.png)
Employee attrition is usually discussed as an HR metric: how many employees left, why they left, and whether turnover is increasing.
For an actuary or CFO, however, attrition has another dimension.
It can directly influence the measurement of gratuity and other long-term employee benefit liabilities.
An actuarial valuation does not normally assume that every employee currently on the payroll will remain with the organisation until retirement. Instead, it considers the probability that employees may resign, retire early, transfer or otherwise leave employment before reaching their expected retirement date.
That probability is reflected through an employee attrition or turnover assumption.
Under Ind AS 19, employee turnover is specifically identified as a demographic actuarial assumption. Actuarial assumptions are required to represent an entity's best estimates of the variables that will ultimately determine the cost of employee benefits and should be unbiased and mutually compatible. (Ministry of Corporate Affairs)
For CFOs, therefore, the attrition assumption is not simply a technical percentage appearing at the back of an actuarial report. It can influence the reported liability, employee benefit expense, actuarial gains and losses, cash-flow expectations and year-on-year movement in employee benefit obligations.
Understanding how it works is essential.
An attrition assumption represents the expected rate at which employees are likely to leave an organisation before retirement or another expected benefit-payment event.
For example, an actuarial valuation may assume that a certain percentage of employees within a particular age or service group will leave the organisation each year.
This does not mean that the actuary is predicting exactly which employees will resign.
Instead, it is a probability-based assumption applied to the employee population.
Consider 1,000 employees in a workforce.
If the assumed annual attrition rate for a particular employee group is 10%, the actuarial model does not simply remove exactly 100 named employees from the valuation.
Rather, it reflects the probability of employees leaving at different points in the future and calculates expected benefit payments accordingly.
This probability becomes part of the projected cash flows used to determine the employee benefit obligation.
Ind AS 19 recognises employee turnover, mortality, disability and early retirement among the demographic assumptions that can affect the ultimate cost of defined benefit plans. (Ministry of Corporate Affairs)
Employee benefit obligations often extend many years into the future.
A 30-year-old employee may potentially remain with an organisation for another 25 or 30 years. But assuming with certainty that the employee will continue until retirement may not reflect realistic workforce behaviour.
Some employees will resign.
Some may leave within a few years.
Others may remain until retirement.
An actuarial valuation therefore incorporates probabilities to estimate the amount and timing of future benefit payments.
For defined benefit plans such as gratuity, Ind AS 19 requires the use of the Projected Unit Credit Method. Under this approach, an entity estimates the ultimate cost of benefits earned through current and past service and considers demographic assumptions such as employee turnover together with financial assumptions such as future salaries. (Ministry of Corporate Affairs)
Attrition matters because leaving employment can change:
the amount of future service an employee completes;
the salary at which a benefit may eventually be calculated;
whether particular vesting or eligibility conditions are satisfied;
the timing of the benefit payment; and
the probability of different benefit outcomes.
The actuarial model therefore needs a reasonable estimate of future employee turnover rather than assuming either zero attrition or immediate departure.
One of the most common questions is:
Should the attrition assumption simply equal the company's historical employee turnover rate?
Not necessarily.
Historical experience is an important starting point, but an actuarial assumption is intended to represent future expectations.
Suppose an organisation experienced the following turnover pattern:
PeriodObserved AttritionYear 112%Year 214%Year 313%Year 418%Year 516%
A simple five-year average might provide useful information.
But automatically adopting that average as the future actuarial assumption may overlook changes occurring in the business.
For example, the organisation may recently have introduced stronger retention programmes, significantly revised compensation, closed a business division, shifted geographically, changed its workforce composition or experienced an unusual period of restructuring.
Historical turnover tells management what happened.
The actuarial assumption needs to represent management's best estimate of what is expected to happen in the future, supported by reasonable evidence.
Under Ind AS 19, actuarial assumptions are best estimates rather than deliberately conservative or optimistic inputs. The standard requires them to be unbiased. (Ministry of Corporate Affairs)
This is why a strong assumption-setting process combines historical data with an understanding of current and expected workforce conditions.
A workforce is rarely homogeneous.
Employee turnover can vary considerably depending on:
age,
length of service,
job level,
business unit,
geography,
employment category,
industry segment, and
career stage.
Consider an organisation with two major employee populations.
Younger employees in operational or technology roles may historically show significantly higher turnover.
Senior employees with 15 or 20 years of service may have much lower turnover.
If the organisation simply calculates an overall company attrition rate of 12% and applies it uniformly to every employee, the assumption may fail to represent either group accurately.
For this reason, actuarial valuations may use graded attrition assumptions.
A simplified example could look like this:
Employee ProfileIllustrative Attrition AssumptionYounger / early-service employeesHigherMid-career employeesModerateLong-service / senior employeesLowerEmployees approaching retirementBased on relevant experience
These are only illustrative patterns. The actual rates should come from the organisation's experience and reasonable expectations rather than from a generic table.
Age and service can be particularly important when analysing attrition.
An employee who joined six months ago may have a different probability of leaving than someone who has already completed fifteen years with the organisation.
This is sometimes described as selective attrition.
Employees may experience higher turnover during the first few years of employment and progressively lower turnover as tenure increases.
Alternatively, particular industries may experience different patterns.
This creates an important actuarial issue.
Two organisations could both report an overall annual attrition rate of 10%, but their employee benefit liabilities may respond very differently if their workforce compositions are different.
Most employees are relatively young and have short service.
Most employees have long service and are approaching retirement.
The same 10% assumption does not necessarily have the same significance for both organisations.
The impact depends on where employees sit in their benefit journey, what benefits they have already earned and what happens to those benefits if they leave.
The relationship between attrition and gratuity liability needs to be understood carefully.
It is tempting to say:
Higher attrition = lower gratuity liability.
That can often be directionally true in certain circumstances, but it is not a universal rule.
The actual effect depends on the plan terms, employee service, benefit eligibility, salary, age and expected timing of payment.
For an employee early in service, an increased probability of leaving before satisfying an applicable minimum-service condition may reduce the expected gratuity obligation.
For an employee who has already satisfied the relevant service condition, leaving employment may result in gratuity becoming payable earlier.
That changes not only the benefit amount but also the timing of the cash flow and the period over which it is discounted.
Similarly, an employee expected to remain until retirement may receive a benefit based on several additional years of service and future salary increases.
If the employee is instead expected to leave earlier, those future service years and salary increases may not be reflected in the same way.
The liability impact is therefore driven by the complete benefit formula rather than the attrition rate in isolation.
Assume two employees both earn the same current salary.
Employee A has completed three years of service.
Employee B has completed fifteen years of service.
If the assumed probability of resignation increases, the actuarial impact on Employee A can be very different from the impact on Employee B.
Employee A may still be affected by minimum-service or vesting considerations applicable to the benefit.
Employee B already has substantial accumulated service, and resignation may trigger a material benefit payment.
The same change in attrition assumption is therefore being applied to two very different benefit positions.
This is why sensitivity to employee turnover cannot be understood merely by looking at a single company-wide percentage.
Attrition also matters for long-term leave benefits, but its impact may differ from gratuity.
Suppose employees accumulate earned leave over several years.
What happens when an employee resigns?
That depends entirely on the leave rules.
If accumulated leave is encashable on resignation, employee departure can trigger payment of the accumulated balance.
If certain accumulated leave is forfeited on resignation, higher employee turnover may reduce the expected amount ultimately payable.
If employees usually utilise leave before departure, the expected cost can follow another pattern.
Consequently, attrition interacts with:
leave accumulation,
leave utilisation,
vesting conditions,
encashment rules,
maximum accumulation limits,
salary escalation, and
the timing of settlement.
A long-term leave valuation therefore needs more than a resignation percentage. It needs an understanding of what resignation means economically under the leave policy.
This principle extends beyond gratuity and leave encashment.
Other long-term or post-employment benefits may include long-service awards, jubilee benefits, post-retirement medical benefits and other service-linked arrangements.
Suppose an organisation pays a long-service award only after an employee completes 20 years.
An employee with 18 years of service is much closer to qualifying than an employee who has completed two years.
The assumed probability of remaining with the organisation therefore materially affects the expected obligation.
For a post-retirement medical benefit, employee turnover may influence the probability that an employee will ultimately become eligible for post-retirement coverage.
Thus, the economic consequences of employee turnover depend on the conditions attached to each benefit.
This is why the same attrition assumption should not be mechanically interpreted in the same manner across every employee benefit plan.
Assume an organisation previously used an attrition assumption of 8%.
After reviewing employee experience, management and its actuary conclude that a rate of 12% better represents expected future turnover.
The actuarial model is then rerun using the updated assumption.
If higher expected turnover reduces the probability of employees earning larger future benefits, the defined benefit obligation may decrease.
If the assumption is reduced instead—for example, because employee retention has improved—the expected period of employee service may increase.
That can increase projected benefits for some plans.
However, again, the direction and size of the change depend on plan design and employee demographics.
For employees already entitled to a benefit on resignation, increased turnover may accelerate benefit payments. Earlier payments receive less discounting than payments expected many years later, which can offset some of the effect of reduced future salary and service.
Consequently, CFOs should be cautious about treating attrition sensitivity as a simple inverse relationship.
There are two distinct ways employee turnover can affect an actuarial valuation.
Management revises its expectation of future employee turnover.
For example:
Previous assumption: 8%
New assumption: 10%
The resulting change in the obligation reflects a change in the demographic assumption.
Actual employee behaviour differs from what the previous valuation expected.
For example, the actuarial model assumed that 100 employees were likely to leave over a particular period, but substantially more or fewer employees actually left.
Under Ind AS 19, actuarial gains and losses can arise from changes in demographic assumptions and from experience adjustments when actual outcomes differ from previous assumptions. (Ministry of Corporate Affairs)
This distinction is useful because it explains why the liability moved, rather than merely showing that it changed.
A valuation assumption should not continue indefinitely simply because it has been used in previous years.
Suppose an organisation has assumed 10% attrition for five consecutive valuations.
Actual experience might have been:
YearAssumedActualYear 110%9%Year 210%11%Year 310%17%Year 410%18%Year 510%16%
One year of higher turnover may simply be volatility.
Three years of materially higher turnover, however, should prompt investigation.
Management should ask why actual experience has changed.
Was there a structural change in the workforce?
Has the industry's labour market changed?
Did salary competitiveness deteriorate?
Was there restructuring?
Did the company expand into locations with different employee behaviour?
Did the employee mix change significantly?
The answer determines whether the deviation is temporary or whether the future assumption should be reconsidered.
Experience analysis is important, but assumption setting should not become mechanical.
Imagine attrition was historically 8%, but during one year it increased to 22% because the organisation closed a major division.
Automatically changing the long-term attrition assumption to 22% may not make sense if the restructuring was a one-off event.
Conversely, dismissing a sustained increase as temporary year after year may also be inappropriate.
The objective is to identify a reasonable expectation of future employee behaviour.
This requires judgement.
An actuarial assumption should therefore consider both:
credible historical experience, and
known factors affecting future experience.
A particularly important improvement for large organisations is employee segmentation.
Suppose a company operates both a corporate head office and a large customer-service operation.
The customer-service workforce may have materially higher employee turnover.
The senior corporate population may have significantly lower turnover.
Applying the same 15% attrition assumption to both groups may distort the expected benefit cash flows.
A more representative approach might consider different assumptions for groups that demonstrate materially different behaviour.
Possible segmentation may include employee age, completed service, grade, location, business unit or employment category.
However, segmentation should have a genuine basis.
Creating dozens of employee groups from limited data does not automatically make a valuation more accurate.
The data should be sufficiently credible to support the distinctions being made.
An attrition assumption is only as reliable as the data used to analyse employee turnover.
Common problems include terminated employees remaining marked as active, employee transfer records being treated as resignations, employees moving between group entities being counted as external exits, incorrect joining dates and incomplete historical employee files.
These errors can produce a misleading turnover percentage.
For example, suppose 500 employees appear to have exited during the year.
Further analysis reveals that 150 merely transferred between group companies.
If all 500 are classified as employee attrition, management may significantly overestimate external turnover.
Before historical attrition is used for actuarial purposes, the underlying employee movement data should therefore be reconciled and cleaned.
Rather than simply checking whether an attrition percentage has changed from last year, CFOs should understand the logic behind it.
The more useful questions are whether actual turnover has been compared with previous assumptions, whether turnover differs materially by age or service, whether unusual events influenced historical data, whether the assumption reflects current workforce expectations and whether the selected rate appropriately interacts with the benefit rules.
This shifts the conversation from:
“Is 10% the correct number?”
to:
“Why is 10% a reasonable representation of expected future employee behaviour for this workforce?”
That is a much stronger governance question.
For a defined benefit gratuity plan, a change in the employee turnover assumption can contribute to an actuarial gain or loss forming part of the remeasurement of the net defined benefit liability or asset.
Ind AS 19 requires remeasurements of defined benefit plans to be recognised in Other Comprehensive Income (OCI). (Ministry of Corporate Affairs)
For other long-term employee benefits, however, the accounting differs.
Ind AS 19 requires service cost, net interest and remeasurements relating to other long-term employee benefits to be recognised in profit or loss, except where another Ind AS requires or permits inclusion in the cost of an asset. (Ministry of Corporate Affairs)
This distinction is particularly important when employee attrition affects both gratuity and long-term leave encashment.
The underlying workforce assumption may be related, but the financial statement impact of changes in that assumption may not appear in the same place.
An attrition assumption should not be changed merely to achieve a preferred valuation result.
Equally, it should not remain unchanged simply for consistency.
A reassessment becomes particularly important when there is sustained deviation between actual and assumed turnover, major organisational restructuring, significant changes in compensation or retention strategy, changes in employee demographics, rapid expansion or contraction, acquisitions or integration of a new workforce, shifts in industry employment conditions or clear evidence that different employee groups display different turnover patterns.
The purpose of the review is not necessarily to change the assumption every year.
The purpose is to determine whether the existing assumption still represents the best estimate available at the reporting date.
Actuarial assumptions should be stable enough to reflect long-term expectations rather than short-term noise.
But consistency should not become rigidity.
Using exactly the same attrition assumption for ten years despite significant changes in workforce behaviour can be just as problematic as changing the assumption dramatically every year.
A good assumption-setting process therefore combines:
historical experience,
current workforce characteristics,
future expectations,
benefit design,
credible management information, and
actuarial judgement.
Ind AS 19's principle that assumptions should be unbiased and mutually compatible provides the appropriate framework: assumptions should reflect the economics of the obligation rather than produce a predetermined accounting outcome. (Ministry of Corporate Affairs)
Consider a company with a rapidly growing technology workforce.
Several years ago, most employees were experienced professionals with relatively long tenure. The company therefore used a comparatively low employee turnover assumption.
Following expansion, the workforce changed substantially.
A large proportion of employees are now younger, recently hired professionals operating in a highly competitive employment market.
Actual turnover has also increased for three consecutive years.
Continuing to apply the old attrition assumption simply because it was used historically may no longer adequately reflect the workforce.
Management and the actuary may instead analyse turnover by age and service bands.
The analysis could show that early-service employees experience significantly higher turnover while long-service employees remain comparatively stable.
A graded assumption may therefore better represent expected future behaviour than one organisation-wide percentage.
The important point is that the assumption changes because the underlying employee population and credible expectations changed - not because management wanted the liability to increase or decrease.
No actuarial assumption operates completely in isolation.
Attrition interacts with salary escalation, retirement age, discount rate and benefit eligibility conditions.
For example, lower expected attrition could mean employees are expected to remain longer.
Remaining longer may mean additional service.
Additional service may lead to higher benefits.
Future salary increases may further raise those benefits.
At the same time, payments may occur further into the future, increasing the period over which they are discounted.
The final effect is the combined result of these interactions.
This is why simply stating that “attrition decreased from 10% to 8%, therefore liability increased by X” can sometimes be an oversimplification unless the actuarial calculation isolates the assumption change.
Employee attrition is often owned operationally by HR, while actuarial valuation sits with finance.
That separation can create gaps.
HR may have detailed information on resignations, employee tenure and retention trends but may not know how those patterns affect employee benefit liabilities.
Finance may review the actuarial report without having access to the workforce insights held by HR.
A stronger process brings the two together.
HR provides reliable employee movement data and explains emerging workforce trends.
Finance understands the reporting implications.
The actuary translates those expectations into an appropriate demographic assumption and measures their financial effect.
That combination produces a more defensible valuation than choosing an attrition percentage in isolation.
Employee attrition is far more than an HR statistic when an organisation has gratuity or other long-term employee benefit obligations.
It affects the probability that employees will remain in service, the benefits they may ultimately earn, the timing of future payments and, ultimately, the value reported in the financial statements.
But there is no universally appropriate attrition percentage.
The right assumption depends on the organisation's actual employee experience, workforce profile, benefit rules and reasonable expectations about future employee behaviour.
A higher attrition assumption does not automatically reduce every liability, just as a lower assumption does not automatically increase every liability by a predictable amount. The impact needs to be understood within the specific benefit design and employee population.
For CFOs, the key is therefore not merely to ask whether the attrition assumption changed.
The more important questions are:
Does the assumption reflect actual workforce behaviour? Has experience been compared with previous expectations? Are different employee groups behaving differently? And does the current assumption remain a reasonable best estimate for the future?
Regular experience analysis, reliable employee data and thoughtful segmentation can make gratuity and long-term employee benefit valuations more robust, explainable and useful for financial planning.
At KA Pandit, actuarial valuation can go beyond calculating the year-end liability. Understanding the reasons behind assumption changes - including employee attrition - helps organisations interpret movements in their employee benefit obligations, strengthen financial reporting and make better-informed long-term decisions.
This article is based on the principles of Ind AS 19 – Employee Benefits, particularly its requirements relating to the Projected Unit Credit Method, demographic assumptions, employee turnover, actuarial gains and losses, and the recognition of defined benefit and other long-term employee benefit obligations. (Ministry of Corporate Affairs)
A particularly important message for this topic is to avoid writing that higher attrition always means lower gratuity liability. The effect depends on service, vesting/eligibility, plan terms and payment timing, so the blog deliberately keeps that distinction clear.
Get in touch with our experts for personalized solutions.
