Restaurant failure rates: what the BLS survival data actually shows

What is being counted?

The industry is NAICS 72: accommodation and food services. It includes lodging as well as restaurants, caterers, bars and other food-service establishments. It does not isolate independent restaurants from chains. Treating the sector percentage as the chance that one independent restaurant will fail adds precision the table cannot support. BLS industry definition

The unit is an establishment, rather than an entire company. A new branch of an existing chain counts as a new establishment even though the company is already operating. BLS determines age from the first reported positive employment and follows March observations. Its opening cohorts are labeled by the year ended March, not the calendar year printed on a restaurant's opening announcement. These are also not exact first-anniversary checks for each location. BLS age and survival methodology

The underlying Business Employment Dynamics data come from establishments covered by state unemployment insurance programs. The table is an employment-based business record, not a survey asking owners whether a location was profitable. It contains no bankruptcy flag, owner return or reason-for-closure column. BLS data source, survival table

The latest one-, five- and ten-year observations

The current sector table, checked October 11, 2026, runs through March 2025. Here are three observations ending in that month:

Source: BLS Table 7, accommodation and food services, linked from the establishment age and survival index. Rates are rounded to one decimal place by BLS. The check date is not the observation period: these are not full-year 2026 results.

These rows describe three different groups. They are useful for seeing the latest available observation at each duration. They are not three stages in the life of the same set of restaurants. Do not join them into a survival curve or interpret their differences as the effect of a particular business practice.

The table also contains a March 2025 opening cohort, but only its initial observation. That row cannot yet supply a one-year survival rate. Likewise, a newer quarterly jobs release does not automatically mean this annual cohort file has been updated.

Follow one cohort to understand the denominator

For a consistent group, follow the establishments in the year-ended-March-2020 cohort:

Source: the same BLS cohort table. The entire sequence includes the pandemic period; it is an observed history, not a forecast for a new location.

Two calculations answer different questions:

• —Survival since the initial observation: 29,826 ÷ 50,259 × 100 ≈ 59.3%. Its complement is approximately 40.7% over the five-year interval.
• —Survival relative to the previous year's survivors: 29,826 ÷ 32,956 × 100 ≈ 90.5%. Its complement is approximately 9.5% for the last interval.

Calling that second complement a 40.7% annual failure rate would mix periods and denominators. Neither calculation identifies why an establishment stopped appearing among survivors. Employment-based openings and closings can include interrupted operation; BLS links records across ownership changes and business reorganizations. Do not turn a decline in the count into a claim that every missing establishment went bankrupt. BLS linking and closing definitions, reopening explanation

Turn the risk question into a sales requirement

Suppose a hypothetical operator is evaluating another location with these monthly assumptions:

• —$30,000 in fixed operating costs, including rent, management and the minimum crew they must staff.
• —Variable costs equal to 60% of sales, including ingredients, additional sales-dependent labor and variable fees. None of the minimum crew is counted twice.
• —Average sales of $25 per guest, excluding sales tax and tips passed through to staff.
• —26 open days per month.

The amount remaining after variable costs is 40 cents per sales dollar. The calculation is:

$30,000 ÷ (1 − 0.60) = $75,000 per month

At $25 per guest, that is 3,000 monthly covers, or about 115.38 covers per open day. If planning for the same whole number of guests each day, round up to 116; across a real month, the total matters and daily traffic will vary.

Now suppose the variable-cost assumption rises to 65%, with everything else unchanged:

The second budget needs about $10,714.29 more monthly sales to cover the modeled costs. That is scenario arithmetic, not evidence that a price increase will produce the additional sales. Check the guest count against available seats, service hours and the demand you can substantiate for that location.

If actual monthly sales were only $65,000, the same models would leave −$4,000 and −$7,250, respectively, after their modeled variable and fixed costs. Those results explain a budget gap; they do not translate into a probability of business failure. Opening expenses, capital purchases, debt principal and other cash needs are outside this example. A full cash plan needs those separately, with no expense counted twice.

Test your location's assumptions

Use the free restaurant break-even calculator with 30,000 fixed costs, 60% variable costs, a 25 average check and 26 open days. Then change only the variable share to 65%. The tool's daily result rounds to the nearest whole cover; the example above rounds upward for an equal daily planning target. Enter average spend per guest for a covers calculation, rather than the total bill for a party of several guests.

Replace the teaching inputs with your own budget before using the result. The restaurant break-even guide explains how to separate fixed and variable costs. If the ingredient part of your budget is uncertain, start with current supplier prices and complete quantities and yields for a dish in the recipe cost calculator. An ingredient-only margin excludes the other costs in this break-even model.

Marji is restaurant food cost software. These calculators can help you examine assumptions; they do not estimate your location's survival probability or guarantee profitability.