US Small Business Failure Rate Statistics 2026
By Axis Intelligence Research
Co-author: Mia Scarlett, Business Analyst | Last updated: August 3, 2026 | License: CC BY 4.0
Of the 988,310 private-sector establishments that opened in the United States in the year ended March 2024, 769,449 were still operating twelve months later — a first-year survival rate of 77.9%, meaning 22.1% closed. Five years out the attrition roughly doubles: the cohort that opened in the year ended March 2020 finished March 2025 with 51.4% still trading, according to Bureau of Labor Statistics survival data.
Quick Answer: What Is the Small Business Failure Rate in 2026?
About 22% of new US businesses fail in their first year, roughly 49% by year five, and roughly 65% by year ten, based on the most recent Business Employment Dynamics establishment survival tables from the U.S. Bureau of Labor Statistics, covering cohorts observed through March 2025. Axis Intelligence Research estimates that only about 29 of every 1,000 business applications filed with the IRS become an employer business that is still operating five years later — a figure we call the Application-to-Survivor Yield. First-year closures alone outnumber all US business bankruptcy filings by about 8.4 to 1, which is why bankruptcy counts badly understate how often businesses actually stop.
Key Findings
- Axis Intelligence Research finds that 22.1% of US private-sector establishments that opened in the year ended March 2024 had closed by March 2025, based on BLS Business Employment Dynamics cohort data — not the 90% first-year failure figure repeated across business media.
- Axis Intelligence Research finds that the establishment cohort opened in the year ended March 2020 recorded 51.4% five-year survival through March 2025, meaning 48.6% failed within five years.
- Axis Intelligence Research finds that accommodation and food services establishments opened in the year ended March 2024 survived their first year at 85.3%, 7.4 percentage points above the all-industry average — a direct contradiction of the widely repeated claim that restaurants fail faster than other businesses.
- Axis Intelligence Research estimates the 2026 baseline Application-to-Survivor Yield at 2.9%: of 1,000 IRS business applications, roughly 56 become employer businesses within four quarters and roughly 29 of those are still trading five years later.
- Axis Intelligence Research finds that the 218,861 establishments that closed within twelve months of opening in the year ended March 2025 outnumbered all 25,960 US business bankruptcy filings recorded for the twelve months ended March 31, 2026 by a factor of 8.4.
What Percentage of Small Businesses Fail Each Year?
The answer depends entirely on which cohort you follow and for how long. The Bureau of Labor Statistics tracks every private-sector establishment born in a given year and reports how many are still operating at each anniversary. That structure — a birth cohort followed forward — is the only defensible way to state a failure rate, and it is the structure used throughout this report.
The table below is the Axis Cohort Attrition Curve: a dated snapshot of the most recent complete observation for each survival horizon, as of the March 2025 data point.
| Years since opening | Cohort (year ended) | Establishments at birth | Still operating March 2025 | Survival rate | Failure rate | Source |
|---|---|---|---|---|---|---|
| 1 year | March 2024 | 988,310 | 769,449 | 77.9% | 22.1% | BLS BED Table 7 |
| 2 years | March 2023 | 1,049,106 | 691,774 | 65.9% | 34.1% | BLS BED Table 7 |
| 3 years | March 2022 | 1,065,228 | 599,752 | 56.3% | 43.7% | BLS BED Table 7 |
| 4 years | March 2021 | 839,297 | 439,753 | 52.4% | 47.6% | BLS BED Table 7 |
| 5 years | March 2020 | 767,573 | 394,689 | 51.4% | 48.6% | BLS BED Table 7 |
| 10 years | March 2015 | 677,876 | 235,071 | 34.7% | 65.3% | BLS BED Table 7 |
| 31 years | March 1994 | 569,387 | 71,595 | 12.6% | 87.4% | BLS BED Table 7 |
Two things fall out of the curve that most coverage misses.
The first is that year one is not a cliff. It is the single worst year — 22.1% of the March 2024 cohort was gone within twelve months — but the decline between years four and five is under one percentage point. Businesses that clear roughly four years enter a much flatter part of the curve.
The second is that the widely repeated “90% of startups fail” claim has no counterpart in federal data at any horizon short of three decades. It takes 31 years of observation for the private-sector cohort to fall to 12.6% survival. A statistic about venture-backed technology startups has been laundered into a statistic about the corner bakery, and the two populations have almost nothing in common.
Mia Scarlett, Business Analyst: The interesting number in this table is the one nobody quotes. Between the fourth and fifth anniversary the March 2020 cohort lost a single percentage point of its base — 439,753 down to 394,689. That is not a business population in distress; that is a population that has finished sorting itself. If you are underwriting a loan, valuing an acquisition, or modeling franchise churn, the cohort’s age is a better predictor than its sector, its size, or the macro headlines in the year you happen to be looking.
Small Business Failure Rate by Industry
Industry is the variable most people assume dominates survival. It matters — but the direction of the effect is close to the opposite of the folklore.
Axis Intelligence Research pulled the BLS survival series for three sectors carrying the strongest public assumptions — accommodation and food services, construction, and health care and social assistance — and compared each against the all-industry private-sector baseline for the same cohorts.
Five-year survival by sector (cohort opened year ended March 2020, observed March 2025)
| Sector | Establishments at birth | Surviving March 2025 | 5-year survival | Axis Sector Survival Gap | Source |
|---|---|---|---|---|---|
| Accommodation and food services | 50,259 | 29,826 | 59.3% | +7.9 pp | BLS BED Table 7, NAICS 72 |
| Construction | 57,753 | 32,608 | 56.5% | +5.1 pp | BLS BED Table 7, NAICS 23 |
| Health care and social assistance | 157,898 | 83,101 | 52.6% | +1.2 pp | BLS BED Table 7, NAICS 62 |
| All private sector | 767,573 | 394,689 | 51.4% | baseline | BLS BED Table 7 |
First-year survival by sector (cohort opened year ended March 2024, observed March 2025)
| Sector | Establishments at birth | Surviving March 2025 | 1-year survival | Axis Sector Survival Gap | Source |
|---|---|---|---|---|---|
| Accommodation and food services | 57,581 | 49,090 | 85.3% | +7.4 pp | BLS BED Table 7, NAICS 72 |
| Health care and social assistance | 219,773 | 180,420 | 82.1% | +4.2 pp | BLS BED Table 7, NAICS 62 |
| Construction | 64,892 | 51,689 | 79.7% | +1.8 pp | BLS BED Table 7, NAICS 23 |
| All private sector | 988,310 | 769,449 | 77.9% | baseline | BLS BED Table 7 |
The Axis Sector Survival Gap is a simple construction — sector survival rate minus all-industry survival rate for the identical cohort and identical observation date, expressed in percentage points — and it exists because raw sector rates get compared across mismatched cohorts constantly, which is how bad industry rankings get built.
Why the restaurant failure myth survives
Accommodation and food services is the sector most often described as a death trap, and in this data it is the sturdiest of the three examined, at both horizons. That is not a rounding artifact: the gap is 7.4 percentage points at one year and 7.9 at five.
Part of the explanation is definitional. BLS counts establishments — physical locations with payroll — not brands, concepts, or menus. A restaurant that changes owners, name, and cuisine while keeping its payroll account continuous is a survivor in this dataset even though the neighborhood experienced it as a closure. Part of it is composition: the sector’s births skew toward chain and franchise units with capital behind them, which survive at higher rates than independents.
The honest reading is that the sector’s establishment survival is above average and its concept survival is probably not — and that no federal series measures the second thing.
Construction has quietly stopped being the worst
Construction has ranked among the lowest-survival sectors for most of the series’ history. Its 2008 cohort recorded 64.5% first-year survival, and its 2007 cohort finished five years at 33.7%. The cohort opened in the year ended March 2020 finished five years at 56.5% — above the all-industry baseline.
Mia Scarlett: Every listicle that ranks construction as the riskiest sector is quoting a cohort that entered the market before the housing collapse. That was true then. It stopped being true somewhere around the 2013 cohort, and the 2020 cohort now sits five points above the national line. When a sector’s survival profile inverts and the published rankings do not move for a decade, that is a data-currency failure, not a construction story.
How Many Businesses Actually Reach the Starting Line?
Failure rate coverage almost always begins at the moment a business opens, which skips the largest filter in the funnel: most business applications never become an employer business at all.
The Census Bureau’s Business Formation Statistics track applications for an Employer Identification Number and project how many convert to businesses with payroll. In June 2026, seasonally adjusted business applications totaled 531,423, of which 149,714 carried high-propensity markers, and projected business formations within four quarters totaled 29,741, according to the Census Bureau’s June 2026 release.
That is a conversion rate of 5.6%.
The Axis Application-to-Survivor Yield (AASY), v1.0 — baseline reading 2.9%
Axis Intelligence Research combines the Census conversion rate with the BLS five-year survival rate to estimate how many business applications ultimately produce a five-year-old employer business.
Formula: AASY = (PBF4Q ÷ BA) × S₅
Inputs:
- PBF4Q = 29,741 (projected business formations within 4 quarters, June 2026, seasonally adjusted, Census BFS)
- BA = 531,423 (business applications, June 2026, seasonally adjusted, Census BFS)
- S₅ = 51.4% (five-year survival, private-sector establishments opened year ended March 2020, BLS BED Table 7)
Calculation: (29,741 ÷ 531,423) = 5.596% → 5.596% × 0.514 = 2.877%
Baseline reading, as of August 3, 2026: 2.9%, or approximately 29 survivors per 1,000 business applications.
This is an estimate, not an observation, and it is the first reading of the metric — a baseline with no history behind it. Its limitations are stated in full in the Methodology section below, and they are material: the conversion rate and the survival rate come from different cohorts and different statistical units.
What the number is useful for is scale. The public conversation treats “starting a business” and “surviving five years” as one continuous event with a roughly even chance attached. Sequenced properly, the odds compound, and they compound against the applicant.
Business Closures vs. Business Bankruptcies: The 8-to-1 Gap
Bankruptcy filings are the most frequently cited proxy for business failure because they are timely, national, and easy to count. They are also a poor proxy, and the gap is quantifiable.
US business bankruptcy filings totaled 25,960 in the twelve months ended March 31, 2026, up 11.4% from 23,309 a year earlier, according to the Administrative Office of the U.S. Courts. Business filings have nearly doubled since the twelve months ended March 2022, when they totaled 13,160.
Set that against establishment closures. Of the 988,310 establishments that opened in the year ended March 2024, 218,861 had closed by March 2025.
Axis Closure-to-Bankruptcy Ratio = 218,861 ÷ 25,960 = 8.4
First-year closures from a single birth cohort outnumber every business bankruptcy filed in the United States — across all cohorts, all ages, all sectors — by more than eight to one. The periods are adjacent rather than identical (closures observed through March 2025, filings through March 2026), which is disclosed in the fact table and does not move the order of magnitude.
| Measure | Value | Period | Source |
|---|---|---|---|
| Business bankruptcy filings | 25,960 | 12 months ended March 31, 2026 | U.S. Courts |
| Business bankruptcy filings | 13,160 | 12 months ended March 31, 2022 | U.S. Courts |
| Chapter 11 filings, all debtors | 9,941 | 12 months ended March 31, 2026 | U.S. Courts |
| First-year establishment closures | 218,861 | Cohort opened yr ended March 2024 | BLS BED Table 7 (Axis calculation) |
Mia Scarlett: Bankruptcy is a legal event with filing fees, counsel, and a schedule of creditors. Most small business failure is an operator quietly ending payroll, returning keys, and dissolving an LLC that owed money to nobody but its founder. Anyone modeling small business risk off bankruptcy counts is measuring the visible eighth of the distribution and calling it the whole.
Are Business Failure Rates Getting Worse in 2026?
Not by the survival measures. The first-year rate has been notably stable across recent cohorts: 79.4% for the March 2018 cohort, 79.2% for March 2019, 80.9% for March 2020, 79.1% for March 2021, 76.3% for March 2022, 78.2% for March 2023, and 77.9% for March 2024.
Two things about that series are worth stating plainly. The March 2020 cohort — establishments that opened into the pandemic’s first weeks — recorded the highest first-year survival of the group at 80.9%. And the March 2022 cohort, born into the largest wave of business formation on record, recorded the lowest at 76.3%. Formation booms dilute cohort quality; crises select for it.
The two indicators moving are formation volume and bankruptcy volume, and they are moving in opposite directions from survival. That combination — record entry, rising formal insolvency, flat survival — is what a churning business population looks like, not a deteriorating one.
Methodology
Data collection. Every figure in this report was fetched from its primary source during production on August 3, 2026, and logged with its URL and retrieval date in the accompanying CSV. Three source families are used:
- U.S. Bureau of Labor Statistics, Business Employment Dynamics, Establishment Age and Survival, Table 7 — total private (NAICS 00), construction (23), health care and social assistance (62), accommodation and food services (72). Cohorts are defined by year ended March; the most recent observation available is March 2025.
- U.S. Census Bureau, Business Formation Statistics, June 2026 release (CB26-115), published July 9, 2026 — business applications and projected business formations, seasonally adjusted.
- Administrative Office of the U.S. Courts, bankruptcy filing statistics, twelve-month periods ended March 31.
Survival and failure rates are computed as survivors ÷ births for the named cohort, and failure rate = 100 − survival rate. Failure in this dataset means the establishment ceased to report employment; it does not distinguish voluntary closure, sale, relocation with a new payroll account, or insolvency.
Formulas for Axis-calculated figures.
- Application-to-Survivor Yield (AASY) v1.0 = (PBF4Q ÷ BA) × S₅. Inputs and arithmetic disclosed inline above.
- Sector Survival Gap = sector survival rate − all-industry survival rate, same cohort, same observation date, in percentage points.
- Closure-to-Bankruptcy Ratio = first-year establishment closures (births − first-anniversary survivors) ÷ business bankruptcy filings.
All arithmetic was independently recomputed from the raw source values before publication.
Limitations.
- BLS BED tracks establishments, not firms. A multi-location company closing one site registers as a closure; a firm relocating with a continuous payroll account may not.
- The dataset covers employer establishments only. The large majority of US small businesses have no employees and are invisible to this series, so these rates describe the payroll-employing segment, not sole proprietorships.
- The AASY combines a forward-looking Census projection with a historical BLS survival rate drawn from a different cohort. It is an estimate of order of magnitude, not a precise conversion rate, and it should not be read as a prediction for applications filed in 2026.
- Census business applications and BLS establishment births are different statistical units measured by different agencies; the two series are sequenced in the AASY, never summed or averaged.
- Sector analysis covers four series, not all nineteen NAICS sectors. Claims about sector performance are stated relative to the all-industry baseline, not as rankings against unexamined sectors.
- The Closure-to-Bankruptcy Ratio compares adjacent, non-identical periods (closures through March 2025; filings through March 2026).
About This Dataset
Contents: 73 rows covering US private-sector establishment survival by cohort (1994–2024 birth years, observed through March 2025), sector-level survival for NAICS 23, 62, and 72, Census business application and formation counts for June 2026, US business bankruptcy filings for twelve-month periods ended March 2022–2026, and four Axis-calculated figures with method notes.
Temporal coverage: 1994-03-31 to 2026-06-30
Spatial coverage: United States
Creator and publisher: Axis Intelligence Research
License: CC BY 4.0 — free to use, republish, and build on with attribution.
Cite This Report
APA: Axis Intelligence Research. (2026). Small business failure rate statistics 2026: Survival by year, industry, and cohort. https://axis-intelligence.com/small-business-failure-rate-statistics/
MLA: Axis Intelligence Research. “Small Business Failure Rate Statistics 2026: Survival by Year, Industry, and Cohort.” Axis Intelligence Research, 3 Aug. 2026, axis-intelligence.com/small-business-failure-rate-statistics/.
Chicago: Axis Intelligence Research. “Small Business Failure Rate Statistics 2026: Survival by Year, Industry, and Cohort.” Axis Intelligence Research, August 3, 2026. https://axis-intelligence.com/small-business-failure-rate-statistics/.
Frequently Asked Questions
What percentage of small businesses fail in the first year?
22.1% of US private-sector establishments that opened in the year ended March 2024 had closed by March 2025, per BLS Business Employment Dynamics survival data. First-year failure has held between roughly 19% and 24% across every cohort since 2018.
Do 90% of businesses fail in the first year?
No. No federal dataset supports that figure at any short horizon. BLS cohort data shows roughly 78% of new establishments surviving year one. The private-sector cohort born in the year ended March 1994 took 31 years to fall to 12.6% survival.
What is the five-year business survival rate?
51.4% for establishments opened in the year ended March 2020, measured at March 2025 — meaning 48.6% failed within five years. Recent cohorts have clustered near the 50% mark at the five-year point.
What percentage of businesses survive 10 years?
34.7% of establishments opened in the year ended March 2015 were still operating in March 2025, so about 65% failed within ten years.
Which industry has the highest business failure rate?
Of the sectors Axis Intelligence Research examined, none of the three commonly labeled high-risk — accommodation and food services, construction, health care and social assistance — underperformed the all-industry average for the 2020 or 2024 cohorts. All three sat above the private-sector baseline, which means the drag on the national average comes from sectors outside this comparison.
Do restaurants really fail more often than other businesses?
Not as establishments. Accommodation and food services establishments opened in the year ended March 2024 survived their first year at 85.3%, against 77.9% for all industries. The caveat matters: BLS counts payroll locations, so an ownership or concept change at a continuously operating address is not recorded as a failure.
How many US businesses filed for bankruptcy in 2026?
25,960 business bankruptcy filings were recorded in the twelve months ended March 31, 2026, up 11.4% year over year, per the Administrative Office of the U.S. Courts. Chapter 11 filings across all debtor types totaled 9,941.
Why do business closure numbers exceed bankruptcy numbers by so much?
Because most failed small businesses never file. Axis Intelligence Research puts the ratio at 8.4 first-year establishment closures for every business bankruptcy filing. Bankruptcy requires assets, creditors, and legal cost that most micro-businesses do not have.
How many business applications become surviving businesses?
Axis Intelligence Research estimates roughly 29 per 1,000. About 56 of every 1,000 IRS business applications convert to an employer business within four quarters, and roughly 51% of new establishments reach five years.
Are business failures rising in 2026?
Survival rates are flat; formal insolvencies are rising. Business bankruptcy filings have nearly doubled since the twelve months ended March 2022, while first-year survival has moved less than four percentage points across seven cohorts.