Restaurant Failure Rate Statistics 2026: The Real Numbers vs the 90% Myth
The Duckhub team builds AI-powered QR menu and online ordering software used by cafes, bars, and restaurants. We write practical guides based on what we see working across thousands of published menus.
Restaurant failure rate statistics are dominated by a claim that is simply false: “90% of restaurants fail in their first year.” The measured reality, from peer-reviewed research and government data, is 17–26% first-year failure — and restaurants survive at almost exactly the same rate as American businesses in general. This page compiles the verified numbers, traces where the myth came from, and covers what actually closes restaurants.
This is the myth-busting hub of our restaurant data series (anchored by the restaurant industry statistics umbrella), and the sourcing standard matters more here than anywhere: every figure below links to the government dataset or the actual study.
TL;DR: the failure-rate numbers that hold up
- First-year failure is ~17–26%, not 90%: 17% in Luo & Stark’s 81,000-restaurant BLS-microdata study; 26% in Parsa’s Cornell study.
- The 90% claim has no data behind it — it traces to a 2003 American Express ad and an NBC reality show, and was never a research finding.
- Five- and ten-year survival matches other industries: roughly half of establishments reach year five; 34.7% of all US establishments born in 2013 were still operating in 2023 (BLS).
- Restaurants are not uniquely risky: new full-service restaurants failed less often (17%) than other new service businesses (19%) in the best available study.
- Closure ≠ failure: many closures are voluntary — owners closing profitable restaurants for personal reasons.
- The real 2026 risk story is margins, not mass death: 42% of operators report being unprofitable, and independents shrank 2.3% in 2025 while chains grew.
Where the “90% fail” myth came from
The most famous restaurant statistic was invented for television. The 90% figure traces to the early 2000s — an American Express advertisement and NBC’s 2003 “Restaurant: A Reality Show” — and American Express later acknowledged it had no data supporting the claim. No academic study, government dataset, or industry census has ever produced a first-year failure rate near 90%.
The myth survives because it is useful: it flatters survivors, dramatizes TV, sells consulting, and justifies lenders’ caution. H.G. Parsa — the researcher who did more than anyone to test it — called it flatly a myth after his Cornell study measured the real number at roughly a quarter. When you see 90% (or its milder cousin, “60% fail in year one”) cited today, you are looking at a two-decade-old advertising claim, not data.
The real restaurant failure rate, measured
The best evidence puts first-year restaurant failure between 17% and 26%, with about half of restaurants surviving past five years. Two studies and one government dataset anchor the range:
| Statistic | What it measures | Source (date) | Quality |
|---|---|---|---|
| 26.16% | Independent restaurants that failed in year one (Columbus, OH cohort, 1996–99) | Parsa et al., Cornell Hotel & Restaurant Administration Quarterly (2005) | Independent (academic) |
| ~59–61% | Cumulative failure by end of year three in the same research program | Parsa et al. (2005, 2015) | Independent (academic) |
| 17% | First-year failure of full-service restaurant startups — 81,000 restaurants in 20 years of BLS microdata | Luo & Stark, “Only the Bad Die Young” (2014) | Independent (academic, BLS data) |
| ~80% | One-year survival of new food-service establishments in BLS cohort data | BLS Business Employment Dynamics, Table 7 | Government |
| ~50% / 34.7% | Five-year survival of establishments generally / ten-year survival of the 2013 cohort | BLS, The Economics Daily (2024) | Government |
Note what the multi-year numbers really say: cumulative three-year failure around 60% is where the “60%” folklore comes from — a three-year figure routinely misquoted as a first-year one. The denominators got lost; the scary number survived.
Are restaurants riskier than other businesses? No.
Restaurants survive at almost exactly the same rate as American businesses overall — and new full-service restaurants slightly outperform other new service businesses. This is the finding that should retire the “riskiest business in America” trope. In Luo & Stark’s BLS-microdata study, 17% of full-service restaurant startups failed in year one versus 19% for other service-sector startups. And BLS cohort curves for the broader food-service sector track the all-industry average within about a percentage point at one, five, and ten years.
| Cohort survival | Restaurants / food service | All private sector | Source |
|---|---|---|---|
| 1 year | ~80% | ~80% | BLS Business Employment Dynamics |
| 5 years | ~49–50% | ~50% | BLS Business Employment Dynamics |
| 10 years | ~34–35% | 34.7% (2013 cohort) | BLS, The Economics Daily (2024) |
The honest framing for anyone opening a restaurant: the odds are the ordinary odds of small business — demanding, but average. What is different about restaurants is the margin structure underneath the survival curve, which is why the same operator data shows 42% of operators unprofitable last year on 2.8–4.0% median margins. Restaurants don’t die more often; they live closer to the line.
Closure is not failure
A meaningful share of restaurant closures are voluntary exits, not insolvencies — and conflating the two inflates every scary number. Parsa’s interview research documented owners closing profitable restaurants for family, health, and lifestyle reasons; the demands of the trade (hours, stress) drive exits that no bankruptcy court ever sees. Brian Headd’s classic Small Business Economics analysis of Census data found a substantial share of “closed” small businesses were successful at closure by their owners’ own account.
This distinction matters practically. Raw closure counts — the numbers that circulate after every downturn — mix insolvency, retirement, relocation, concept changes, and lease-expiry walkaways. Actual financial failure (unpaid creditors, bankruptcy) is a fraction of gross closures. When a statistic says “X restaurants closed,” the honest follow-up question is always: how many failed?
Why restaurants actually close
The research-backed causes are unglamorous: money management, experience, and concept — not cuisine. Across Parsa’s qualitative work and small-business research generally, the recurring failure factors are:
- Undercapitalization and cash-flow management — the dominant, repeatedly confirmed cause; thin margins leave no buffer for the first slow quarter.
- Lack of business (not culinary) experience — failed operators skimped on budgeting, costing, and planning; survivors ran numbers.
- Fuzzy concept — “good food” without a defined market and operating model.
- Cost structure shocks — on 2.8–4.0% median margins, spikes in labor, food, rent, or fees flip the P&L; a 15–30% delivery commission can exceed the entire profit line.
- Owner burnout — the top complaint in exit interviews, and a reminder that many “failures” are resignations.
Notably absent from the evidence: any single silver bullet. Technology shows up as a survival tool — the efficiency-vs-profitability data suggests it buys operators room rather than guarantees — but no study yet ties any specific tool causally to survival.
Closures 2020–2026: spike, normalization, consolidation
The pandemic caused a real closure spike — roughly a doubling in 2020 — and closure rates then normalized, leaving consolidation as the current story. BLS data recorded food-service closures surging to roughly twice the pre-COVID annual pace in 2020, with the excess concentrated in Q2 2020; by 2021, quarterly closures were back near historical norms. The lasting damage was financial (debt loads) rather than a permanent elevation of the failure rate.
The 2025–2026 pattern is not mass death but sorting: independent locations declined 2.3% in 2025 (a net loss of 9,500+) while chains grew 1.4% (Technomic), and analysts flag about 9% of full-service restaurants as at risk of closure in 2026 amid soft traffic and high costs. That is a margin-pressure story — the same one told by the 42%-unprofitable statistic — not a return of the 2020 shock, and nothing in it resurrects the 90% myth.
Failure-rate claims you should not cite
This niche is the industry’s worst offender for zombie statistics. The claims to drop, with corrections:
| Popular claim | Why it fails | Use instead |
|---|---|---|
| “90% of restaurants fail in year one” | Advertising claim from ~2003; AmEx admitted no data; no study supports it | 17–26% first-year failure (Luo & Stark; Parsa) |
| “60% of restaurants fail in the first year” | A three-year cumulative figure (~60%) misquoted as first-year | “~60% cumulative by year three; ~20% in year one” |
| “Restaurants are the riskiest business” | BLS cohorts show survival within a point of the all-industry average | “Restaurant survival is about average; margins are below average” |
| “Every closure is a failure” | Research documents voluntary exits of profitable restaurants | Separate insolvency from owner-initiated closure |
| “Half of restaurants fail in the first year post-COVID” | Closure rates normalized by 2021 in BLS data | The 2020 spike + normalization, with dates |
One meta-lesson worth citing in its own right: the 90% myth persisted for twenty years after being debunked in a peer-reviewed journal, because a dramatic wrong number beats a boring right one. Statistics pages that link primary sources — the reason this series exists — are the antidote.
The honest risk picture for 2026
Opening a restaurant in 2026 carries ordinary small-business odds and extraordinary margin discipline requirements. About four in five new restaurants survive year one; half reach year five; a third reach ten — the same curve as the rest of the economy. What kills the ones that fail is rarely the food and usually the finance: too little capital, too loose a grip on prime cost, and revenue channels that hand 15–30% to intermediaries. The controllable variables are the boring ones — costing, scheduling, direct channels, and tools that stretch a small team. For those levers, see the profit margin benchmarks and the technology data pillar; for the structural backdrop, the industry statistics hub.
Survival is a margin game, and margins are won on the boring lines. Duckhub helps independents hold them: direct online ordering at 0% commission, a QR menu that updates instantly instead of reprinting, and AI handling descriptions, photos, and translations. The free Egg plan includes 70 products and 30 QR table codes; paid plans start at $39/month. More data and guides on the Duckhub blog.