Franchise Failure Rate: What the Research Actually Shows

Franchise failure rate studies, graded

Sixteen studies and official data series compare franchise and independent outcomes or measure franchise failure directly. One earns an A grade, eleven earn a B and four earn a C under the rubric below. The table lists each study’s design, sample, period, outcome measure and headline finding, and how we checked it.

Two definitions matter throughout. A closure (or discontinuance) means a business stopped operating. A failure implies a loss; Headd (2003) found that about a third of closed businesses in Census data were successful at closure, so closure rates overstate failure for all business types. Studies also differ on the unit counted: a franchise unit can survive under a new owner while the original franchisee’s firm has exited.

How we graded each study.

• A: nationally representative or population-scale microdata that follow new businesses from startup, compare franchise and independent businesses directly, and control for who chooses to franchise. We read the full text.
• B: large or representative data with a clear design but one major limit: older data, firm-level only, a small franchise subsample, loan losses instead of closures, no comparison group, or only the abstract available to us.
• C: franchisor-reported or self-selected data, commentary or a review without new data, or a finding we could only verify through another paper’s summary.

The strongest study: a first-year edge that does not keep growing

The highest-quality evidence finds a real but small survival advantage for franchised startups, concentrated in the first year.

Lafontaine, Zapletal and Zhang (Census CES Working Paper 17-21, 2017) linked the Census Survey of Business Owners to the Longitudinal Business Database for 5,240 single-establishment startups from 2002 and 1,690 from 2007, in retail, accommodation and food, and other services. In the 2002 cohort, 98.3% of franchised startups survived one year against 92.3% of independents. Two-year survival was 90.4% against 81.7%.

After controls for owner education, age, financing and location, the franchise advantage was 5.1 percentage points at one year and 6.4 points at two and three years. Among businesses that survived year one, franchised businesses were 2.1 points more likely to survive year two (significant only at the 10% level), and the year-three difference (1.0 point) was not statistically different from zero. The published version (Journal of Economics and Management Strategy, 2019) describes the difference as small compared with claims in the trade press and short lived.

Limits: the sample covers single-establishment, majority-owned businesses in franchise-intensive industries only, and the 2007 cohort is small. The Census data record a business sold to a new owner as an exit, which the authors argue makes their estimate of the franchise advantage conservative.

Older Census studies: franchises did no better, and sometimes worse

Census research on businesses started in the 1980s found young franchise firms closed more often than young independent firms, while individual franchise units owned by established operators did well.

Bates (Census CES Working Paper 94-3, 1994) followed 20,554 young firms from the Census Characteristics of Business Owners survey. Of firms started from 1984 to 1987 and active in 1987, 34.7% of franchisee firms and 28.0% of independents had closed by late 1991, a survival rate of 65.3% against 72.0%. The franchise firms started larger, with mean startup capital of $86,493 against $29,822. In retail, Bates (1995) reported that 45.1% of young franchise firms had closed by 1991 against 23.4% of young independent retailers.

Bates (Census CES Working Paper 97-5, 1997) reconciled the conflicting evidence. Among 52,088 young corporate restaurant establishments opened in 1986 and 1987, franchise units survived better than independents, but 84% of the new franchise units belonged to multi-establishment companies, most of them mature. Among young firms, 38.1% of franchisees and 31.9% of independents had closed by 1991. In Bates’s words, among true newcomers franchise survival rates were low, while units of entrenched multi-unit franchisees survived at high rates.

This work is contested. Castrogiovanni and Justis (2007) argue that flaws in Bates (1995) allow other interpretations, and Stanworth and colleagues (1998) concluded from US and UK evidence that franchise and conventional failure rates are strikingly similar once definitions are fixed.

Newer survey evidence: no significant difference by 2011

The most recent peer-reviewed comparison we found, published in 2024, shows franchised startups ahead in the first year and behind by the seventh, with no statistically significant difference.

Cotei and Farhat (Journal of Entrepreneurial Finance, 2024) used the Kauffman Firm Survey, which tracked 4,928 businesses started in 2004 through 2011. By 2011, 60% of franchise startups, 63% of independent startups and 74% of purchased existing businesses were still operating. The test comparing franchise and independent survival curves gave p = 0.83, meaning no significant difference. Franchises were 2.77 times as likely as independents to exit through a sale or merger rather than a closure.

Limit: the survival comparison includes only 53 franchise startups (representing about 1,697 businesses), so the franchise estimates are imprecise. The article’s introduction transposes the 60% and 63% figures; we use its Table 3. An earlier Kauffman study by Welsh, Desplaces and Davis (2011) likewise concluded that franchises face impediments that may affect first-year survival.

Franchisors fail too: evidence on new franchise systems

A franchisee’s outcome depends partly on whether the franchise system itself survives, and young systems exit often.

Bates (1997) summarizes Shane (1996): of franchisors that first sold franchises in 1983, 24.6% were still operating by 1993. Shane and Foo (1999) studied 1,292 new franchisors established from 1979 to 1996 and found that institutional legitimacy helps explain which ones survive. Kosova and Lafontaine (2010) found that chain age and size predict franchised-chain survival, and Madanoglu and Castrogiovanni (2018), using nearly 5,000 franchising firms, found a U-shaped link between the share of outlets a system franchises and network failure.

At the franchisee level, Michael and Combs (2008) found across 88 restaurant chains that franchisor contract design and strategic resources reduce franchisee failure. The practical point for a buyer: system age, size and franchisor support are measurable in the Franchise Disclosure Document, and the research links them to survival.

New evidence: SBA 7(a) loan outcomes, FY2010 to FY2019

SBA loans to franchise businesses were charged off somewhat more often than loans to non-franchise businesses in the FY2010 to FY2019 cohorts, and the gap comes from startups.

We analysed every 7(a) loan approved from FY2010 to FY2019 and disbursed, using the SBA 7(a) FOIA file as of June 30, 2026. A loan is a franchise loan when the lender recorded an SBA Franchise Directory code. Of 42,558 franchise loans, 8.7% had been charged off, against 7.0% of 436,405 non-franchise loans. Weighting non-franchise charge-off rates to the franchise mix of 6-digit industries and approval years raises the non-franchise comparison to 7.3%, so industry mix explains only a small part of the gap.

The gap depends on the business’s age at approval. For startups and businesses under one year old, 11.3% of franchise loans were charged off against 7.7% of non-franchise loans (8.8% after industry adjustment). For existing businesses two years or older, the pattern reverses: 4.9% against 6.6%. Change-of-ownership loans were level at 4.6% against 4.5%. By size, the gap is widest for loans of $150,000 to $499,999 (11.2% against 6.7%), and franchise loans of $1,000,000 or more were charged off less often (3.1% against 4.0%).

Our franchise flag matches SBA’s own count closely. GAO reported 32,323 franchise 7(a) loans worth $10.6 billion for FY2003 to FY2012; the FOIA file shows 31,941 disbursed franchise-coded loans worth $10.6 billion for the same years (98.8% of GAO’s count). For those earlier cohorts, which include the 2008 to 2009 recession, charge-off rates were level: 22.4% for franchise loans and 22.3% for non-franchise loans. GAO (2013) separately reported that SBA paid its guaranty on about 28% of the FY2003 to FY2012 franchise loans.

For more on how franchise borrowers use SBA loans, see our companion report on SBA franchise loans.

Charge-off rates in the five largest franchise sectors

Within accommodation and food services, the largest franchise sector in SBA lending, franchise and non-franchise loans were charged off at the same rate, 7.8%. Retail franchise loans did slightly better (7.4% against 8.1%). The franchise gap is concentrated in other services (10.5% against 6.7%) and arts, entertainment and recreation, which includes fitness (10.7% against 8.7%).

Survival in the most franchised sector (BLS, sector level)

Establishments in accommodation and food services, the sector that holds 46.3% of franchise establishments, have survived slightly longer than private-sector establishments as a whole since the 1990s. This is a sector-level comparison and says nothing directly about franchises.

The 2017 Economic Census counted 498,234 franchise establishments, of which 230,689 were in accommodation and food services; franchise establishments made up 36.0% of establishments in the industries of that sector covered by the franchise tabulation. BLS Business Employment Dynamics tables, last updated December 17, 2025, follow every cohort of new private establishments from the year ended March 1994. Averaged across cohorts, 82.9% of new accommodation and food establishments survived one year against 78.7% of all private establishments (31 cohorts), 53.6% survived five years against 49.3% (27 cohorts), and 37.7% survived ten years against 33.9% (22 cohorts).

The sector gap widened for recent cohorts: for establishments opened in the year ended March 2020, five-year survival was 59.3% in accommodation and food against 51.4% for all private establishments. BLS counts establishments, so a new outlet of an existing chain counts as a birth, and BED has no franchise flag. The figures fit the Bates (1997) finding that units of established multi-unit operators survive well; they cannot show whether franchising causes the difference.

Where the claim that franchises rarely fail comes from

Every ‘franchises rarely fail’ statistic we could trace rests on data supplied by franchisors, not on tracking franchisees. The earliest figure we found is an IFA-published 1968 estimate; the most cited is the U.S. Commerce Department’s Franchising in the Economy series, a survey of franchisors.

Bates (1997) documents the history. Atkinson’s 1968 book, published by the International Franchise Association (IFA), put franchise discontinuance at 1.6% a year; a 1971 US Senate Select Committee on Small Business report called its comparisons with retail failure rates erroneous. Ashman (1988) cited the Commerce Department for a claim that 92% of franchises were in business after five years against 23% of independents. Staff of the House Committee on Small Business reviewed the Commerce reports and found no comparable failure or success rates; the reports noted that the number of failures was unknown (1992 hearing, as quoted by Bates).

Castrogiovanni, Justis and Julian (1993) described the Commerce surveys, conducted by Andrew Kostecka, as the primary referent and said they indicated that fewer than 4% of franchises fail each year, a figure the authors said needed corroboration. Entrepreneur (2013) reported that the survey covered nearly 2,000 franchisors who chose to respond, that the data were not audited, and that some analysts read it as 5% of units closing over five years, which became the 95% success claim. The same article reported that the IFA asked franchisors and brokers to stop using the figure.

Later industry figures share the design. An IFA-sponsored 1992 Arthur Andersen survey reported that 96.9% of franchise units opened in the prior five years were still operating, based on franchisor responses. An IFA Educational Foundation study of 444 systems reported mean unit discontinuance of 8.10% in 1994 (median 4.65%), higher than earlier industry figures. We found no SBA publication behind the widely repeated claim that the SBA puts franchise five-year success at about 90% against 70% for independents; the SBA Office of Advocacy (February 2026) reports 49.2% five-year survival for all new establishments from 1994 to 2022, with no franchise split.

What FDD Item 20 can and cannot tell you

The closest thing to a current, brand-by-brand closure measure is Item 20 of the Franchise Disclosure Document. The FTC Franchise Rule requires each franchisor to report, for its last three fiscal years, outlets opened, terminated, not renewed, reacquired and closed for other reasons. These are franchisor-reported counts. A third-party analysis by VetMyFranchise (2026) of Item 20 tables in 858 recent FDDs reported a median one-year closure rate of 4.7% of franchised units, with the 90th percentile at 16.2%; we did not check that analysis. A one-year unit closure rate in established systems is a different measure from the survival of a new franchisee’s business, so it should not be compared with startup survival rates. A separate report on franchisee turnover is planned. See also our guide to buying a franchise.