom or never solicited"), New Jersey 2024 (~240k contracts).

Claim A

More competition lowers what governments pay — and advertising a solicitation more widely causally increases the number of bidders.

Coviello & Mariniello, "Publicity requirements in public procurement: Evidence from a regression discontinuity design," Journal of Public Economics 109 (2014), 76–100
✓ Confirmed

A stricter publicity requirement causally induced more bidders, and more bidders is the channel through which it lowered the price the government paid.

The cleanest causal evidence that wider advertising = lower cost — the exact mechanism of Supplier Discovery.

Scope: Italian public-works auctions; regression-discontinuity design. One country / one contract type.

Hong & Shum, "Increasing Competition and the Winner’s Curse: Evidence from Procurement," Review of Economic Studies 69(4) (2002), 871–898
✓ Confirmed

In common-value settings, procurement costs can rise with more bidders (the winner’s curse). Florida highway construction data.

We cite this against ourselves — the honest boundary. Our real claim is moving a solicitation from 1–2 bidders to 4–5, not chasing an unlimited count.

Scope: Common-value environments; does not overturn the general result that more bidders lower price. Cite as nuance, not headline.

Also in file as leads (confirm figures first): Gupta 2002 (US, savings up to ~6–8 bidders), Iimi 2007 (~0.2%/bidder), Tas 2020 (cost-minimizing counts), Kenny & Crisman 2016 (publicity→competition).

Claim B

Firms don’t bid mainly because they never learned the opportunity existed. Reaching them is a distribution problem, not a posting-compliance one.

Blum, Datta, Fazekas, Samaddar & Siddique, "Introducing E-Procurement in Bangladesh: The Promise of Efficiency and Openness," World Bank Policy Research WP 10390 (2023)
✓ Confirmedabstract extracted from the primary PDF and confirmed firsthand

Making tenders findable online raised the number of bidders by 1.6–2.2, cut single-bidder tenders by 7.8–13.5 percentage points, and raised the discounts firms offered by 7.4–8.0 points.

The cornerstone number — findability delivers BOTH more bidders AND lower price, from the exact intervention we run.

Scope: Bangladesh, 2011–16, fixed-effects on matched contracts. Developing-country e-GP — directional for US local, not literal.

OECD, "SMEs in Public Procurement: Practices and Strategies for Shared Benefits," OECD Public Governance Reviews (2018)
✓ Confirmed

"Finding opportunities on tendering platforms online" and suppliers who "do not know whom to contact" rank among the top barriers to SME participation.

Authoritative, cross-country confirmation that finding the opportunity — not price or capacity — is a leading barrier.

Scope: 37 OECD/partner countries; a policy diagnosis, not a causal estimate.

U.S. GAO, "Government Contracting: Federal Efforts to Assist Small Minority Owned Businesses," GAO-12-873 (2012)
✓ Confirmed

GAO names "knowledge of the federal contracting process" and "difficulty gaining access to contracting officials" as significant barriers. (A distinct GAO report from any we already cite.)

Scope: US federal, minority-owned focus; supports the awareness/information-cost family.

Claim C

Fragmentation and paywalled access suppress competition. Open, findable procurement raises it.

Bandiera, Prat & Valletti, "Active and Passive Waste in Government Spending: Evidence from a Policy Experiment," American Economic Review 99(4) (2009), 1278–1308
✓ Confirmed

83% of measured government overspending is "passive waste" — failure to shop the market — not corruption.

The single most powerful frame in the file: governments overpay mostly because buyers don’t reach the better option — exactly the access problem we attack. Also disarms the "this is really about corruption" objection.

Scope: Standardized goods, Italian public bodies; top-journal quasi-experiment.

Lewis-Faupel, Neggers, Olken & Pande, "Can Electronic Procurement Improve Infrastructure Provision? Evidence from India and Indonesia," AEJ: Economic Policy 8(3) (2016), 258–283
✓ Confirmed

E-procurement brought in winners from outside the local region and higher-quality contractors (no measured price drop — an honest result).

Evidence that making opportunities findable pulls in bidders beyond the usual local pool — our "reach the firms two counties over" pitch.

Scope: Public works, two developing countries; benefit shows up as entry + quality, not price.

Electronic Frontier Foundation, "The Time Has Come to End the PACER Paywall" (2020)
✓ Confirmed

Charging $0.10/page to view legally public federal court records turns the system into a "gatekeeper to information."

A documented, on-record instance that paywalling legally-public government information is a recognized barrier — the direct analogy to bid-aggregator paywalls. Use as analogy (it’s advocacy, and about courts).

Scope: US federal court records; advocacy source.

Also in file: Duguay/Rauter/Samuels (JAR 2023, EU open data → more open competitive bidding — direction confirmed, magnitude to verify); OECD (2018) & World Bank (2020) leads (the effect is conditional on institutions — an honest nuance).

Claim D — contrast only

Governments already treat supplier concentration as a named cost-and-risk problem — even in the hard sole-source markets that are not ours.

SourceWhat it establishesStatus
GAO-22-104154 (2022)DOD names "reliance on foreign and single-source suppliers" as a risk to mitigate (200,000+ supplier base).✓ confirmed
GAO-24-106129 (2023)Consolidation "may reduce competition and increase the risk of higher costs"; ~400 defense M&A/yr vs ~40 reviewed.✓ confirmed
GAO-25-107283 (2025)Dependence treated as a national-security risk requiring supply-chain visibility.✓ confirmed
SBA (2024)Federal initiative to "reverse a decade-long decline in the contracting base."✓ confirmed
DoD "State of Competition" (2022)Of ~51 A&D prime contractors in the 1990s, ~5 remain; remedy = broaden the base + small-biz outreach.▲ lead
The size of the prize

If under-competed contestable solicitations across U.S. state & local government drew more qualified bidders, the savings run to billions a year.

The transparent model

annual savings ≈ [ addressable S&L procurement $ ] × [ share under-competed AND contestable ] × [ cost reduction from real competition ]
Conservative
~$5B/yr
$0.9T × 18% × ~3%
Midpoint
~$26B/yr
$1.2T × ~27% × ~8%
Optimistic
~$68B/yr
$1.5T × ~34% × ~13.5%

These are arithmetic on the sourced inputs — not measured outcomes. The "contestable" haircut (~⅔ of single-bid buys) removes genuine sole-source. We present the method and the range; we never quote one number. Even the conservative floor is billions.

InputValue / rangeConfidence
Total S&L expenditure$3.7T (FY2021; $1.8T state + $1.9T local) — Census / Urban Institutemeasured (hard)
Procurement share of that~20–33% → ~$0.9T–$1.5T addressableassumption band
Under-competed AND contestable~40% single-bid, LESS genuine sole-source (~⅔ contestable)assumption — nets out sole-source
Savings when de-monopolized~2–3% of spend / ~1% per bidder → ~13.5% on affected contracts → 8.3%+ optimisticrange, each end sourced

The data gap is part of the pitch. There is no clean, primary, published distribution of bid counts for U.S. municipal/local solicitations. The best proxies (US federal single-bid ~44%, EU single-bid >40%) both land near ~40% — stated as borrowed from federal + EU analogues, not measured on US local data. That missing dataset is exactly what a tool like the Mindy map begins to produce.

On this document

Every source above is real and independently checked. ✓ Confirmed = the finding was read directly off the source or extracted from the primary PDF; ▲ Lead = the source is real and its figure is to be confirmed before printing in a government-facing document. New sources only — additive to the Institute white papers. Mindy is a product of GovCon Giants AI, an independent platform, not affiliated with or endorsed by any government agency. Figures are cited as published; none are modeled except the explicitly-labeled size-of-prize range.