A dairy in Idaho’s Magic Valley with 12 workers lost 5 percent of the hours they were due to work, 576 hours over 24 weeks. Relief milkers for those hours cost $10,368, at an invented $18 an hour. If a few long absences caused the loss, that money fills the gap and leaves the cause on the job. The rate cannot tell you which it was. Before you hire relief, split the rate into how often people were away and how long each absence lasted, and write the longest absences and their causes on one page.
The absenteeism calculator does the arithmetic from your time sheets. The page, called the absence sheet here, holds what the calculator cannot work out. It lists the period you measured, what you counted as absence, the longest absences with their causes, and the jobs only one person can do.
How do I work out the absence rate on my farm?
The absence rate is the hours lost to absence divided by the hours scheduled, which are the hours the crew was due to work in the period. On the dairy, 12 workers scheduled for 40 hours a week over 24 weeks come to 960 hours each and 11,520 hours for the crew. With 576 hours lost, the rate is 5 percent, or 48 lost hours per worker.
What counts as absence is a choice you make before counting. The International Labour Organization (ILO), the United Nations agency for work, defines absence hours as the number of contractual hours of work not actually worked during a short reference period. Where no contract exists, it counts the number of hours usually worked but not actually worked.
The ILO list of reasons runs from illness and injury to bad weather, public or other holidays, or another reason. Your count does not need all of it, because rain and holidays are not something you can fix. Write down what you leave out before you calculate, or the choice turns into a defense of whatever number comes out.
The choice moves the result. In a study of 782 service workers at a Brazilian oil company, followed for three years, the average sick leave lasted 6.6 days with maternity leave counted. It fell to 5.64 days after maternity leave was taken out. That is almost 15 percent less, with nothing changed on the crew.
Divide by scheduled time, not by paid time or calendar days. A period with more scheduled days gives a smaller rate for the same absences. Compare periods of the same kind, because eight weeks at the peak of harvest and eight weeks off-season do not measure the same thing.
The same 5 percent can have two shapes. The rate is how often people were away times how long each absence lasted, divided by the scheduled hours. Dairy A has 4 absences per worker at 12 hours each, which is 48 hours, and 48 over 960 is 5 percent. Dairy B reaches the same 5 percent with 6 absences of 96 hours.
| What was counted | Dairy A | Dairy B |
|---|---|---|
| Absences recorded | 48 | 6 |
| Absences per worker on the crew | 4 | 0.5 |
| Average length of one absence | 12 hours, a day and a half | 96 hours, twelve working days |
| Workers absent at least once | 9 of 12 | 4 of 12 |
| Absences for each of those workers | 5.33 | 1.5 |
| Absence rate | 5 percent | 5 percent |
Before you count anything else, write what you count as absence and what you leave out.
Do many short absences or a few long ones carry the loss?
Three sets of real records show how different the two patterns can be. None of them comes from a dairy. Two come from sugarcane, where the work runs against a harvest deadline, as farm work does, and one from the service staff of an oil company.
| Where the records came from | How many absences per worker | How long each one lasted |
|---|---|---|
| 1,230 doctor’s notes from 400 sugarcane workers in western São Paulo, Brazil, over nine months | about three per worker | three quarters of them for a single day |
| 9,443 absence records from a sugar agribusiness with 3,150 workers in Peru, over two years | about three per worker | 4.24 days on average; the longest 6.64 percent (ten days or more) averaged 27.3 days |
| 2,564 sick-leave episodes from 782 service workers at an oil company in Brazil, over three years | 3.3 per worker; 69.3 percent of workers had at least one | 6.6 days on average; the five leaves for illness caused by work added up to 837 days |
The sugarcane study, published in 2014 by Aline Ceccato and colleagues in Cadernos de Saúde Pública, a Brazilian public health journal, is the clearest case of the short pattern. Seventy-five percent of the doctor’s notes were for one day. No single one of those absences would have stood out, and together they were three quarters of the notes.
The Peruvian study, published in 2018 by Wilfor Aguirre Quispe and a colleague in Horizonte Médico, a Peruvian medical journal, shows what a few long absences do inside a record that is mostly short. The authors call an absence prolonged at ten days or more. The average absence lasted 4.24 days, and only 6.64 percent were prolonged, with a mean of 27.3 days.
The oil company study sorted its leaves by cause. Of the 2,546 leaves with complete data, 2,447 (96.11 percent) were classed by the company doctors as non-occupational illness with leave of up to 15 days. Five workers had leaves for illness caused by work, and those leaves added up to 837 working days, about 167 days each (837 divided by 5, our arithmetic).
Count absences and days lost in two separate columns, so the few long ones do not hide behind the many short ones.
Why does the same rate need different fixes?
Many short absences usually point to the schedule and to how people feel about the work. A 2019 study by Tesfaye Mekonnen and colleagues in BMC Research Notes, a medical research journal, interviewed 444 flower farm workers in Ethiopia. Having worked while sick was the factor most tied to later sick absence, at about three times the odds, and job dissatisfaction came with about 1.6 times the odds.
The authors give a reason: working while sick “may exacerbate workers’ health conditions resulting in subsequent repeated away off work because of lack of the necessary recuperation”. That points to a crew sized so that one person can be missing, and to days off written on the calendar before they are taken by surprise.
A few long absences usually point to the job. In the Peruvian study, the factors most tied to prolonged absence were work accidents, moderate to intense physical activity with load handling, and atypical work schedules. How much a job asks a worker to lift, how shifts rotate and what leads to an accident are decisions someone made about the work.
The sugarcane authors add that other studies tied the high incidence of absence for muscle and joint problems to payment by production, with heavy workloads and no pauses, which growers call piece rates. They say their own findings agree.
The two patterns also connect. In the oil company records, most leaves for illness caused by work had been preceded by leave for the same or an equivalent diagnosis, filed as non-occupational. For one worker, the work cause was recognized only after the fifteenth episode with the same or a similar diagnosis.
So short absences point to the schedule only while they are spread across different workers and different complaints. When one worker keeps coming back with the same ache, a long absence may be starting, and a relief milker no longer answers it.
To see which case you have, count how many different workers were away. In the oil company records, 69.3 percent of the workers had at least one leave. Of the 72 workers with more than 50 days of absence, 27 had ten leaves or more. Of the 109 with three days or less, 79 had a single leave.
The authors sum it up: the more leave episodes, the longer the absence. On two crews of 30 workers, three workers away ten times each and thirty workers away once each both give 30 absences. In the first crew the problem is three workers, and in the second it is spread across the whole crew.
Sort the doctor’s notes by complaint and by worker, and mark every worker who comes back with the same complaint.
What does a lost hour cost at harvest?
The usual instinct is to value a lost hour at the hourly wage, and on a farm that undercounts it. A 2017 study of Canadian firms, not farms, by Zhang and colleagues in Health Economics Review found that the productivity loss due to worker absence exceeds the wage for team workers, especially in small firms. Team workers are people whose work depends on the rest of a crew.
The authors say why the small operation is exposed: large firms can hire extra employees to hold output when a team worker is absent, and small firms may not be able to afford this expense.
They also cite an economic model, not a measurement on farms, with two conditions common on farms. One is crew work that depends on skills hard to replace. There, the value of lost output to the firm from an absence will exceed the daily wage of the absent worker. The loss, the model says, could be as large as the total output of the team. The other is a deadline, because the cost of an absence exceeds the wage when the firm pays a penalty for missing an output target.
Farms rarely pay that kind of penalty, but harvest works the same way: work not done while the weather allows costs part of the crop. In both conditions, the loss shrinks if replacements are found who are either inexpensive or are close substitutes for the absent worker.
The same lost hour is worth little in a quiet month and more in the third week of harvest, and absences do not fall evenly. In the sugarcane record, doctor’s notes for muscle and joint problems were more frequent at the end of the harvest than in the off-season.
Count the jobs on your place where nobody else could step in. That count tells you more than the hourly wage about whether a day off costs one day’s pay or part of the harvest.
What goes on the absence sheet first, and how long does it take?
Set aside an hour with the time sheets from the period that just ended, and the doctor’s notes if you keep them. Without notes, the time sheets alone have how often and how long. Enter the numbers in the calculator and copy its results to the top of the sheet: the rate, absences per worker, the average length and how many workers were away.
The six lines below are what the calculator cannot work out. Absences go on the sheet by week and not by name, because the sheet is for fixing how the work is organized and is not a file on anyone. Names go only next to the jobs nobody else can cover.
A rate that falls is not always good news. If it falls because workers are afraid to stay home sick, the cost is only postponed, and the Ethiopian result on working while sick suggests it can come back as more sick days. There is no target rate to aim at, so compare with your own previous period.
Once the sheet is filled in, the next question is what turnover costs, which has its own article and calculator.
If you have a farm org chart, write next to each job who can stand in for it. If you do not have one yet, the list of one-person jobs on this sheet is a start. These articles are among the articles on managing people, and setting this period beside the last is a habit the rest of farm management runs on too.