Every major climate hazard on a risk report damages something you own. Flood ruins the ground floor and the contents in it. Wind peels off the roof. Wildfire takes the whole structure. In each case there is a building to repair, a claim to file, a photograph of the loss.
Heat is the exception. A heatwave can run for a week and leave the building untouched: no broken windows, no water line on the wall, nothing to insure against. And yet the business inside it loses real money. The loss hides inside the working day. People slow down. They take more breaks. The afternoon shift gets less done than the morning one. By the time the heat passes, output is down and payroll is unchanged, but there is nothing to point a camera at.
That makes heat the hardest hazard to take seriously and the easiest to underprice. So here is the question this piece answers: if there is nothing to repair, how do you put a number on it?
Heat business interruption, the labour-productivity cost of extreme heat, is the rare climate loss with nothing to photograph: the building is fine, but the work slows down. You can still put a dollar figure on it. Read a heat-stress profile against a work-capacity curve, weigh the lost hours by the productive labour at risk, and add it up across the year, the same accounting flood uses for expected annual damage. The catch: the curves disagree, so any credible number has to travel with its method.
Heat turns into a cost through one channel: lost work
For a business, the heat bill is mostly a labour bill. As the air gets hotter and more humid, the human body has to spend more of its energy shedding heat and less of it doing work. Past a point, a worker physically cannot keep up the same pace without overheating, so the body forces the only fix available: rest. More of every hour goes to recovering instead of producing. That lost time is the cost.
To measure it you need the right thermometer, and the everyday one on the wall is not it. What actually drives heat strain is a combination of temperature, humidity, sun, and wind, not the air temperature alone. Humidity matters most. Sweat is how the body sheds heat, and sweat only works if it can evaporate; in humid air it cannot, so a humid 32°C day is far more punishing than a dry 32°C day. Risk professionals use a heat-stress measure that folds all four factors into a single number. Think of it as the temperature your body actually feels while working, not the temperature on the thermometer. It is the heat equivalent of flood depth: the single intensity number that the damage curve is read against.
For physical work, the loss curve is gentle, then a cliff, then flat
Take heavy outdoor work first: construction, warehousing, farming, anything sustained and physical. Researchers have measured how much work capacity people lose as the heat-stress number climbs, and the curve has a consistent shape. Nothing happens for a long while, then loss sets in and accelerates fast, then it flattens near the top because there is little capacity left to lose.
The rough anchors for heavy work, on the heat-stress measure:
| Heat-stress level | Work capacity lost |
|---|---|
| Below about 26°C | Little to none |
| About 27°C | About 10% |
| About 32 to 33°C | About half |
| About 38°C | Almost all |
The shape is the important part: a long flat start, a steep middle, a saturating top. The middle is where a modest rise in conditions produces a large jump in lost work. A site sitting at the foot of that climb is far more exposed to one more degree of warming than the average temperature alone would suggest.
These anchors come from the occupational-health literature on physical labour. Separately, a closed-form curve fit to 338 controlled lab work sessions traces the same shape, with a high goodness of fit: independent evidence that the pattern is measured, not assumed.
The part most heat ratings miss: same building, different loss
Here is where heat stops behaving like other hazards. Flood damage depends on the building and where it sits. Heat damage depends on the building, where it sits, and what people do inside it. It is a two-dimensional problem, and the second dimension is the one that gets dropped.
The reason is physics. A person doing heavy manual work generates roughly twice the internal body heat of someone at a desk. That internal heat stacks on top of the heat coming from the environment. So at the very same outdoor conditions, the labourer reaches the danger zone long before the office worker: the office worker’s curve sits well to the right of the labourer’s. Same weather, same coordinate, two completely different loss curves.
What that means in practice:
Two businesses at the same address, on the same hot afternoon:
- Office (desk work): a few percent of capacity lost
- Warehouse or yard (heavy work): roughly a third lost
Illustrative, to show the shape of the effect, not a figure from our live system. Real numbers depend on local humidity, sun, and the work being done. The heat-stress measure also runs several degrees below the temperature on the thermometer, so a “hot afternoon” lands lower on the curve than the air temperature suggests.
This is why a heat rating pinned only to a map coordinate, the standard output across most climate-risk tools, including a qualitative “High”, can only ever be half an answer. The location tells you the weather. It cannot tell you whether the building is full of people lifting and loading or people typing. What the workforce actually does is the other half of the exposure, and any number that skips it is describing a different building than the one being underwritten.
From lost hours to dollars
Once you have a loss curve and a heat-stress profile for a location, the path to money is direct. The curve converts conditions into a fraction of work lost. Multiply that fraction by the productive labour at risk during those hours, and you have lost output: the heat business interruption cost.
Three refinements keep the figure defensible rather than alarmist:
- A full shift loses more than a single hot hour. Heat strain accumulates over the day as core temperature and dehydration build, so an estimate based on one peak hour undercounts a long shift.
- Some of the loss is recoverable. Shifting the heaviest work out of the two or three hottest hours claws back a meaningful share, though that escape route narrows as the climate warms and the cool hours disappear.
- Acclimatised workers hold up better. People regularly exposed to heat tolerate more of it than those who are not, which is why the same conditions cost a desert worksite less than a temperate one facing its first heatwave.
The annual number comes the same way a flood’s expected annual damage does. For flood, you weigh the damage at each depth by how often that depth occurs and add it up. For heat, you weigh the work lost at each level of heat by how many hours a year the location spends at that level, and add it up. Different hazard, same accounting: a loss curve integrated over an exposure distribution, ending in one defensible annual dollar figure.
| Same accounting | Flood | Heat |
|---|---|---|
| The intensity you measure | How deep the water gets | How hard the heat feels while working |
| What the damage curve outputs | Share of the building’s value lost | Share of the working hours lost |
| Added up over | How often each flood depth occurs | How many hours a year reach each level of heat |
| The result | Expected annual damage, in dollars | Expected annual cost of lost work, in dollars |
A heat number has to come with its method
Flood has FEMA’s HAZUS: a single, government-backed set of depth-to-damage curves that everyone can point to. Heat has no equivalent. The published work-capacity curves disagree by as much as a factor of two on where loss begins and how steeply it falls. Some of that is real scientific uncertainty; some is that the curves answer slightly different questions: a few are workplace safety limits that mandate rest at set thresholds, others are direct measurements of how much work still gets done.
That disagreement is not a reason to avoid the number. It is the reason any credible heat-cost figure must travel with its method attached: which curve, which assumptions about the work being done, which heat-stress profile for the site. A heat business interruption estimate that hides those choices behind a single confident number is the one to distrust. A credible estimate shows its work and lets you change the inputs. On heat, transparency beats false precision, because there is no single authority to defer to.
What this means for risk and disclosure teams
A “High” heat rating is a starting point, not an answer. Climate disclosure regimes such as TCFD and IFRS S2 ask for the financial impact of physical risk, and heat business interruption is exactly that kind of line item: recurring, quantifiable, and growing. It can be estimated today, provided the estimate is anchored to a published method, the local heat-and-humidity profile, and the real work done at the site rather than a generic label. The teams that move first from a heat rating to a heat number are the ones who will have an answer ready when the question is asked.
The invisible peril you can still price
Heat is the climate hazard that costs you money without leaving a mark. That is precisely why it gets underpriced: there is no wreckage to force the issue. But the loss is real, it is measurable, and it follows the same logic as the physical hazards: an intensity measure, a damage curve, and an exposure distribution integrated into an annual figure.
At Continuuiti we quantify physical climate risk in dollar terms at the individual building level: expected annual damage, value-at-risk, return-period losses, starting with flood and FEMA’s depth-damage curves. Heat business interruption is the same machinery pointed at a different peril: the one that slows your people instead of flooding your floor. If you are pricing physical climate risk across a portfolio or building out climate disclosure, we welcome the conversation.
The work-capacity relationships described here come from published, peer-reviewed research: Foster et al. 2021 (a closed-form physical-work-capacity model, PWC = 100/(1 + 33.63·WBGT^−6.33), R² = 0.94); Parsons et al. 2021; the Kjellström/Hothaps labour-capacity lineage adopted by the WHO and the Lancet Countdown; and the ISO 7243 and ACGIH occupational heat-stress reference limits. “WBGT” is the Wet-Bulb Globe Temperature, the heat-stress measure referenced in plain language throughout. All figures in the worked sketch are illustrative. No client or location data was used.
Frequently Asked Questions
Can you put a dollar figure on heat business interruption?
Yes. Heat rarely damages the building, but it slows the work inside it. Read a heat-stress profile for the site against a work-capacity curve to get the fraction of work lost, multiply by the productive labour at risk, and add it up across the hours of the year. The result is an expected annual cost of lost work, the same accounting flood uses for expected annual damage.
What is the heat-stress measure, and why not just use air temperature?
Heat strain depends on temperature, humidity, sun, and wind together, not air temperature alone, and humidity matters most because sweat only cools you if it can evaporate. Risk professionals fold all four into one heat-stress number (technically WBGT, the Wet-Bulb Globe Temperature). It is the heat equivalent of flood depth: the single intensity value the damage curve is read against, and it runs several degrees below the thermometer reading.
Why do two businesses at the same address lose different amounts to heat?
Because heat damage depends on what people do inside, not just the weather outside. Heavy manual work generates roughly twice the internal body heat of desk work, so a warehouse crew hits the danger zone long before an office at the very same conditions. That second dimension is why a heat rating pinned only to a map coordinate can only ever be half an answer.
Does climate disclosure require the financial impact of heat?
Frameworks such as TCFD and IFRS S2 ask for the financial impact of physical risk, and heat business interruption is exactly that kind of recurring, quantifiable line item. It can be estimated today, provided the estimate is anchored to a published method, the local heat-and-humidity profile, and the real work done at the site rather than a generic label.
Is there a HAZUS for heat?
No. Flood has FEMA’s HAZUS, a single government-backed set of depth-to-damage curves. Heat has no equivalent: the published work-capacity curves disagree by as much as a factor of two on where loss begins. That is why any credible heat figure must travel with its method, which curve, which assumptions about the work, and which heat-stress profile, rather than hide behind one confident number.
