Founder Data ReportWeather Doesn’t Hit Your Business Once. It Bills You Twice a Year.
Join us Fridays at 12 PM ET to ask TLE CEO Michael Dermer your questions live.
Federal survey data has tracked small-business weather losses every two weeks for three years. The line is not a series of disasters. It is a season, it peaks in January, and it has done so three years running.
Quick answer: Yes, and on a schedule. In the six months to August 2026, 6.9% of businesses with 1 to 4 employees reported losing money to an extreme weather event. Six months earlier that reading was 13.0% – the annual peak, reached in January or February in each of the last three years. The cause changes with the calendar too: among small businesses reporting a loss, snow and ice were named by 77.2% in February 2026 and by only 14.5% in August, while extreme heat or cold rose from 27.5% to 51.1%.
- Census BTOS asks every employer business, every two weeks, whether extreme weather cost it money in the last six months. The series now runs three years.
- For businesses with 1 to 4 employees it peaks every January. 5.0% to 9.9% in 2024, 8.8% to 12.6% in 2025, 4.9% to 12.8% in 2026. Three for three.
- Peak to trough is a 2.7x swing: 13.0% in February 2026 against 4.8% at the end of December 2025, and 2.9% at the low.
- The event mix flips with the season. Snow and ice: 77.2% of loss reports in February 2026, 14.5% in August. Extreme temperature: 27.5% to 51.1%.
- The events people insure against are the rare ones. In the latest cycle hurricanes were named in 4.4% of small-business loss reports and earthquakes in 0.7%.
- The smallest businesses report the LOWEST loss rate – 6.9% against 9.8% at 20-49 employees. That is dollars lost, not disruption suffered.
It is not an event. It is a season.
Weather gets discussed in small business the way earthquakes get discussed: a thing that either happens to you or does not, and if it does you find out whether your policy was any good. That framing is why almost nobody plans for it. You cannot budget for a coin flip.
The Census Bureau has been quietly undermining that framing since 2023. Its Business Trends and Outlook Survey asks every employer business, every two weeks, whether extreme weather cost it money in the previous six months. Three years of fortnightly readings is long enough to see a shape, and the shape is not random. For businesses with 1 to 4 employees the reading climbs into January, peaks in January or February, falls through spring, lifts again in late summer, and bottoms out in late autumn. Then it does it again.
The January step is the clearest part of it. In the fortnight ending Jan 11, 2026, 4.9% of the smallest businesses reported a weather loss. Four weeks later it was 12.8%. The same step appears in January 2024 (5.0% to 9.9%) and January 2025 (8.8% to 12.6%). Three years, three steps, same fortnight.
That last point deserves a word about how the question is built, because it is what makes the reading fair. Businesses are asked about the last six months, so any two neighbouring readings share five and a half months of history. A step between two adjacent cycles cannot come from old weather that was already inside both windows. It comes from the weeks that were just added – which, in this case, is January.
Each point is a six-month look-back, so the line is a rolling window rather than a set of independent periods. The three marked peaks all fall in January or February. The shaded band is a join, not a reading: the x-axis is spaced by survey cycle rather than by date, and the three cycles cancelled by the October-November 2025 funding lapse are missing entirely, so about six weeks of calendar closes to nothing there.
Source: U.S. Census Bureau, Business Trends and Outlook Survey (BTOS), question 14 (monetary losses from an extreme weather event in the last six months) and question 15 (which event caused them), tabulated by employment size class. Downloaded from census.gov/hfp/btos/downloads on September 6, 2026. Latest cycle 202617, collected August 10-23, 2026; the series runs fortnightly back to September 2023. Two cautions apply to every figure here. First, question 14 asks about the LAST SIX MONTHS, so neighbouring fortnightly readings share five and a half months of window and are not independent observations – a step between two adjacent cycles means the newly added weeks did the work. Second, question 15 is asked ONLY of businesses that answered yes to question 14, so its shares are of that subset rather than of all businesses, and it is a select-all question, so the shares do not sum to 100. Estimates the Census Bureau suppressed for disclosure are omitted rather than read as zero. Three cycles in October and November 2025 were not collected because of the federal funding lapse. BTOS measures reported MONETARY loss, which is not the same thing as disruption.
Two seasons, two different bills
If the rate were the only thing that moved, this would be a scheduling note and not much more. What makes it operationally useful is that the cause moves with it. The survey asks businesses that reported a loss which kind of event caused it, and the answer in February is a different answer from the one in August.
In February 2026, 77.2% of small businesses reporting a loss named snow, blizzard or ice. By August that was 14.5%. Running the other way, extreme temperature – heat or cold – went from 27.5% to 51.1%, and is the single most-named cause in the latest cycle. These are not two readings of one risk. They are two risks that take turns.
Read the two columns below as a job list rather than a weather report. The winter column is a story about access and utilities: staff who cannot get in, a delivery that does not arrive, a pipe, a closure. The summer column is a story about load and spoilage: cooling that runs all day, stock that will not keep, outdoor work that stops. Almost nothing you would do about one helps with the other.
Neither column is the good outcome, so neither is highlighted. Shares are of the businesses that reported a loss in that cycle – roughly one in eight in February and one in fourteen in August – and the question is select-all, so they do not sum to 100.
Source: U.S. Census Bureau, Business Trends and Outlook Survey (BTOS), question 14 (monetary losses from an extreme weather event in the last six months) and question 15 (which event caused them), tabulated by employment size class. Downloaded from census.gov/hfp/btos/downloads on September 6, 2026. Latest cycle 202617, collected August 10-23, 2026; the series runs fortnightly back to September 2023. Two cautions apply to every figure here. First, question 14 asks about the LAST SIX MONTHS, so neighbouring fortnightly readings share five and a half months of window and are not independent observations – a step between two adjacent cycles means the newly added weeks did the work. Second, question 15 is asked ONLY of businesses that answered yes to question 14, so its shares are of that subset rather than of all businesses, and it is a select-all question, so the shares do not sum to 100. Estimates the Census Bureau suppressed for disclosure are omitted rather than read as zero. Three cycles in October and November 2025 were not collected because of the federal funding lapse. BTOS measures reported MONETARY loss, which is not the same thing as disruption.
The events you insure against are the rare ones
Rank the causes in the latest cycle and the order is almost the reverse of the order they appear on the news. Extreme temperature leads at 51.1%. Severe storm or hail is next. Then high winds, then flood, then snow. Hurricane, the event that owns the phrase ‘weather risk’ in most people’s heads, was named in 4.4% of small-business loss reports. Earthquake was named in 0.7%.
That ordering is not an argument against insurance. It is an argument about where the unbudgeted money actually goes. A hurricane produces a claim, a file and a conversation with an adjuster. A fortnight of heat that spoils stock and doubles a cooling bill produces none of those – it produces a slightly worse month that nobody writes down as weather. Multiply that by twice a year, every year, and it stops being an act of God and starts being a line item you have simply never named.
A hurricane produces a claim. A fortnight of heat produces a slightly worse month that nobody writes down as weather.
Extreme temperature is marked only because it leads this particular cycle, not because it is the standing top cause – in February snow and ice led it by more than two to one. Bars are scaled to the highest share shown. Shares are of the small businesses that reported a loss, and the question is select-all, so they sum to well over 100. A further 12.8% chose ‘other’ and described the event themselves; that answer is not charted here because it names no event.
Source: U.S. Census Bureau, Business Trends and Outlook Survey (BTOS), question 14 (monetary losses from an extreme weather event in the last six months) and question 15 (which event caused them), tabulated by employment size class. Downloaded from census.gov/hfp/btos/downloads on September 6, 2026. Latest cycle 202617, collected August 10-23, 2026; the series runs fortnightly back to September 2023. Two cautions apply to every figure here. First, question 14 asks about the LAST SIX MONTHS, so neighbouring fortnightly readings share five and a half months of window and are not independent observations – a step between two adjacent cycles means the newly added weeks did the work. Second, question 15 is asked ONLY of businesses that answered yes to question 14, so its shares are of that subset rather than of all businesses, and it is a select-all question, so the shares do not sum to 100. Estimates the Census Bureau suppressed for disclosure are omitted rather than read as zero. Three cycles in October and November 2025 were not collected because of the federal funding lapse. BTOS measures reported MONETARY loss, which is not the same thing as disruption.
Why the smallest number here is the one to be careful with
Cut the latest cycle by size and the smallest businesses report the lowest loss rate in the economy: 6.9% at 1 to 4 employees, against 9.8% at 20-49 and 8.3% at 250 or more. The spread across every size band is 2.9 points.
It would be easy, and wrong, to read that as smaller businesses being less exposed. The question asks about monetary loss. A two-person consultancy with a laptop and no premises has very little that weather can put a dollar figure on; a twenty-person operation with a building, a van, stock and a payroll that runs whether the doors open or not has a great deal. The gap is at least partly a gap in what there is to lose, not in what gets disrupted.
Which is the actual planning point. If you are at 1 to 4 people today and intend to be at 20 in three years, you are not looking at your own number in that chart. You are looking at 9.8% – and at a fixed cost base that will not pause when the weather does.
The 1-4 band is marked because it is this article’s comparison point, not because a low reported-loss rate is the good outcome. It is not: this question measures dollars lost, and a business with less to lose reports less without being any safer. Bars are scaled to the highest rate shown, so a series that only spans 2.9 points still fills most of the track – read the printed values, not the bar lengths.
Source: U.S. Census Bureau, Business Trends and Outlook Survey (BTOS), question 14 (monetary losses from an extreme weather event in the last six months) and question 15 (which event caused them), tabulated by employment size class. Downloaded from census.gov/hfp/btos/downloads on September 6, 2026. Latest cycle 202617, collected August 10-23, 2026; the series runs fortnightly back to September 2023. Two cautions apply to every figure here. First, question 14 asks about the LAST SIX MONTHS, so neighbouring fortnightly readings share five and a half months of window and are not independent observations – a step between two adjacent cycles means the newly added weeks did the work. Second, question 15 is asked ONLY of businesses that answered yes to question 14, so its shares are of that subset rather than of all businesses, and it is a select-all question, so the shares do not sum to 100. Estimates the Census Bureau suppressed for disclosure are omitted rather than read as zero. Three cycles in October and November 2025 were not collected because of the federal funding lapse. BTOS measures reported MONETARY loss, which is not the same thing as disruption.
What to actually do with a number that has a calendar
A risk with a season can be handled the way you handle any other seasonal cost, which is to say: in advance, in the diary, and cheaply. Three things follow from the data rather than from general advice.
Put the two peaks in the calendar, not the budget. The readings say December and midsummer are the last quiet fortnights before each climb. That is when a generator gets serviced, a supplier gets a second source, and somebody checks whether the cover you bought in 2019 still matches the stock you hold now.
Name the small losses. The reason weather feels like an act of God is that only the catastrophic version ever gets recorded as weather. Give it a line in your accounts – spoilage, a lost day, an energy spike – and within a year you will know your own number instead of the national one.
Plan for the size you are growing into. The loss rate rises with headcount up to 20-49 employees, and the reason is that growth buys you premises, stock and fixed costs. That is worth knowing before the first lease, not after the first closure.
None of this makes the weather any more predictable. It makes your response to it predictable, which is the part that was ever in your control.
Frequently Asked Questions
How many small businesses lose money to extreme weather?
In the six months to August 2026, 6.9% of US businesses with 1 to 4 employees reported a monetary loss from an extreme weather event, according to the Census Bureau's Business Trends and Outlook Survey. That figure moves with the season: it reached 13.0% in February 2026 and fell to 2.9% in late November 2025. Across all employment sizes the latest readings run between 6.9% and 9.8%.
What time of year are small business weather losses worst?
January and February. In each of the last three years the reported-loss rate for businesses with 1 to 4 employees stepped up sharply across January: from 5.0% to 9.9% in 2024, 8.8% to 12.6% in 2025, and 4.9% to 12.8% in 2026. There is a second, smaller rise in late summer. The quietest readings fall in late spring and late autumn.
What kind of weather actually costs small businesses money?
Mostly the mundane kind, and it changes with the season. In the August 2026 cycle, small businesses reporting a loss named extreme temperature most often (51.1%), followed by severe storm or hail and high winds. In February 2026 the leader was snow, blizzard or ice at 77.2%. Hurricanes were named in 4.4% of small-business loss reports in the latest cycle and earthquakes in 0.7%. Note that this question is asked only of businesses that reported a loss and allows more than one answer.
Are small businesses more exposed to extreme weather than large ones?
Not on reported monetary loss, which is what this survey measures. Businesses with 1 to 4 employees reported the lowest rate in the latest cycle at 6.9%, against 9.8% at 20-49 employees and 8.3% at 250 or more. The likely reason is that the very smallest businesses have less that weather can put a price on – no premises, no stock, no fleet – rather than that they are disrupted less often.
Where does this data come from and how often is it updated?
The Census Bureau's Business Trends and Outlook Survey, a biweekly survey of US employer businesses running since 2023. Questions 14 and 15 cover weather-related monetary losses and the event that caused them, and both are published cut by employment size class. The files are free at census.gov/hfp/btos/downloads and need no API key. The figures in this article come from cycle 202617, collected August 10-23, 2026.
Published by The Lonely Entrepreneur — the community and coaching platform for entrepreneurs who are building alone. lonelyentrepreneur.com
This article is for educational purposes and is not a substitute for professional financial or legal advice.