Post-employment benefits are defined as “employee benefits (other than termination benefits and short-term employee benefits) that are payable after the completion of employment”. Post-employment benefit plans are defined as “formal or informal arrangements under which an entity provides post-employment benefits for one or more employees”.
Post-employment benefit plans are classified either as defined contribution plans or defined benefit plans, depending on the economic substance of the plan.
Defined contribution plans are “post-employment benefit plans under which an entity pays fixed contributions into a separate entity (a fund) and will have no legal or constructive obligation to pay further contributions if the fund does not hold sufficient assets to pay all employee benefits relating to employee service in the current and prior periods”.
The entity’s legal or constructive obligation is limited to the contributions it has agreed to pay, together with investment returns arising from contributions. Normally, the rate of contribution to be paid by the employer will be specified in the plan’s rules. The employee takes both the actuarial risk (that benefits will be less than expected) and the investment risk (if the investments have performed badly, the benefit will be lower).
In June 2019, the IFRS IC issued an agenda decision addressing the classification of an employee benefit plan and the effect of a potential discount.
Distinguishing defined contribution and defined benefit plans: Effect of a potential discount on plan classification The IFRS IC, issued an agenda decision in June 2019. The Committee concluded that the existence of a right to a potential discount on fixed future contributions would not in itself preclude classification as defined contribution plan.
The agenda decision also states clearly the importance of considering all terms and conditions of a plan (including any informal practices that might give rise to a constructive obligation) to determine the classification of a plan.
The manner and frequency with which a potential discount is calculated should be considered to determine whether actuarial and investment risk is transferred to the entity.
A plan is a defined contribution plan only when the employer’s legal or constructive obligation is limited to the fixed amount that it contributes to the fund.
It is a defined benefit plan if the entity is obliged to provide agreed benefits to the employees. In our view, the existence of a discount will, in many cases, suggest there is a defined benefit arrangement.
Defined benefit plans are the residual category. They are defined as “post-employment benefit plans other than defined contribution plans”.
The entity’s obligation under a defined benefit plan is to provide agreed benefits to current and former employees. The actuarial risk and/or investment risk falls on the entity, at least to some extent. The entity’s obligations might increase if actuarial or investment experience is worse than expected.
Distinguishing defined contribution plans and defined benefit plans: voluntary enhancements to state pension benefits The national pensions authority administers a state pension plan in the country. Mandatory contributions from employers fund the plan on a payas-you-go basis.
Pensions legislation determines the minimum benefits paid to each employee when they retire, and the national pensions authority sets annual employer contributions at the level required to meet expected pension obligations in the same period.
Employers have no obligation to the state pension plan beyond their annual contributions, but the legislation permits employers to voluntarily enhance the state pension benefits paid to their former employees.
Entity A pays annual contributions to the state pension plan on behalf of all of its employees. Entity A has elected to enhance the benefits received by certain classes of employees:
- Entity A makes annual payments to an insurance entity equal to 5% of the salary of each employee based overseas. The insurance policy pays out when the employee retires, and the employee should use the proceeds to purchase additional post-retirement benefits. Entity A has no obligation to the employee or the insurance company beyond the annual payment.
- The employment contracts of management-grade employees state that their pension will be at least 1% of final salary for each year of service. Entity A will pay to each retired employee each year an amount equal to the difference between the guaranteed benefit and the state pension. Entity A has no obligation if the state pension is greater than the guaranteed benefit.
The arrangement’s substance determines the accounting treatment applied to the different components of the pension plan.
Employers fund the basic state pension on a pay-as-you-go basis, and entity A has no obligation to pay benefits beyond its annual contribution. The state takes actuarial and investment risk. Therefore, this is a defined contribution plan.
Employees based overseas receive additional benefits, based on the proceeds of an insurance policy.
Entity A has no obligation beyond paying the annual insurance premium. The insurance company and the employee bear the investment and actuarial risk. This is a defined contribution arrangement.
Management-grade employees receive a pension equivalent to 1% of their final salary for each year of service. Entity A has a contractual obligation to make a top-up payment to former employees when the state pension is less than the guaranteed minimum pension.
Entity A has the investment risk and the actuarial risk. This is a defined benefit arrangement, and the entity should recognise a pension liability for the additional benefit (that is, the amount of any shortfall where the guaranteed benefit is greater than the state pension).
Distinguishing defined contribution plans and defined benefit plans: pension plan with guaranteed interest An entity provides a pension plan with the following characteristics:
- The employees contribute 3% of their salaries to the plan.
- Entity A’s contribution matches the employees’ contribution.
- The retirement age is 65.
- The employee’s account balance consists of accumulated contributions, credited with interest guaranteed by the employer at 4% per year.
The entity has the investment risk that the return from assets invested will be less than the guaranteed return. The entity’s obligation is not limited to the amount that it agrees to contribute.
Although the entity might anticipate that the fund’s investment returns will exceed the guaranteed rate of 4%, there is a risk that investment returns will fall short and the employer will be required to make additional contributions. This is, therefore, a defined benefit plan.
Classification of plans containing features of both defined benefit plan and defined contribution plan Question:
Some plans contain features of both defined contribution and defined benefit plans. How should such plans be classified?
Solution:
IAS 19 requires that such plans are categorised as either defined contribution or defined benefit plans. IAS 19 is clear that all plans fall into one category or the other, and they default to defined benefit if they fail to meet the defined contribution definition. So, if there are any defined benefit features in a plan, the plan is a defined benefit plan.
For example, a plan might provide money purchase benefits, but with a guaranteed level of benefit based on a proportion of final salary. If investments perform satisfactorily, the employer will have no obligation in excess of the contributions that it has agreed to make.
The employer bears downside investment risk and might be required to make further contributions in order to fulfil the final salary guarantee. The criteria for classifying as a defined contribution plan are not met, so this plan is classified as a defined benefit plan.
Defined benefit plans might be wholly or partly funded by contributions by an entity, and sometimes its employees, into an entity, or fund, that is legally separate from the reporting entity or through a qualifying insurance policy.
Unfunded post-employment benefit plans are plans where no plan assets are set aside in advance to provide for future liabilities; instead pension liabilities are met out of the employer’s own resources as they fall due.
Accounting for defined contribution plans is straightforward. The contribution payable for the service rendered by an employee during a period is recognised as an expense (unless included in the cost of an asset, as required or permitted by another IFRS).
Except for outstanding or pre-paid contributions, an entity has no assets or liabilities in respect of a defined contribution plan.
Effect of vesting conditions on accounting for defined contribution plans An employer might contribute, or be obliged to contribute, an amount to an employee’s defined contribution plans that vest over a certain period subsequent to the contribution.
If the employee leaves within that period prior to vesting, the employer is entitled to a refund of the contribution, or its then investment value.
In July 2011, the IFRS IC issued an agenda decision that clarifies the effect of vesting conditions on the accounting for such defined contribution arrangements set out in IAS 19.
The IFRS IC concluded that each contribution to a defined contribution plan is recognised as an expense or recognised as a liability (accrued expense) over the period of service that obliges the employer to make the contribution to the plan.
This means that the expense would be recognised once the contribution is made (or the obligation to make the contribution arises), rather than over the subsequent vesting period.
In its conclusion, the IFRS IC made a distinction between the period that creates the employer’s obligation to pay the defined contribution and the period of service that entitles an employee to receive the benefit from the defined contribution (that is, the vesting period).
Refunds are recognised as an asset and income when the entity becomes entitled to them (that is, by the employee failing to meet the vesting condition). The IFRS IC’s agenda decision is only applicable in circumstances where employer contributions to a defined contribution plan have vesting conditions.
Contributions that are payable more than 12 months after the end of the period to which they relate should be discounted using the rate specified.
Accounting for defined benefit plans involves the following steps, which are applied separately to each material plan:
The net defined benefit liability (asset) is “the deficit or surplus adjusted for any effect of limiting a net defined benefit asset to the asset ceiling”. The deficit or surplus is defined as:
“(a) the present value of the defined benefit obligation less
(b)The fair value of plan assets (if any)”.
A surplus is regarded as an asset to the extent that the employer can gain an economic benefit from it. The employer does not have to own a surplus in order to recognise an asset. It is sufficient that the employer has access to future economic benefits that it controls via, for example, the ability to reduce future employer contributions. A surplus should be recognised as an asset to the extent that the employer is able to recover it through reduced future contributions or refunds, either directly to the employer or indirectly to another plan in deficit.
Plan assets are defined as:
Assets held by a long-term employee benefit fund, that is, assets, other than non-transferable financial instruments issued by the reporting entity, that:
Qualifying insurance policies. These are policies issued by an insurer that is not a related party (as defined in IAS 24) of the reporting entity, where the proceeds of the policy:
Utilisation of plan assets Question:
Can the employer utilise the proceeds from plan assets towards meeting expenses other than employee benefits?
Solution:
In the case of assets held by a long-term employee benefit fund or a qualifying insurance policy, the assets should be held solely for the purpose of paying or funding employee benefits and cannot be used by the employer for any other purpose (unless they represent surplus assets), including settlement of other liabilities on the employer’s liquidation.
This is important, because plans in some countries might contain clauses that give other creditors access to plan assets.
In such circumstances, the assets are not plan assets for the purposes of IAS 19. Assets that are not plan assets are accounted for under other relevant IFRSs (For example, IFRS 9 if they are financial instruments).
Assets held by the entity to fulfil future pension obligations Entity A operates a defined benefit pension plan. The plan is unfunded, but the entity holds a number of investments to fulfil its future pension obligations, which are treated as available-for-sale financial assets as defined in IFRS 9.
Although the investments are identified as relating to the defined benefit obligation, the fact that they are not held by a separate legal entity and not protected from other creditors means that they are not plan assets as defined by IAS 19.
Investments that are not held by a separate legal entity could still be treated as plan assets if they are qualifying insurance policies.
Trust assets available to creditors on winding up Entity B provides retirement benefits to its senior executives through a trust. The trust agreement’s legal form is such that the assets held to cover the pension liabilities are available to the general creditors of the entity on winding up. The assets held by the trust are not plan assets as defined by IAS 19.
Pension plan controlled by the entity Entity A established entity B as its subsidiary and is its only shareholder. The articles of incorporation of entity B state that entity B is a pension plan and that its only purpose is to collect and invest pension contributions made by entity A and to pay pensions to former employees of entity A under specified defined benefit pension plan terms.
All members of entity B’s board of directors are nominated by entity A. There are no laws that would prohibit the controlling shareholder of the entity from changing the articles of incorporation or the purpose of existence of the subsidiary pension entity.
A change in the articles of incorporation could allow the pension assets to revert to entity A prior to satisfying the benefit payments to employees.
Entity B does not satisfy the criteria for pension plan assets in IAS 19, because entity A can change the articles of incorporation and the purpose of entity B.
Entity A can thus require that the assets are returned to it before all pension benefit payments to the current and former employees are satisfied.
Entity A should therefore consolidate entity B. The assets held by entity B will be accounted for as corporate assets in entity A’s consolidated balance sheet, and the pension obligation will be recognised as a separate liability.
A qualifying insurance policy does not necessarily have to be an insurance contract under IFRS 4 (or IFRS 17). An investment contract that does not include significant insurance risk, and is therefore not an insurance contract under IFRS 4 (or IFRS 17), could still be a qualifying insurance policy.
Plan assets do not include:
Non-transferable financial instruments issued by the reporting entity Non-transferable financial instruments issued by the reporting entity include:
- loans and shares subject to restrictive conditions;
- unlisted corporate bonds that are redeemable but not transferable without the reporting entity’s permission;
- insurance policies which cannot be sold; and
- loans to the reporting entity that cannot be assigned to a third party.
The principle underlying the definition of plan assets requires that the assets are moved beyond the reporting entity’s control so that they are only available to meet employee benefit obligations.
Non-transferable financial instruments issued by the reporting entity would be available to meet benefit obligations only if the reporting entity were to repay or redeem them. Without this restriction, an entity could reduce the benefit liability by issuing non-transferable equity instruments to a plan.
Similarly, promised employer contributions are excluded from plan assets, because they are not creditor protected until paid. Non-transferable debt issued by the employer to a plan is the same as a promise to pay future contributions.
Unpaid contributions due from the reporting entity Amounts due from the reporting entity in respect of unpaid contributions are not plan assets because, until paid, they are not creditor protected. However, plan assets do include other assets and liabilities that do not relate directly to the payment of benefits – For example, rental on property, plant and equipment used by the plan, professional fees incurred by the plan but not yet paid, and liabilities (or assets) arising from a derivative financial instrument or other futures contract.
Parent company guarantee of subsidiary contributions It is common for a group entity to guarantee to third parties the liabilities of fellow group members, associates and joint ventures. An example is a parent company issuing a contingent guarantee in respect of its subsidiary’s defined benefit pension liability.
Such a guarantee reduces the risk that the subsidiary will not make payment when due (For example, due to cash flow difficulties). In our view, the parent issuing the contingent guarantee should account for it as an insurance contract in accordance with IFRS 4 (or IFRS 17 if applicable) (that is, record a liability).
The subsidiary benefiting from the guarantee should disclose it as a contingent asset of the pension plan.
It is not a plan asset, since the plan’s right to receive payment under a guarantee is no different from its right to receive contributions from the employer, and unpaid contributions are not plan assets.
The plan deficit should not be reduced, because a subsidiary is able to secure the support of its parent in making contributions.
An entity might look to another party, such as an insurance company, to settle all or part of a defined benefit obligation. Plan assets include qualifying insurance policies. However, IAS 19 also contains guidance in respect of insurance policies that do not satisfy the definition of qualifying insurance policies.
Where it is virtually certain that another party will reimburse some or all of the expenditure required to settle a defined benefit obligation, even though this does not represent a qualifying insurance policy or a plan asset, it is termed as a reimbursement right. It is important to identify reimbursement rights, because they are not accounted for in the same way as other assets of the entity.
Reimbursement rights are recognised as separate assets, rather than as a deduction, in determining the net defined benefit liability. In all other respects, reimbursement rights and the income that they yield are treated in the same way as other plan assets. The asset representing the reimbursement right is measured at the same amount as the liability that it reimburses.
Treatment of difference between cost of purchasing an insurance policy and the present value of defined benefit obligation Question:
How should the difference between the cost of purchasing an insurance policy and the deemed fair value equal to the present value of the defined benefit obligation that is matched by the policy be treated?
Solution:
The cost of purchasing an insurance policy will typically be larger than the present value of the defined benefit obligation that it is intended to reimburse.
Any difference between the cost of purchasing an insurance policy and the present value of the defined benefit obligation to which it relates should generally be viewed and treated as an actuarial loss.
This is independent of whether the insurance policy is purchased by the pension fund or the entity itself, provided that the policy itself is an asset of the plan or meets the definition of a qualifying insurance policy or reimbursement right.
IAS 19 stipulates that a reimbursement right is treated in the same way as plan assets, except for its recognition as a separate asset rather than deducting it to determine the defined benefit liability.
A different treatment will apply where an insurance policy is purchased to settle a benefit obligation.
Pension obligation covered by an insurance policy Entity A operates a defined benefit pension plan for its senior management. It has decided to cover its pension obligation with an insurance policy taken out with a leading insurance entity that is not a related party.
The policy requires the insurer to reimburse entity A in full for all benefit payments. The policy is a qualifying insurance policy, because its proceeds can only be used to fund employee benefits and are not available to entity A for any other purpose.
Pension obligation covered by an insurance policy where the insurer is a related entity Entity A operates a defined benefit pension plan for its senior management. It has decided to cover its pension obligation with an insurance policy taken out with a leading insurance entity, which is a related party of entity A.
The policy requires the insurer to reimburse entity A in full for all benefit payments.
The definition of plan assets in IAS 19 provides that, where an insurance policy is issued to a reporting entity by a related party of the reporting entity (as defined in IAS 24), it cannot be treated as a plan asset.
Accordingly, in this example the insurance policy is not a plan asset. It will, however, represent a reimbursement right.
The definition of plan assets in IAS 19 has a hierarchical structure, and so it is only necessary to consider when an insurance policy is a qualifying insurance policy where it is not held by a separate employee benefit fund.
Policies held by a separate employee benefit fund could be treated as plan assets, regardless of whether they are ‘qualifying’ insurance policies.
Insurance policies issued by insurers for their employee benefit arrangements Question:
How should insurance companies account for insurance policies issued for their employee benefit arrangements?
Solution:
In practice, some insurance companies have set up separate legal entities, such as trusts, for their employee benefit arrangements. Insurance policies issued by the employer are then held by these entities. Such policies that are ‘non-transferable financial instruments issued by the reporting entity’ do not meet the definition of plan assets under IAS 19. This is regardless of whether all of the other conditions for classification as plan assets are met.
IFRS 4 provides clarity in this area by concluding that, where an insurance contract has been issued by an insurer to a defined benefit plan covering the employees of the issuer, or of another entity consolidated within the same financial statements as the issuer, such contracts should generally not be treated as plan assets.
On the other hand, policies held by a long-term employee benefit fund that are freely transferable, or issued by a related party that is not consolidated, could still meet the definition of plan assets under IAS 19.
Insurance policy with a right to terminate the policy Entity B operates a defined benefit pension plan, and it has decided to cover its pension obligation with an insurance policy taken out with a leading insurance entity.
The policy requires the insurer to reimburse entity B in full for all benefit payments. It also permits entity B to cancel the arrangements, in which case the insurer repays a pre-determined amount to entity B.
The existence of the contractual right to terminate the policy and receive the proceeds means that entity B can apply the proceeds of the policy in its business, regardless of the benefit obligation.
The use of the proceeds is not restricted to paying employee benefits. Hence, this policy is not a qualifying insurance policy, but it does represent a reimbursement right as defined in IAS 19.
Plan assets should be measured at their fair value at the balance sheet date. IFRS 13 provides guidance on fair value.
Plan assets might include qualifying insurance policies or insurance policies that represent reimbursement rights. These give rise to specific valuation issues. If the insurance policies exactly match the amount and timing of some or all of the benefits payable under a plan, the fair value of those insurance policies is deemed to be the present value of the related obligations.
This is subject to any reduction required if the amounts receivable under the insurance policies are not recoverable in full. The determination of fair value for an insurance policy that does not exactly match the amount and timing of some or all of the benefits payable under a plan is made following the guidance in IFRS 13.
The defined benefit obligation reflects “expected future payments required to settle the obligation resulting from employee service in the current and prior periods”. It comprises not only legal obligations under the formal terms of the plan, but also constructive obligations arising from an employer’s informal practices. A constructive obligation might arise where a change in an entity’s informal practices would cause unacceptable damage to its relationship with employees.
Valid expectation to make additional payments to pension plan Entity A has used its active employees’ medical plan to meet the medical costs of retired employees in each of the last 10 years.
Entity A has met all pensioners’ medical expenses and has not made any communications to indicate that this is a limited action on the employer’s part, without any commitment to continue doing so.
The entity has no legal obligation to meet its pensioners’ medical costs; however, its actions have created an expectation that it will continue to meet these costs.
The entity has, therefore, created a constructive obligation, and should account for the costs of providing medical cover as a defined benefit plan.
Established practice of granting annual increases to pensions An entity has a practice of granting annual increases, to pensions in payment and deferred pensions, over and above any increases that might be required by law or the pension trust deed. Such increases are granted as a measure of protection against inflation.
The entity has an established practice and has created a valid expectation that it will continue with this practice in the future, even though it has no legal or contractual obligation to grant the discretionary increases.
It has no realistic alternative but to pay the increases in order to avoid damaging employee relations. The entity has a constructive obligation to continue to grant annual increases, so their cost should be factored into the measurement of the defined benefit obligation.
Occasional discretionary increases to pensions in payment An entity has, historically, granted occasional discretionary increases to pensions in payment. These increases have been made on an ad hoc basis and have no established pattern. Management considers that the entity has not created an expectation of similar increases in the future.
Therefore, the entity has not established a constructive obligation, so the cost of any increases will result in additional past service liabilities when they are granted. If the employer’s practice changes, such that a valid expectation of future increases is created, the cost of future increases should be factored into the measurement of the defined benefit obligation.
An example where such practice could create a constructive obligation would be where the increases were granted following actuarial valuations that disclosed a funding surplus and not granted when the actuarial valuation disclosed a deficit. This could create an expectation regarding the use of surplus, in line with IAS 19.
The formal terms of a defined benefit plan might permit an entity to terminate its obligation under the plan. It is usually difficult for an entity to cancel a plan if the entity wishes to retain its employees. It should be assumed that an entity that is currently promising postretirement benefits will continue to do so over the remaining working lives of employees, in the absence of evidence to the contrary.
Employee service gives rise to an obligation under a defined benefit plan, even if the benefits are conditional on future employment (not vested). Employee service before the vesting date gives rise to a constructive obligation, because the amount of future service that an employee will have to render at each successive balance sheet date before becoming entitled to the benefit is reduced.
The probability that some employees might not satisfy vesting conditions is taken into account in the measurement of the defined benefit obligation. However, this probability does not determine whether the obligation exists.
Similarly, although certain benefits (for example, post-employment medical benefits) become payable only if a specified event occurs after retirement, an obligation is created when the employee renders services that give entitlement to those benefits.
Benefit should be attributed to periods of service according to a plan’s benefit formula, unless an employee’s service in later years will lead to a materially higher level of benefit than in earlier years.
Benefits are attributed on a straight-line basis for plans in which service in later years leads to a materially higher level of benefit. Benefit is attributed to the current period in order to determine current service cost. Benefit is attributed to the current and prior periods in order to determine the present value of the defined benefit obligation.
Annual pension based on final salary A plan provides an annual pension of 1/60th of final salary for each year of service. The pension is payable from the age of 65. The benefit attributed to each year of service is equal to the present value, at the expected retirement date, of an annual pension of 1/60th of the estimated final salary, payable from the expected retirement date until the expected date of death.
Employee service gives rise to an obligation under a defined benefit plan, even if the benefits are not vested. However, when attributing benefit to periods of service, the probability that some employees might not satisfy vesting conditions (which includes performance hurdles) before becoming entitled to benefits should be taken into account. The probability of employees departing before vesting impacts measurement, and not the existence of the liability.
Annual pension that does not vest until employee completes 10 years of service A plan provides a lump sum benefit of C1,000, payable on retirement, for each year of service. An expense of C1,000 is attributed to each year (ignoring discounting). The benefit does not vest until an employee completes 10 years of service.
The measurement of the defined benefit obligation in each of the first 10 years should reflect the probability that some employees might not complete 10 years of service.
Effect of pension promises based on performance hurdles Entities might specify certain performance targets with respect to pension promises. Performance targets might relate to various forms of pension promise, ranging from additional pensionable earnings from performance bonuses to more complex arrangements relating to additional sponsor contributions or years of deemed service.
In 2008, the IFRS IC issued an agenda decision to clarify the impact of pension promises which are based on such performance hurdles.
The IFRS IC concluded that the requirements of IAS 19 were clear and it did not expect any divergence in practice. It noted that old IAS 19 (para 76 in the current version of IAS 19) states that “Actuarial assumptions are an entity’s best estimates of the variables that will determine the ultimate cost of providing post-employment benefits”.
Performance targets are variables that will affect the ultimate cost of providing the post-employment benefits. They should therefore be included in the determination of the benefit.
The IFRS IC also noted that paragraph 67 of old IAS 19 (para 70 in the current version of IAS 19) requires benefits to be attributed to periods of service according to the benefit formula, unless an employee’s service in later years will lead to a materially higher level of benefit than in earlier years. Where benefits are affected by performance hurdles, the effect on the attribution of benefits should also be considered.
A plan’s benefit formula might set a point beyond which no further benefit can be earned. The benefit earned is attributed over the period that it is earned, and not the entire period of the employee’s service.
Single lump sum on retirement for first 10 years of service A plan pays a single lump sum of C1,000 after 10 years of service. A benefit of C100 would be attributed to each of the first 10 years of employee service (taking into account the fact that some employees might not complete 10 years’ service, and before the impact of discounting); but no benefit would be attributed to service after 10 years.
A plan’s benefit formula might result in service in later years giving rise to greater benefit than earlier years. IAS 19 requires that the benefit in these circumstances is attributed on a straight-line basis from:
Benefit increases after 20 years of service A plan provides an annual pension of 1/60th of final salary, but this increases to 1/40th of final salary after more than 20 years’ service.
For those employees expected to remain with the employer for at least 20 years, benefit equal to the present value, at the expected retirement date, of an annual pension of 1/40th of the estimated final salary, payable from the expected retirement date until the expected date of death, is attributed to each year of service.
For all other employees, benefit equal to the present value, at the expected retirement date, of an annual pension of 1/60th of the estimated final salary, payable from the expected retirement date until the expected date of death, is attributed to each year of service.
In all cases, the measurement of the defined benefit obligation will reflect the probability that an employee might leave or die before the normal retirement date.
Benefit increases slightly in later years A post-employment medical plan reimburses 40% of an employee’s postemployment medical costs if the employee leaves after more than 10 and less than 20 years of service, and it reimburses 50% of those costs if the employee leaves after 20 or more years of service.
Under the plan’s benefit formula, the entity attributes 4% of the present value of the expected medical costs (40% ÷ 10) to each of the first 10 years and 1% (50% − 40% ÷ 10) to each of the second 10 years. For employees expected to leave within 10 years, no benefit is attributed.
Material increases in benefits in later years Entity A operates a pension plan that pays a pension of:
- C100 for each of the first three years of service;
- C500 for each of the years of service from years 4 to 6; and
- C2,400 for each of the years of service from years 7 to 9.
So, after nine years of service, an employee will be entitled to a pension of C9,000 (C100 × 3 + C500 × 3 + C2,400 × 3). Further service after nine years does not give rise to additional pension.
Entity A’s employees’ services in later years will lead to a materially higher level of benefit than in earlier years.
Entity A should, therefore, attribute benefit to years of service on a straight-line basis, from the date when service by the employee first leads to benefits under the plan until the date when further service by the employee will lead to no material number of further benefits under the plan, other than from further salary increases.
Entity A should, therefore, attribute C1,000 to each year of service (C9,000 ÷ 9 years, ignoring discounting) for employees expected to complete nine years of service, rather than following the attribution formula in the plan.
Future salary increases will affect the amount required to settle the obligation that exists for service before the balance sheet date, but do not create an additional obligation for benefits that are expressed as a constant proportion of final salary for each year of service. Salary increases do not lead to further benefits, even though the amount of benefit is dependent on final salary.
The extent to which actual final salary at the date of retirement is different from that estimated in the calculation of the defined benefit obligation is reflected as a remeasurement, because it derives from a change to an actuarial estimate.
Effect of expected salary increases on allocation of benefits to period of service The basis of conclusions to IFRIC D9, ‘Employee benefit plans with a promised return on contributions or notional contributions’, explains that expected increases in salary should be taken into account in determining whether a benefit formula expressed in terms of current salary (such as a career average plan) allocates a materially higher level of benefit to later years of service.
This means that, if salary increases do not lead to a materially higher benefit in later years, benefits are allocated to periods of service according to the benefit formula; but, if salary increases are significant, a straight-line allocation should be made.
IFRIC D9 was never finalised as an interpretation and was withdrawn in November 2006. This was because, at that time, the issue was included in the scope of the Board project to revise IAS 19.
However, the specific issue of salary increases was not addressed in the final amendments to IAS 19 issued in June 2011. Our view is that including future salary increases in determining whether a benefit formula allocates a materially higher level of benefit to later years is appropriate.
The first step in accounting for a defined benefit plan is to determine the cost of the benefit that the employees have earned in return for their service, in the current and prior periods, using an actuarial technique. The entity should determine how much benefit is attributable to current and prior periods, and it should make estimates about demographic variables and financial variables that will influence the cost of the benefit. These estimates are broadly described as actuarial assumptions.
Measurement of the defined benefit obligation usually involves the expertise of a qualified actuary, although this is not required by the standard.
Actuarial assumptions include both demographic and financial assumptions, and they reflect an entity’s best estimate of the variables that will determine the ultimate cost of providing post-employment benefits. Some defined benefit pension promises are based on achieving specific performance targets. Such targets are also variables that will affect the ultimate cost of providing the benefit and should be included in the determination of the benefit.
Demographic assumptions concern the future characteristics of the members of the plan, such as:
Financial assumptions deal with matters such as:
Financial assumptions should be based on market expectations at the balance sheet date for the period over which the obligations are to be settled.
Financial assumptions should be determined in nominal terms, unless estimates in real (inflation-adjusted) terms are more reliable (for example, in a hyper-inflationary economy, or where benefits are index linked and there is a deep market in index-linked bonds of the same currency and term).
Actuarial assumptions should be both unbiased and mutually compatible. Unbiased means neither imprudent nor excessively conservative. Mutually compatible assumptions should reflect the economic relationships between factors such as inflation, rates of salary increase and discount rates.
Inflation and discount rate The measurement of long-term employee benefits typically requires an assumption regarding future inflation and its impact on salary and benefit levels. IAS 19 requires actuarial assumptions to be mutually compatible. This implies that salary/benefit increases and discount rates should reflect compatible assumptions about inflation.
However, the Interpretations Committee noted in an agenda decision (in 2017) that it is not possible to assess whether, and to what extent, a discount rate derived by applying the requirements in IAS 19 is compatible with other actuarial assumptions.
Accordingly, the entity applies the requirements in IAS 19 when it determines the discount rate.
Where a common currency is used by different countries, the discount rates used in those countries should be based on the same underlying market data (subject to variation due to differences in the duration of the liabilities).
However, the inflation assumption might differ, particularly over the short to medium term, to reflect country-specific factors.
The discount rate reflects the time value of money, based on the expected timing of the benefit payments. The discount rate does not reflect investment risk or actuarial risk, since other actuarial assumptions deal with these items.
The discount rate does not reflect the specific risk associated with the entity’s business, nor does it reflect the risk that future experiences might differ from actuarial assumptions.
The discount rate is determined by reference to market yields, at the appropriate balance sheet date, on high-quality corporate bonds of equivalent currency and term to the benefit obligations.
Determining high-quality corporate bonds Question:
At what rating would a bond be considered as ‘high-quality’?
Solution:
IAS 19 requires employee benefit obligations to be discounted using the yield on high-quality corporate bonds where there is a deep market in such bonds. There is no explicit definition of this phrase in IAS 19, but most entities have determined ‘high-quality’ to mean bonds rated AA or better.
The IFRS Interpretations Committee (‘IC’) confirmed, in a November 2013 agenda decision, that the concept of ‘high-quality’ is an absolute not a relative measure, and that a reduction in the number of high-quality corporate bonds overall (or of particular durations) does not justify a change in an entity’s interpretation of what is high-quality.
In our view, consistent with the agenda decision, we would not expect management to change its interpretation of what is high-quality.
In some currencies, the rating agencies provide two classifications: a global rating, and a local rating. The purpose of the local rating is to provide a greater level of distinction between bonds reflecting risk relative to the local government debt.
In line with the IC rejection that concluded that ‘high-quality’ is an absolute measure and not a relative one, global ratings, rather than local ratings, should be used as the benchmark for ‘high-quality’.
Is a single discount rate used to measure the benefit liabilities? IAS 19 requires employment benefit liabilities to be measured using the yield on high-quality corporate bonds, for a currency where there is a deep market in such bonds, of the same duration.
The yield on government bonds of the same duration as the liability is used where there is no deep market in high-quality corporate bonds. The standard requires observable yields to be used.
Employee benefit obligations are often measured by discounting projected future benefit payments using a single discount rate based on weighted average market yields reflecting the timing and amount of the benefit payments. his approach is reasonable in an environment where yield curves are flat in the medium term and rates are relatively high.
It can also be used in market conditions where interest rates are low, or even negative, and yield curves have a more pronounced gradient, provided that the discount rate also reflects the spread of benefit payments around the average duration.
A single discount rate should be a weighted average and should not reflect the high point on the yield curve.
Lower rates and variable yields have led many entities to use more sophisticated approaches that better reflect the shape of the yield curve and the timing of benefit cash flows.
These approaches involve using different discount rates for different benefit payments (For example, discounting projected payments in each future year using the spot rate for that year from the yield curve). Such approaches are also appropriate.
Setting the discount rate A range of techniques could be used to derive a single discount rate or the yield curve used to measure employee benefit obligations. The observable data points used in constructing a yield curve or deriving a single discount rate might include negative yielding bonds.
The standard requires observable yields to be used, and so the yield on these bonds should be incorporated into the estimate of the discount rate and cannot be ignored.
Other long-term benefits, rather than post-employment benefits, typically have a shorter duration. It is possible that yields for the entire duration of such benefits could be negative.
Judgement should be applied consistently when assessing the discount rate under IAS 19. The starting point and process for determining the discount rate should be consistent from year to year, unless a change can be justified by a change in circumstances.
When looking at relatively short-term obligations, the materiality of any time value impact should be considered. Low interest rates (whether positive or negative) might mean that the impact of discounting is not material.
An entity determines if there is a deep market in high-quality corporate bonds in the currency in which the liabilities are denominated, rather than corporate bonds of a particular country. Government bonds denominated in that currency should be used where there is no deep market in high-quality corporate bonds in that currency.
Where a common currency is used by more than one country and there is no deep market in high-quality corporate bonds, the standard does not specify which government bond rate should be used.
The IFRS IC confirmed this in an agenda decision issued in June 2017, stating that an entity needs to consider high-quality corporate bonds denominated in the currency of the obligation. This is the case even if the high-quality corporate bonds are issued in local or other markets or countries.
Where there is no deep market in high-quality corporate bonds, market yields on government bonds denominated in that currency are used. The IFRS IC also states that judgement is required to determine the appropriate populations for both high-quality corporate bonds and government bonds to use when determining the discount rate.
The currency and the terms of the bonds must be consistent with those of the post-employment benefit obligations.
Can a synthetic high-quality corporate bond rate constructed with reference to a bond market in another country be used to determine the discount rate? Question:
In a currency where there is no deep market in high-quality corporate bonds, can an entity use corporate bond yields in another currency and currency swap rates to synthetically create bond yields?
Solution:
In June 2005, the IFRS IC observed that, where there was not a deep market in high-quality corporate bonds, the standard requires government bond rates to be used.
A synthetically constructed equivalent to a high-quality corporate bond, by reference to currency swap rates and the bond market in another currency which has a deep market in high-quality corporate bonds, cannot be used to determine the discount rate.
Change in basis or reference for determining the discount rate A conclusion as to whether a deep market exists will be made on a case-by case basis for each currency. Corporate and government bond markets develop over time.
As a result, a deep corporate bond market might develop where one did not previously exist, or the reverse might occur, and such a development might result in a change in the basis or reference for determining the discount rate to be applied.
Where a change in the discount rate occurs because of the development of a deep corporate bond market in a currency that previously did not have one, this change would be accounted for prospectively.
There might be no deep market in bonds with a sufficiently long maturity to match the estimated maturity of all of the benefit payments. In such cases, the discount rate for longer maturities should be estimated by extrapolating current market rates of shorter-term payments along the yield curve.
Determining corporate bond yields There might be wide spreads of yields on AA rated bonds, especially during times of global market liquidity problems. This is likely to lead to both greater volatility and lower defined benefit liabilities, as a result of the higher discount rate. Therefore, it is important to consider the following:
- Whether a robust methodology has been used to determine the discount rate, and whether this methodology has changed since the previous reporting date.
- What allowance has been made for plan-specific factors, such as age profile?
For example, if the yield on an index has been used, how has the yield been adjusted to reflect differences between the term of bonds comprising the index and the term of the plan’s liabilities?
- If the yield on an index has been used, what consideration has been given to whether the constituents of the index are appropriate?
A change in an index, or the exclusion of some constituents from an index to reflect current circumstances, is a change in accounting estimate and should be applied prospectively.
The accounting policy to apply a high-quality corporate bond yield of appropriate duration has not changed, merely the approach considered appropriate to determine that yield.
Management should consider the impact of changes in the constituents of indices to ensure that such changes are appropriately taken into account at the balance sheet date.
For example, if a bond that is a constituent of a particular index is downgraded on 15 December 20X1, but the index is not published until 5 January 20X2, reporting entities should reflect the downgrade as an adjustment in the index yield used to determine liabilities at 31 December 20X1, were material.
Estimates of future salary increases take account of, for example, inflation, seniority, promotion and supply and demand in the employment market. Assumptions about future salary and benefit levels should reflect an entity’s legal and constructive obligations.
The measurement of the obligation should include any formal requirement or constructive obligation to increase benefits in future periods. An entity might have an obligation to increase benefits where:
Estimates of future cost-of-living increases to be reflected in measurement Pension plan X has a long-established practice of providing cost-of-living increases to pensions in payment in line with the movement in a consumer price index (CPI). However, these are only awarded to the extent that the investment returns on plan assets are above a specific rate.
There is no catch-up in future years for subsequent higher returns where an increase has been restricted by the rate of return in a specific period.
Assumed future pension increases should reflect both expectations for future changes in the CPI and expected returns on plan assets, together with the variability in those returns (see IAS 19 (Conditional indexation of IAS 19)).
Some defined benefit plans limit the contributions that an entity is required to pay. The ultimate cost of the benefits takes account of the effect of a limit on contributions. The effect of a limit on contributions is determined over the shorter of: the estimated life of the entity; and the estimated life of the plan.
A defined benefit plan might require employees or third parties to contribute to the cost of the plan. Contributions by employees reduce the cost of the benefits to the entity. An entity considers whether third-party contributions reduce the cost of the benefits to the entity or are a reimbursement right.
Contributions by employees or third parties are either set out in the formal terms of the plan, arise from a constructive obligation that goes beyond those terms, or are discretionary. Discretionary contributions by employees or third parties reduce service cost on payment of those contributions to the plan.
Contributions from employees or third parties set out in the formal terms of the plan either reduce the service cost (if linked to service), or reduce re-measurements of the net defined benefit liability (asset) (for example, if required to reduce a deficit arising from losses on plan assets or actuarial losses). Contributions from employees or third parties in respect of service are attributed to periods of service as a negative benefit.
Measurement of deficit where there is an expectation that employee contribution rates will be increased Some pension plans include provisions that specify employee contributions not as a fixed amount but as a proportion of the total contribution rate (For example, a 60:40 split between employer and employees).
These arrangements typically call for surpluses to be shared between the employer and its employees, and for contributions for both employer and employees to be increased proportionally where there is a deficit.
The employer’s share of any surplus could be used to reduce employer contributions, whilst the employee share could be used either to reduce contributions or to enhance benefits.
Whilst the plan terms might specify a shared cost approach, legally and practically it is often not possible to force employees to pay increased contributions, because legislation often gives employees a right to opt out of employer-sponsored pension plans. If they consider the required contribution level to be onerous, employees could elect to do so.
For such plans, the issue is how to measure the deficit where there is an expectation that employee contribution rates will need to be increased to help meet the deficit.
In order to determine whether a liability for the full deficit or only a part of the deficit should be recognised, an entity should consider whether or not there is an asset (that is, a reimbursement right) or some other way to avoid the obligation.
The amendments made in 2013 address the treatment of employee contributions, depending on whether those contributions are linked to service or not, and whether they vary with the length of service or not.
Employee contributions that are linked to service and do not vary with the length of service reduce the current period service cost. Employee contributions that are not linked to service impact the measurement of the defined benefit obligation.
Although IAS 19 says “An example of contributions that are not linked to service is when the contributions are required to reduce a deficit arising from losses on plan assets or from actuarial losses”, this does not make it clear how the determination of whether or not contributions are linked to service is made.
In our view, where contributions are expressed as a percentage of salary, they are linked to service, because salary will only be paid where there is service. If members will be required to contribute towards a deficit, whether they are still in service or not, such contributions will not be linked to service.
If the employer can force employees to fund their share of the deficit, either through payment of additional contributions or a reduction in benefits, it would seem clear that the employer cannot be obliged to meet the employees’ share of the deficit.
Accordingly, it might be appropriate to recognise only its portion of the deficit (in the above example, 60%). This is consistent with IAS 19, which specifies that the defined benefit obligation should reflect any limit on the employer’s share of the cost of the future benefit.
However, where the employer cannot enforce a benefit reduction or force employees to pay future contributions at the pre-specified rate, because the employees could choose to opt out of the pension plan, the plan does not require employees to make contributions to reduce or eliminate an existing deficit.
This suggests that the entity would generally record the full deficit. In substance, the plan includes a provision that makes it likely that employees will help to fund the deficit, but future expected employee contributions do not meet the definition of an asset (Conceptual framework) and would not be recognisable as a reimbursement right under IAS 19.
An alternative view, which could also be considered, is to regard the rate of future expected employee contributions as an actuarial assumption which might lead to recording less than the full deficit.
Where this is the case, this assumption should be consistently applied, correlated logically with the plan’s assumption about future service, and appropriately disclosed.
The accounting for contributions from employees or third parties depends on whether they are linked to service. If they are linked, there is a further distinction based on whether the amount of contributions varies depending on the number of years in service:
Contribution is linked to service and varies with the length of service A plan provides a lump sum benefit on retirement of 10% of final salary for the first 10 years of service, plus 20% of final salary for each subsequent year of service, and it requires employee contributions equal to 5% of salary for the first 10 years of service and 8% thereafter. In this plan, contributions are linked to the length of service.
The contributions vary with the length of service, as well as salary, and so they have to be recognised over the working life. The benefit earned and the employee contributions would be recognised on a straight-line basis over the employee’s working life.
Contribution is linked to service but does not vary with the length of service A plan that requires employees to contribute 4% of salary if they are below age 40, and 7% of salary if they are 40 or above, is an example of a plan in which employee contributions are not linked to the length of service.
The contributions are linked to age and salary, but are not dependent on the length of service. So, the contributions can be recognised as a reduction of pension expense in the year in which the related service is delivered.
Contribution is not linked to service A post-employment medical insurance plan requires the employee to contribute an amount equal to the first C20 per month of the insurance premium. This is an arrangement in which contributions are not linked to service.
The expected future contributions from the employee would be payable after retirement. This would be included in the measurement of the defined benefit obligation.
Measurement of the defined benefit obligation should not reflect future changes that are not set out in the formal terms of the plan (or a constructive obligation). Such changes, when they occur, will result in:
Some post-employment benefits are linked to variables such as the level of state retirement benefits or state medical care. The measurement of such benefits reflects expected changes in such variables if, and only if, either:
Assumptions about medical costs consider the effect of technological advances, changes in health care utilisation and changes in health status of plan participants.
Medical costs assumptions should take into account the level and frequency of future claims and the cost of meeting those claims, which are affected by, for example, age, gender, geographical locations and health status. Some plans require employees to contribute towards the medical costs covered by the plan, and this should also be taken into account in the actuarial assumptions.
Changes in employee contributions will be reflected as a past service cost. The cost of meeting claims might also be reduced by benefits from the state or other medical providers and this, too, should be reflected in the actuarial assumptions.
The next step in the process of measuring the defined benefit obligation is to discount the estimated benefit to present value using an actuarial method. IAS 19 requires the whole amount of the obligation to be discounted to present value, even if part of the obligation falls due within 12 months of the balance sheet date.
An entity uses the projected unit credit method to determine the present value of its defined benefit obligation. The projected unit credit method views each period of service as giving rise to an additional unit of benefit entitlement, with each unit being measured separately, to build up the total obligation. The projected unit credit method is the only actuarial method permissible under IAS 19.
Assumptions used in applying projected unit credit method An entity has a statutory obligation to operate a termination indemnity scheme, under which employees who leave for any reason receive a lump sum based on one month’s salary for each year of service. Senior staff receive enhanced benefits. The benefits are not conditional on future employment, and they vest at the balance sheet date.
The entity has a legal obligation to pay the termination indemnity, calculated by reference to an employee’s salary and length of service. The employer has the actuarial risk, so this is a defined benefit arrangement.
The liability for termination indemnities should not reflect an assumption that all eligible staff resign at the balance sheet date; instead, the employer should calculate the liability using the projected unit method. The calculation of the liability will include a projection of the benefit earned to date to each future point that the benefit could be paid (For example, end of each year), with allowance for salary increases and probabilities of payment.
Each payment would then be discounted back to the valuation date, using the appropriate discount rate, based on the expected payment date of benefits.
A liability calculated without reference to future salary increases or the expected payment date would misstate the liability. The accrued liability would be overstated if the effect of discounting the payment exceeded the impact of future salary increases.
This was also clarified by the IFRS IC in April 2002, where it opined that the measurement of the liability for the vested benefits should reflect the expected date of employees leaving service and be discounted to a present value.
Applying projected unit credit method for a plan that pays a lump sum on termination of employment An entity operates a defined benefit plan that pays a lump sum on termination of service of 1% of final salary for each year of service. An employee joins the entity at the beginning of year 1 at a salary of C50,000 per year. Salaries are assumed to increase at 7% per annum and the discount rate is 8%. It is expected that the employee will retire after five years of service.
The following table shows how the defined benefit obligation builds up for this employee over the five years of service. It is assumed that there are no changes in actuarial assumptions, and the possibility that the employee might leave the entity before retiring is ignored.
Year 1 2 3 4 5 Estimated salary (7% growth) 50,000 53,500 57,245 61,252 65,540 Benefit attributed to each period 655 655 655 655 655 Cumulative benefit payable 655 1,311 1,966 2,622 3,277 Opening obligation 0 482 1,041 1,686 2,428 Interest (8% on opening obligation) 0 39 83 135 194 Current service cost (discounted at 8%) 482 520 562 607 655 Closing obligation 482 1,041 1,686 2,428 3,277 At the end of year 5, the employee’s final salary is C65,540 (C50,000*1.074). The cumulative benefit earned at the end of five years is 1% of the final salary times five years of service (C65,540*1%*5 = C3,277). The current year benefit earned in each of the five years is one-fifth of this (C3,277/5 = C655). The current service cost is the present value of the benefit attributed to the current year. So, in year 2, For example, the current service cost is C520 (C655/1.083).
Applying projected unit credit method for a plan that pays a lump sum on termination of employment with materially higher benefits in later periods An entity operates a defined benefit plan that pays a lump sum on termination of service of 1% of current salary for each year of service.
An employee joins the entity at the beginning of year 1 on a salary of C50,000. Salaries are assumed to increase at 30% per annum and the discount rate is 8%. It is expected that the employee will retire after five years of service.
Because the benefit is materially higher in later years due to the significant salary increases, the benefit attributable to each period is spread on a straight-line basis.
The following table shows how the defined benefit obligation builds up for this employee over the five years of service. It is assumed that there are no changes in actuarial assumptions, and the possibility that the employee might leave the entity before retiring is ignored.
Year 1 2 3 4 5 Estimated salary (30% growth) 50,000 65,000 84,500 109,850 142,805 Benefit earned in each period 500 650 845 1,098 1,428 Benefit attributed to each period 904 904 904 904 904 Cumulative benefit 4,521 Opening obligation 0 665 1,436 2,326 3,349 Interest (8% on opening obligation) 0 53 115 186 268 Current service cost (discounted at 8%) 665 718 775 837 904 Closing obligation 665 1,436 2,326 3,349 4,521 The current year benefit earned in each year is 1% of the current year salary, but the current year benefit attributed to each period is the cumulative benefit earned spread on a straight-line basis, C904 (C4,521/5). The current service cost is the present value of the benefit attributed to the current year. So, in year 2, For example, the current service cost is C718 (C904/1.083).
The measurement of the defined benefit obligation can be complicated in the case of plans with a promised return on contributions, sometimes known as ‘cash balance’ plans.
Two approaches have developed in practice: one values the liability on the fair value of the relevant assets; and the other uses an actuarial approach to determine the liability and discount it using the relevant discount rate. An entity should make a policy choice between these alternatives and apply it consistently.
Employee benefit plans with a promised return on contributions or notional contributions The measurement of the defined benefit obligation can be complicated in the case of plans with a promised return on contributions, sometimes known as cash balance’ plans. These include the following types of plans:
- Plans that promise a fixed return on actual or notional contributions (For example, a plan that provides a benefit equal to specified contributions plus a return of 4% a year over a specified future period).
- Plans that promise a return on actual or notional contributions based on specified assets or indices (For example, a plan that provides a benefit of an amount equal to specified notional contributions plus or minus the return on a notional holding of a basket of shares).
- Plans that combine both of the above features.
The IFRS IC published a draft interpretation D9, ‘Employee benefits with a promised return on contributions or notional contributions’, in July 2004.
The draft interpretation was subsequently withdrawn because the IASB included consideration of this topic in its IAS 19 project. The IASB published a discussion paper in March 2008.
However, the proposals that were in the discussion paper were not included in the revised standard. Whilst the IASB is not expected to address this topic in the near future, the IFRS IC has indicated that it believes that the Board did not intend for the accounting in this area to change as a result of the revision to IAS 19 in June 2011.
Accordingly, even though D9 has been withdrawn, it still contains useful guidance, and we believe that its concepts can be applied under revised IAS 19, pending further developments on the accounting for contribution-based promises.
Most respondents to D9 agreed that such plans are defined benefit plans. This is because the promise of a specified return (whether fixed or variable) means that the employer might have to make additional contributions, so the plans cannot be defined contribution plans (in which an entity has no legal or constructive obligation to pay further contributions relating to current or past service).
Plans that promise a fixed return on actual or notional contributions do not present any problems with regard to the accounting. The benefit to be paid in the future is estimated by projecting forward the contributions or notional contributions at the guaranteed fixed rate of return.
It is then attributed to periods of service and discounted to present value, in the same way as any other defined benefit plan. Plan assets are also treated in the same way as any other defined benefit plan.
The accounting for plans that promise a variable return on actual or notional contributions based on specified assets or indices is more complicated.
If the usual projected unit credit method for measuring the defined benefit obligation is applied, benefits are projected forward at the expected rate of return on the assets and discounted back to a present value.
Hence, if an entity established such a plan and immediately contributed C100, the defined benefit obligation on the same day would exceed C100 if the expected return on assets was greater than the discount rate.
This method will always give rise to a deficit where the projected returns expected to be earned on the reference assets exceed the discount rate. We believe that it would also be acceptable to apply the alternative approach outlined in D9.
The D9 alternative is that the plan liability is measured on the balance sheet date, using the fair value of the assets or notional assets on which the benefit depends. No projection is made for the benefits and there is no discounting.
If benefits are unvested, the measurement of the plan liability takes into account the probability that some employees might not satisfy the vesting conditions. The fair value of plan assets (if any) is recognised and measured according to the normal requirements of IAS 19.
An entity should make an accounting policy choice between the two alternatives outlined above. Even though an entity does not have to measure the defined benefit obligation in line with the fair value of the assets for a variable return promise, applying the D9 alternative will typically result in a more intuitive outcome.
If the D9 concepts are followed, the measurement of the plan liability follows the fair value of the assets on which the benefit is specified. The change in the plan liability should be split into an expected increase and a remeasurement. The increase in the liability is equal to the sum of:
- implied returns on assets;
- current service costs (that is, the contributions due for the period); and
- any re-measurements (For example, actuarial gains and losses from actual versus implied returns on assets, movement in the reference asset and differences from expected vesting to actual).
Liability increases arising from implied returns on assets and current service costs should be recognised in the income statement, whereas any re-measurements arising from actuarial gains and losses should be reflected in other comprehensive income.
Careful consideration is required as to whether the benefit is an ‘other long-term benefit’ or a ‘post-employment benefit’, because remeasurements are accounted for differently.
A plan with a promised return might be funded or unfunded, and benefits might be vested or unvested. Where there is a promised return, this might be based on plan assets or notional assets.
In some cases, the related invested assets might not meet the definition of plan assets. An entity needs to consider IFRS 9 for appropriate classification and measurement of such assets.
The accounting for associated asset costs might differ, depending on which standard the assets are accounted under. For plan assets, guidance for asset management costs is found in IAS 19. For non-plan assets, an entity follows the guidance in IFRS 9 for costs associated with financial non-plan assets.
Plans that provide a combination of a fixed guarantee and benefits that depend on future returns on assets are sometimes described as ‘defined contribution plans with a defined benefit underpin’.
Such schemes should be accounted for under the higher of the defined benefit obligation relating to the fixed guarantee and the obligation arising from the variable return. The latter is determined using the approach set out above for variable schemes.
An entity is required to determine the present value of the defined benefit obligation with sufficient regularity that the amounts recognised in the financial statements do not differ materially from the amounts that would be determined at the balance sheet date. Annual actuarial valuations are not required as at the balance sheet date.
Frequency of valuations The employer’s and the pension plan’s financial statements might have different accounting periods. In practice, it is likely that at least two full actuarial valuations will be required for funded pension plans: one for IAS 19 accounting, and another (on a funding basis) for the pension plan trustees. The funding valuation might use different, possibly more conservative, assumptions in relation to the liabilities. The IAS 19 valuation need not necessarily be done as at the employer’s balance sheet date.
In fact, a valuation as at an employer company’s year-end might not be available in time for the completion of the company’s financial statements. It could be done at an earlier date or the same date as the valuation required by the trustees, and it could then be updated, as necessary, to the employer’s year end.
For example, a company with a December year-end might have a pension plan with a March year end and obtain actuarial valuations as at 31 March. The full valuation would then have to be updated to each company year-end for any material transactions and other material changes in circumstances (including changes in market prices and interest rates).
An update is, in effect, an estimate of a full valuation. IAS 19 indicates that some aspects of the valuation should be updated at each balance sheet date of the reporting employer.
For example, the financial assumptions underpinning the valuation of the plan liabilities should be updated to reflect changes in market conditions. Thus, the discount rate should always be the current rate of return on an appropriate bond at the employer’s balance sheet date.
A change in the discount rate might also require other financial assumptions, such as the inflation assumption, to be updated. Other aspects of the valuation of the liabilities can be estimated from the previous full valuation by rolling the valuation forward and updating it for changes to the plan, such as benefit improvements.
Assumptions that are not directly affected by changes in market conditions need not be updated annually.
Individual circumstances will dictate whether a full valuation is required in between the previous full valuation, or whether an update is sufficient. If the latest full valuation was a long time ago and many changes have occurred since, the actuary might not be confident that an update will produce a reliable current estimate of the plan liabilities. In such circumstances, a full valuation might be appropriate.
Factors considered in deciding on frequency of valuations A UK-based multi-national group operates a large number of defined benefit pension plans in different countries. The pension plans include:
- A group pension plan, which provides generous benefits but is open only to UK employees. The pension cost recorded in respect of this plan is equivalent to 5% of the group’s total employee cost and 10% of net profit.
- A US pension plan, which provides benefits only to US-based employees. The group’s operations in the US are not significant, and investment returns have been stable for a number of years.
- A Latin America pension plan, which provides pension and medical insurance benefits to all employees in Latin America. The operations in Latin America are material to the group, and the stock markets in Brazil and Venezuela are extremely volatile. Venezuela is a hyperinflationary economy.
- An Africa pension plan, which provides limited benefits to expatriate staff in Africa. The plan is unfunded and covers a limited number of employees. The group announced a benefit enhancement in the current year.
Deciding on the frequency of valuations requires judgement. Factors to be taken into account include the size of the pension obligation, the volatility of the economic environment in which each plan operates, changes in plan benefits, and the overall impact of employee benefit costs on the financial statements. The same frequency does not necessarily apply to all plans.
It will also be necessary to determine the extent of the valuation. Sometimes, a full actuarial valuation will be necessary, but often it will be sufficient to update an existing valuation, as discussed above.
In this example, management concluded that the following valuation frequency was appropriate for each of its plans:
- The cost of the group pension plan is material to the financial statements. Although both the economic environment in the UK and the benefit package are stable, the income statement is sensitive to changes in the employee benefit cost. Accordingly, the present value of the defined benefit obligation and the fair value of plan assets should be determined each year.
- The US pension arrangements are not significant and the economic environment is stable. The fair value of quoted plan assets and updates to the discount rate, for example, are determined every year, but a full valuation of the defined benefit obligation is obtained only every three years.
- The Latin America pension plan is significant and it operates against a volatile economic background. The scale of the changes in the economic environment that occur every year might have a significant effect on the actuarial assumptions. Accordingly, the present value of the defined benefit obligation and the fair value of plan assets should be determined every year.
- The pension obligation in Africa is not significant and the liability is unfunded. Hence, there are no plan assets to be valued. The group usually determines the fair value of the defined benefit obligation every three years. However, in view of the announcement of the benefit enhancement, it might be necessary to obtain a valuation in the current year.
IAS 34 requires that an interim financial report should be prepared for a discrete period, with items of income and expenses recognised and measured on a basis consistent with that used in preparing the annual financial statements. The entity’s pension assets and liabilities should be measured in the same way as they would be at a year-end.
The net balance recognised in respect of a defined benefit pension plan could be an asset. The amount of asset that could be recognised is restricted. A pension plan surplus is regarded as an asset to the extent that the employer can obtain economic benefits. The amount recognised as an asset cannot exceed its recoverable amount, measured as the lower of:
Limit on amount recognised as an asset A defined benefit plan has the following characteristics at the balance sheet date:
C’000 Present value of defined benefit obligation (1,000) Fair value of plan assets 1,200 Surplus determined in accordance with IAS 19 200 Present value of available future refunds and reductions in future contributions 170 The amount that could be recognised as a net defined benefit asset in this example is C170,000, being the present value of available future refunds and reductions in future contributions.
The Interpretation Committee was asked to provide additional guidance on how to apply the asset ceiling in IAS 19. This is provided in IFRIC 14, which addresses:
IFRIC 14 is not relevant if a defined benefit plan is in deficit and there is no minimum funding requirement. An entity can recognise an asset if a plan is in surplus and the employer has an unconditional right to the surplus. .There is no need to consider the impact of any minimum funding requirement. IFRIC 14 might, however, lead to an increase in liabilities, even where a scheme is in deficit under IAS 19. This happens where contributions under a minimum funding requirement to reduce the existing deficit will create surplus that is not recoverable once they are made.
IFRIC 14 does not affect an entity’s ability to obtain a refund. This is determined by legislation and the plan rules. The interpretation provides guidance on how to account for any restrictions imposed by legislation or plan rules.
An economic benefit, in the form of a refund or a reduction in future contributions, is available if the entity can realise it at some point during the life of the plan or when the plan liabilities are settled. It is not necessary for a surplus to be immediately realisable at the balance sheet date for it to be recognised. A refund does not have to have been agreed for it to be available.
A refund is ‘available’ only if the entity has an unconditional right to the refund. There is no unconditional right, and an asset is not recognised if the refund of a surplus depends on the occurrence or non-occurrence of one or more uncertain future events that are not wholly within the entity’s control.
Assessing the existence of an unconditional right to a refund of surplus Interpreting whether an unconditional right to a refund of surplus exists can be difficult in practice. An entity does not have an unconditional right to a refund where the payment of the refund is subject to approval by another party (For example, a regulator or governing body of the plan).
However, if a regulator’s approval is required for a refund, but that approval is perfunctory (For example, for a closed fund with no more members), this could be treated as an unconditional right. This is due to the fact that the approval process is not substantive.
Plan rules should be considered in establishing whether there is an unconditional right. Legal advice might have to be obtained to determine whether an unconditional right exists.
A third party (For example, plan trustees) might have discretionary power but not an obligation to wind up a plan or grant benefit improvements on the occurrence of a particular event. The question arises as to whether the exercise of that power should be anticipated, or whether the effects should be recognised only when they occur. We believe that there are two acceptable views:
- The existence of an asset at the balance sheet date depends on whether the company has a right to obtain a refund or a reduction in future contributions. This right is not affected by future acts that could change the amount of the surplus that could ultimately be recovered. The fact that the trustees could choose to wind up the scheme or grant benefit improvements (and thus reduce the surplus) should not be anticipated and would not remove the company’s unconditional right to the surplus. This is based on IFRIC 14.
- The entity does not have an unconditional right to the surplus if the right to a refund resulting in economic benefit depends on actions by a third party, even if these actions are discretionary. If the trustees could force the plan to wind up before the last benefit is paid, or grant benefit improvements and thus effectively eliminate any surplus, the company does not have an unconditional right to the surplus. This is based on IFRIC 14.
The policy that a company adopts might represent a critical accounting judgement. This could require a disclosure in accordance with IAS 1.
The economic benefit available as a refund is the surplus at the balance sheet date that the entity has a right to receive as a refund, less any costs associated with payment of the refund. The refund should not be discounted.
A refund might not be available until after a plan is wound up, in which case the costs of winding up the plan are deducted from the asset.
The economic benefit available as a future reduction in contributions, absent any minimum funding requirement, is the future service cost to the entity for each period over the shorter of the plan’s expected life and the entity’s expected life.
The future service cost, excluding amounts that will be borne by employees, should be calculated using actuarial assumptions that are consistent with those used to determine the defined benefit obligation.
Statutory minimum funding requirements exist in many countries to improve the security of the post-employment benefit promise. Such requirements normally stipulate a minimum amount or level of contributions that should be made to a plan over a given period. Therefore, a minimum funding requirement might limit the ability of the entity to reduce future contributions.
IFRIC 14 does not change statutory or contractual funding rules, but it clarifies how entities should account for the effect of any such requirements.
Minimum funding requirements can result from a funding plan agreed by the entity and the plan’s trustees or a regulator, where such an agreement creates an obligation on the employer to pay the amounts specified. They do not have to be enforced by statute or by a regulator. Such agreements are often referred to as a ‘schedule of contributions’ where they specify the amount or rate of contributions that will be paid over a specified period.
A schedule of contributions that constitutes a minimum funding requirement is likely to be revised regularly by the plan’s actuary (typically, every three years). The contribution schedule cannot be viewed as having a three-year duration because it will be replaced by a new schedule after three years. It is extrapolated forward indefinitely on the basis that it will be replaced by an equivalent schedule at the end of the three years.
The limit on the measurement of a defined benefit asset might cause a minimum funding requirement to give rise to a liability. A requirement to make contributions to a plan would not normally affect the measurement of the defined benefit asset or liability. Contributions paid become plan assets, and so there is no additional liability.
However, a minimum funding requirement might give rise to a liability if the required contributions will not be available to the entity as a refund or a reduction in future contributions once they have been paid. Such an additional liability arises only if two conditions are satisfied:
Minimum funding contributions might be required to cover future service, as well as any existing shortfall relating to past service. The two sub-components are treated differently.
A pre-payment provides an economic benefit to the entity where it relieves the entity of an obligation to pay future minimum funding contributions that exceed future service cost. An entity should consider those economic benefits in measuring an asset. The economic benefit available as a reduction in future contributions, for minimum funding contributions relating to future service, is the sum of:
A liability should be recognised if an entity has an obligation to pay contributions to cover an existing shortfall in respect of services already received and some or all of those contributions will not be available after they are paid. The liability and its subsequent remeasurement should be recognised immediately by a charge to other comprehensive income.
Refunds where access to a surplus is limited Entity A participates in a multi-employer pension plan that provides benefits to employees based on final salary and years of service. Independent trustees administer the plan in accordance with relevant pension legislation. The trustees provide the information that each participating employer requires to apply defined benefit accounting.
The most recent actuarial valuation revealed a surplus. Pension legislation permits the trustees to refund surplus contributions to participating employers, subject to a legal maximum.
The pension legislation also requires that the plan maintains a solvency margin, determined as a percentage of the plan assets. The solvency margin is an insurance for contingencies that is ultimately available for refunds. The trade unions have challenged the trustees in connection with previous refunds, and the proposed refunds were subject to litigation.
The refunds that the participating employers finally received, following negotiations with trade unions, were significantly less than the legal maximum and were paid in instalments.
Entity A proposes to use its share of the surplus in the plan to calculate the asset limit.
Entity A should take account of the legal maximum, ignoring the solvency margin. Local legislation restricts the proportion of the surplus that the trustees can refund. Entity A, therefore, restricts the asset limit to the amount of the surplus that is available for refund.
Trade union action might restrict the refund and force the employer to provide additional benefits from the surplus. When the terms of the pension plan change and employees obtain additional benefits for their past service, the entity recognises the past service costs.
The IAS 19 asset ceiling is based on whether a refund is possible and not on whether a refund is probable.
IAS 19 surplus and minimum funding contributions fully refundable Entity A has a funding level on a minimum funding requirement (MFR) basis (which is measured on a different basis from that required under IAS 19) of 82% in plan B.
Under the minimum funding requirements, entity A is required to increase the funding level to 95% immediately and, as a result, it has a statutory obligation at the balance sheet date to contribute C200 to plan B immediately.
The plan rules permit a full refund of any surplus to entity A at the end of the life of the plan. The year-end valuations for plan B are set out below:
C Fair value of plan assets 1,200 Present value of defined benefit obligation (1,100) Surplus 100 The net defined benefit asset, before consideration of the MFR, is C100. Entity A should recognise a liability to the extent that the contributions payable is not fully available. Payment of the contributions of C200 will increase the IAS 19 surplus from C100 to C300.
Under the plan’s rules, this amount will be fully refundable to the entity with no associated costs. Therefore, no liability is recognised for the obligation to pay the contributions.
IAS 19 deficit and minimum funding contributions not fully available Entity C has a minimum funding requirement (MFR) basis of 77% in plan D. Under the MFR, entity C is required to increase the funding level to 100% immediately and, as a result, it has a statutory obligation at the balance sheet date to pay additional contributions of C300 to plan D.
The plan rules permit a maximum refund of 60% of the IAS 19 surplus to entity C, and entity C is not permitted to reduce its contributions below a specified level, which happens to equal the IAS 19 service cost. The yearend valuations for plan D are set out below:
C Fair value of plan assets 1,000 Present value of defined benefit obligation (1,100) Deficit 100 The net defined benefit liability, before consideration of the MFR, is C100.
The payment of C300 would change the IAS 19 deficit of C100 to a surplus of C200. Of this C200, 60% (C120) is refundable.
Therefore, of the contributions of C300, C100 eliminates the IAS 19 deficit, and C120 (60% of C200) is available as an economic benefit. The remaining C80 (40% of C200) of the contributions paid is not available to entity C.
The entity recognises a liability to the extent that the additional contributions payable is not available to it. Entity C should increase the net defined benefit liability by C80.
C80 is recognised immediately in other comprehensive income, and entity C recognises a net balance sheet liability of C180. No other liability is recognised in respect of the statutory obligation to pay contributions of C300. When the contributions of C300 are paid, the net balance sheet asset will be C120.
Any change in the effect of the asset ceiling, other than amounts included in net interest on the net defined benefit liability (asset), should be recognised through other comprehensive income.
These amounts are not recycled through profit or loss in a subsequent period. A settlement or curtailment loss that is recognised in the income statement, and leads to a reduction in an irrecoverable surplus within other comprehensive income, represents two different, although connected, transactions and does not represent recycling.
The amount of pension expense (income) to be recognised in profit or loss comprises the following individual components:
Accounting entries for recognising a net defined benefit asset The actuarial valuation of an entity’s pension plan at 31 December 20X3 showed a net surplus of C39 million. The profit and loss charge for 20X4 is C66 million (that is, the aggregate of current service cost, and net interest on the net defined benefit asset, which is not analysed here into different components).
The entity decided to reduce its employer contributions for 20X4 to C50 million. The surplus in the plan was measured at C23 million at 31 December 20X4. There are no minimum funding requirements and the surplus is fully refundable. There were no remeasurement gains or losses in the year.
The double-entry for the year 20X4 would be as follows:
C’m C’m Dr Profit and loss account (net pension cost) 66 Cr Defined benefit asset 66 Dr Defined benefit asset 50 Cr Cash (contributions paid) 50 The movements in the defined benefit asset are as follows:
C’m Defined benefit asset brought forward 39 Pension cost (profit and loss account) (66) Contributions paid 50 Defined benefit asset carried forward 23 The asset of C39 million at 31 December 20X3 is increased by contributions of C50 million during 20X4 and reduced by the pension cost of C66 million. At 31 December 20X4, an asset of C23 million is recognised, reflecting the ending net surplus of the plan.
Appropriate portions of employee benefit costs can be capitalised within the cost of assets, such as inventories or property, plant and equipment.
Current service cost is defined as “the increase in the present value of the defined benefit obligation resulting from employee service in the current period”.
It represents the actuarially calculated present value of the pension benefits earned by the active employees in each period, and it is supposed to reflect the economic cost for each period based on current market conditions. This cost is determined independently of the funding of the plan.
For a given set of employees and benefit formula, the current service cost should be the same, irrespective of whether the plan is in surplus, in deficit or unfunded.
Current service cost is not necessarily a stable percentage of pensionable pay year-on-year. For example, current service cost will vary if the discount rate changes. It will also increase year-on-year, as a proportion of pay, if the average age of the workforce is increasing. This is likely where a plan is closed to new entrants.
The valuation of the plan liabilities and, hence, the calculation of current service cost for each year should be based on the plan’s benefit formula. An exception is where a disproportionate amount of total benefits relates to later years of service, in which case the benefit should be allocated on a straight-line basis over the period in which it is earned.
Current service cost should be based on the most recent actuarial valuation that is available at the beginning of the year. The financial assumptions underlying the calculation of the present value of the benefits earned (that is, the rate used to discount liabilities, rate of inflation, rate of salary increase, and rates of pension and deferred pension increases) should be current as at the beginning of the year.
The current service cost for the year is based on the financial assumptions set at the beginning of the year. However, if a plan amendment, curtailment or settlement occurs during the reporting period and the net defined benefit liability (asset) is remeasured, the entity should determine current service cost for the remainder of the reporting period using the actuarial assumptions used to remeasure the net defined benefit liability (asset).
At the end of the year, the financial assumptions should always be updated for the purpose of remeasuring the plan liabilities. The adoption of new assumptions at the end of the year does not affect the current service cost for the past year, but it sets the assumptions underlying the current service cost for the next year.
Net interest on the net defined benefit liability (asset) is defined as “the change during the period in the net defined benefit liability (asset) that arises from the passage of time”.
The net interest cost can be viewed as comprising theoretical interest income on plan assets, interest cost on the defined benefit obligation (that is, representing the unwinding of the discount on the plan obligation) and interest on the effect of the asset ceiling.
Net interest on the net defined benefit liability (asset) is calculated by multiplying the net defined benefit liability (asset) by the discount rate. The entity should use the net defined benefit liability (asset) and the discount rate determined at the start of the annual reporting period.
However, if a plan amendment, curtailment or settlement occurs during the reporting period and the net defined benefit liability (asset) is remeasured, the entity determines net interest for the remainder of the reporting period using:
In addition, when determining net interest, the entity should take into account any changes in the net defined benefit liability (asset) during the period resulting from contributions or benefit payments.
Costs that would be expensed as services are provided Question:
What are some examples of costs that would be expensed as services are provided?
Solution:
These could include:
- payments to members of the trust’s management board who administer the pension payments in respect of that role;
- administration costs incurred to administer the pension plan participants’ database; and
- actuarial valuation costs.
Costs of administering the benefit plan should be expensed in the period in which the administration services are provided. However, expenses such as future medical costs (including claim handling costs, in the case of medical benefit) would be included in the actuarial assumptions used to measure the defined benefit obligation.
Indirect costs arising in relation to plan assets Question:
Can certain indirect costs arising in relation to plan assets be considered as costs of managing plan assets?
Solution:
Entities might view other indirect costs that arise in relation to plan assets as costs of managing plan assets. An example could be trustees’ time devoted to meetings discussing investment strategies. This would be acceptable if it is clear that such costs are incurred in managing plan assets.
However, we expect that, in most circumstances, entities will find that the effort necessary to analyse and track such costs sufficiently to support such an approach will not be cost beneficial.
Treatment of administration costs other than costs for of managing plan assets The following table summarises the acceptability of the potential methods of treating costs that do not qualify as costs of managing plan assets:
Method of recognising other costs in the income statement Acceptable Commentary Recognise in operating expenses (For example, in staff costs), but leave outside the pensions note reconciliation1 ✓ · Most aligned with the situation where plan expenses are paid directly by the entity. · In this case, even though not required, it would be desirable to have some disclosure in the pensions note of the amount and where it has been included.
· Amounts could also be included within the pensions note reconciliation, but this might be somewhat artificial, because expenses are not paid out of plan assets.
Recognise in operating expenses, and also show within the pensions note reconciliation as a separate line item ✓ · Most aligned with the situation where plan expenses are paid directly by the plan. · In this case, disclosure in the pensions note would be required.
· Amounts should be included within the pensions note reconciliation, because they are paid out of plan assets.
Recognise in operating expenses, but as part of current service costs within the pensions note reconciliation ✗ · Not acceptable, because expenses are not included within the definition of current service cost. · But amounts could be immaterial for amounts recognised in financial statements, even though recognised in operating results; so relegated to essentially a disclosure issue compared to other approaches above.
· More useful for users of financial statements to see plan expenses explicitly stated; so, a separate line item would be clearer.
Recognise in finance costs, but as part of net interest within the pensions note reconciliation ✗ · Not acceptable, because not within the definition of net interest under IAS 19. Recognise in finance costs as a separate line item, and within the pensions note reconciliation ✗ · Not appropriate. · Since the line where expenses should be recognised is not specified under IAS 19, in general, there is some flexibility, in accordance with IAS 1.
· Particularly if a plan is closed, it could be argued that plan expenses are more appropriately presented within financing results.
Include present value of future expenses in the measurement of the defined benefit obligation ✗ · Not acceptable, because IAS 19 requires an entity to recognise administration costs at the time when services are provided. 1All references to ‘pensions note reconciliation’ in the table refer to the requirement in IAS 19.
Costs of managing the plan assets are deducted from the return on plan assets. Such costs might include fees paid to the bank for asset management services, salaries of the management board who manage the trust, investment consultant fees and any tax payable by the plan itself. This tax excludes any tax included in the actuarial assumptions used to measure the defined benefit obligation (for example, social security and payroll tax on contributions).
Costs incurred in managing plan assets are recognised as part of the return on assets. This results in such costs being recognised directly within other comprehensive income, rather than profit or loss. It does not matter if the costs are paid directly by the sponsoring entity on behalf of the plan or by the plan itself.
Accounting for social security and payroll taxes Entity A incurs social security and payroll tax charges of 10% of its cash contributions to a defined benefit pension plan administered by independent trustees.
The taxes are payable in the month following the cash contribution to the trust. Pension payments from the trust to retirees are not subject to further social security or payroll taxes.
Social security and payroll tax charges are linked to the pension, and they do not constitute a separate employee benefit.
Consequently, they should be included in actuarial assumptions used to measure the defined benefit obligation, rather than expensed in the year in which the cash contributions are made.
Payments of the social security contributions reduce the pension liability, because they would have previously been factored into the calculation of the defined benefit obligation.
Taxes related to defined benefit plans are included either in the return on plan assets or the calculation of the benefit obligation, depending on the nature of the taxes. Taxes on the return on plan assets will be part of the actual investment return and will be recognised in other comprehensive income. Social charges or other taxes levied on benefit payments or contributions to the plan will be included in the measurement of the benefit obligation to the extent that they relate to benefits in respect of service before the balance sheet date.
Past service costs are defined as “the change in the present value of the defined benefit obligation for employee service in prior periods, resulting from a plan amendment (the introduction or withdrawal of, or changes to, a defined benefit plan) or a curtailment (a significant reduction by the entity in the number of employees covered by a plan)”.
Benefit changes might produce both an increase in the cost of future service relating to active members (reflected in a higher annual current service cost) and an increased liability for past service relating to current and ex-employees.
The cost (that is, the capitalised present value) of benefit changes that relate to past service and that have not previously been allowed for in the valuation of the plan liabilities, including the unvested portion related to past service, should be recognised in full as an expense in the income statement at the earlier of the following dates:
Increase in benefit of a pension plan But administration expenses are not a finance item and they are not interest on a liability; they are costs for a service that does not arise from a borrowing. In our view, financing is rather the entity’s management of cash and debt in a broader sense.
An entity operates a pension plan that provides a pension of 1% of final salary for each year of service, subject to a minimum of five years’ service. On 1 January 20X4, the entity improves the pension to 1.25% of final salary for each year of service, including prior years.
As a consequence, the present value of the defined benefit obligation increased by, say, C500,000, as shown below:
C Employees with more than five years’ service at 1 January 20X4 300,000 Employees with less than five years’ service at 1 January 20X4 (average of three years of service, so two years until vesting) 200,000 Increase in defined benefit obligation 500,000 The entity has changed the present value of the defined benefit obligation as a result of an amendment to the plan. Therefore, a past service cost of C500,000 should be recognised in the income statement immediately.
The increase in the defined benefit obligation of C200,000, relating to employees who have not yet vested, will include an assumption about how many of them will leave before they complete the five-year vesting condition and thus forfeit their benefit.
A plan amendment occurs when an entity introduces or withdraws a defined benefit plan or changes the benefits payable under an existing defined benefit plan. Plan amendments might arise from a constructive obligation. Informal communications made to employees might result in the entity having no choice but to pay the benefits.
Any loss associated with the amendment should be accounted for at the point when there is a constructive obligation. This treatment applies irrespective of whether the additional past service liability relates to pensioners, ex-employees entitled to benefits that are not yet in payment or current employees, or whether the benefit changes are funded by a surplus or give rise to a deficit.
Timing of past service cost due to plan amendment Entity B announced its intention to amend a pension plan such that benefits are reduced. Under the current plan, employees earn a pension of 1% of final salary for each year worked. The plan will be amended, so that the benefit is based on an employee’s average salary measured over the period from 1 January 20X4 to leaving.
Entity B published amended terms of the pension plan on its website on 20 December 20X3. The entity’s human resource department prepared amended employment contracts in early January 20X4 and sent them to employees for signature.
All amended employment contracts were signed by the end of January 20X4. The amendment is not contractually binding under local legislation until signed by the employee.
The past service cost (negative in this case) should not be recognised before the amended terms become contractually binding. The terms of the defined benefit plan were amended in January 20X4, when the new terms became contractually binding through signature by the employees.
Entity B should account for the effects of the amendment in January 20X4. The date from which arrangements are contractually binding will vary, depending on local legislation.
A curtailment occurs when an entity significantly reduces the number of employees covered by a plan and might arise from an isolated event (such as the closing of a plant), discontinuance of an operation or termination or suspension of a plan.
The gain or loss relating to a past service cost arising from a plan amendment or curtailment is calculated by remeasuring the net defined benefit liability (asset), using the current fair value of plan assets and current actuarial assumptions (including current market interest rates and other current market prices), reflecting the benefits offered under the plan before the amendment or curtailment. This remeasured net defined benefit liability (asset) is compared to the new net defined benefit liability (asset) immediately after the amendment or curtailment. The difference is the gain or loss on plan amendment or curtailment.
The liability does not need to be remeasured if the effect of the remeasurement (including the effect of recalculating current service cost and net interest) is immaterial in accordance with IAS 8. Remeasurement should be considered on a plan-by-plan basis. A liability is not remeasured and therefore current service cost and net interest for the remainder of the period are not recalculated, for plan amendments, settlements and curtailments that have no substance or are clearly insignificant to the plan as whole.
The following are not past service costs:
Distinguishing past service costs from actuarial losses It is important to distinguish between past service costs and actuarial losses. A larger than expected increase in salaries that are considered for pension payments, for example, will increase the plan liabilities relating to past service. This is an actuarial loss rather than a past service cost, since it is a remeasurement of the existing commitment to pay retirement benefits based on final salary.
Similarly, the cost of a higher than anticipated increase to pensions in payment will generally be regarded as an actuarial loss (that is, a remeasurement of the employer’s existing commitment to provide cost of living increases).
The introduction of a new or additional commitment to protect pensions from inflation, where none previously existed, would give rise to a past service cost. This is because such pension increases would not previously have been allowed for in the valuation of the plan liabilities.
In practice, the distinction between past service costs and actuarial losses requires judgement.
Some transactions do not fit easily into either category (For example, the introduction of a government subsidy for the provision of benefits). This gives rise to a change in the cost of the benefit, without a change in the benefit actually received.
Past service costs might be either expenses or income. Past service costs might be positive (expenses) when new benefits are introduced or existing benefits changed so that the present value of the defined benefit obligation increases, or negative (income) when existing benefits are changed so that the present value of the defined benefit obligation decreases.
The treatment of a ‘negative’ past service cost (that is, income) is consistent with the normal treatment of past service costs described above. The net impact can be recognised as a single net expense or credit, if an entity increases certain benefits, and reduces others available to the same employees, in the same period.
Like curtailments, settlements are events that materially change the liabilities relating to a plan and that are not covered by the normal actuarial assumptions. A settlement is a payment of benefits not set out in the terms of the plan. A settlement arises when “an entity enters into a transaction that eliminates all further legal or constructive obligation for part or all of the benefits provided under a defined benefit plan”. Settlements have the effect of extinguishing a portion of the plan liabilities, usually by transferring plan assets to or on behalf of plan members (for example, when a subsidiary is sold or when assets and liabilities are transferred into a defined contribution plan).
Insurance policy for defined benefit pension obligations On 1 January 20X3, entity A insured its defined benefit pension obligations with a third-party insurer (entity B). The insurance policy is in the name of entity A and not in the name of specific plan participants.
However, entity B will pay pensions directly to entity A’s former employees. Entity A can revoke the insurance policy at any time and obtain the policy’s cash surrender value. The insurance policy does not form part of entity A’s bankruptcy estate under existing legislation.
Entity A believes that it retains no further legal or constructive obligation towards its employees for the defined benefit pension plan. While the policy is in effect, the entity is not required to pay further amounts if the insurer does not pay the pension benefits.
The acquisition of the insurance policy is not a settlement of the pension obligation. The insurance policy has eliminated the legal or constructive obligation, and entity A is no longer exposed to actuarial risks.
However, entity A can revoke the policy at any time and obtain the policy’s cash surrender value. The transaction cannot be considered a settlement of the existing pension obligation.
The policy is a non-qualifying insurance policy, because the proceeds of the policy can be paid to entity A. It should be recognised as a reimbursement right, if appropriate.
Option to receive lump sum payment on retirement instead of monthly pension payments Entity A operates a final salary pension plan that gives a right to a pension of 1% of final salary for each year of service. The terms of the plan allow an employee to choose to receive a lump sum payment at the time of retirement, instead of receiving monthly pension payments. The option for the lump sum payment can only be exercised within three months prior to retirement.
The option for a lump sum is not a settlement. It is a normal employee benefit entitlement. The actuarial valuations should include assumptions about the number of employees that are expected to choose the lump sum payments.
Difference between curtailment and settlement The management of entity C has decided to restructure one of the entity’s three business segments and announce the plan at 31 December 20X5. The terms of the plan are as follows:
- In Germany, it will close three of entity C’s six factories and relocate 500 of the 1,000 employees. As part of the relocation package, unions have agreed that, in exchange for retraining, the affected employees’ benefits will be frozen and new employees will not be covered by the plan.
- In Sweden, it will discontinue all activities. Management has agreed with the unions that it will make a one-off payment to all of its employees in exchange for cancelling their pension entitlement.
The negotiated agreement in Germany will trigger a curtailment, since management is making a significant reduction in the number of employees covered by a plan. The closure of half of the factories in Germany is likely to have a material impact on entity C’s consolidated financial statements.
The discontinuation in Sweden will be both a curtailment and a settlement, since the entity will enter into a transaction that eliminates all liabilities for the benefits provided under the pension plan, and will significantly reduce the number of employees in the plan.
An entity recognises a gain or loss on settlement of a defined benefit plan when the settlement occurs as a part of the service costs recognised in the income statement.
‘Before’ and ‘after’ measurement in case of restructuring The facts are the same as in previous FAQ. Immediately before the announcement of the restructuring plan, the net pension liability in each of the countries was as follows:
Germany Sweden C’m C’m Present value of defined benefit obligation (5.0) (7.5) Fair value of plan assets 3.5 5.0 Net pension liability (1.5) (2.5) After the restructuring, the present value of the pension obligation in Germany is reduced to C4 million. The one-off payment made to the employees in Sweden was C3 million.
The gain or loss is the aggregate of the change in the present value of the defined benefit obligations and the change in the fair value of plan assets. Hence, the entity recognises a negative past service cost (gain) of C1 million in respect of Germany, being the reduction in the present value of the defined benefit obligation (C5 million − C4 million). It recognises a loss of C500,000 in respect of the Swedish discontinuance, being the difference between the net pension liability before the settlement (C2.5 million) and the payment made to settle the liability (C3 million). In Sweden, the curtailment has no net impact on the defined benefit obligation, because the curtailment and settlement occur concurrently.
Settlement loss on sale of a subsidiary A wholly owned subsidiary is sold for C100 million. The book value of the subsidiary’s net assets on the date of the sale was C70 million. The subsidiary participated in the group defined benefit pension plan. There was no contractual agreement or stated policy for charging the net defined benefit cost to individual group entities.
The subsidiary, in its individual financial statements, recognised only its contributions payable, in accordance with paragraph 41 of IAS 19. At the date of the sale, the plan was in surplus by C35 million.
As part of the sale agreement, a bulk transfer of plan assets was made to the purchaser’s pension plan. The ‘before’ and ‘after’ measurements of the group plan are as follows:
Before sale After sale Settlement loss C’m C’m C’m Plan assets 160 130 (60) Plan liabilities (125) (100) 25 Plan surplus 35 30 (5) The profit on sale is as follows:
C’m C’m Sale proceeds 100 Less: – subsidiary’s net assets at date of sale (70) – settlement loss (5) (75) 25
Purchase of an insurance policy to fund some or all of the employee benefits relating to employee service in the current and prior periods is not a settlement if the entity retains a legal or constructive obligation to pay further amounts (in case the insurer does not pay the employee benefits specified in the insurance policy). Such a policy will be treated as a plan asset or a reimbursement right.
Calculating the gain or loss arising from a settlement requires a ‘before’ and ‘after’ measurement similar to the measurement of a gain or loss on a curtailment event. The gain or loss on a settlement is the difference between:
Involuntary termination of employee service Entity A announced in December 20X3 that 100 employees would have their employment terminated in February 20X4 as part of a restructuring project that the entity is demonstrably committed to carry out.
Entity A has given the 100 employees an option to have their pensions settled by a lump sum payment of C75,000 per employee, payable at the date of termination in February 20X4. Of these, 90 employees made an irrevocable decision to accept the payment in December 20X3. Entity A expects that the remaining 10 employees will not accept the offer.
Entity A had a pension liability towards the 90 employees of C6,300,000 at 31 December 20X3, prior to considering the lump sum settlement offer. The plan is wholly unfunded.
A curtailment has occurred, giving rise to a past service cost in December 20X3 as a result of the restructuring. The future service of the 90 employees who exercised the lump sum option will not earn them any additional benefits, and the plan is curtailed.
The entity entered into a transaction with the 90 employees that changed the pension obligation at the same time. Entity A should recognise, in its financial statements for the year ended 31 December 20X3, a liability of C6,750,000 (90 × C75,000) towards the 90 employees that accepted the offer.
This results in the recognition of additional expense of C450,000 (C6,750,000 − C6,300,000). This liability should be adjusted, as necessary, when final settlement actually occurs, by way of payment.
Timing of curtailment due to replacement by a new defined contribution plan and subsequent settlement by way of a lump sum payment Entity C operates a non-contributory defined benefit pension plan for its senior employees. Entity C has not funded its benefit obligation. Entity C closes the defined benefit plan to new entrants and benefit accrual, and replaces it with a defined contribution plan on 1 January 20X3.
The benefits with respect to services provided up to 1 January 20X3 are not affected. The terms of the defined contribution plan are still to be agreed with employees at this stage.
Entity C agrees with employee representatives to make a lump sum cash payment of C5,000 on 30 June 20X3. The payment is made in connection with the introduction of the defined contribution plan in exchange for cancelling their pension entitlement under the previous defined benefit plan.
The pension liability recognised in the balance sheet on 30 June before the agreement was C5,500.
The discontinuation of the old plan is a curtailment event.
Entity C recognises any change in the defined benefit obligation, related to the curtailment in January 20X3, when the curtailment occurs.
Entity C recognises a settlement gain of C500 on 30 June 20X3, when a legally binding agreement has been reached, that eliminates all further legal or constructive obligations for the benefits provided under the pension plan in exchange for the lump sum payment.
The settlement gain is the difference between the accrued pension liability before the settlement and the cash payment to senior management.
Settlement with an unrecognised surplus Entity A has a defined benefit plan. At 31 December 20X1, the following amounts were attributable to the plan:
C Fair value of plan assets 135 Defined benefit obligation (100) Surplus 35 This surplus has been deemed irrecoverable in accordance with IFRIC 14; the surplus was, therefore, written off through other comprehensive income.
Two years later, entity A settles its defined benefit plan. When the settlement occurs, all plan assets, with their fair value of C135, are used to settle the defined benefit obligation of C100. The impact, if any, of this settlement on the income statement needs to be assessed.
A settlement loss of C35 should be recognised in the income statement. This was clarified by amendments to IAS 19 issued in February 2018 and applicable to plan amendments, curtailments and settlements occurring in reporting periods that begin on or after 1 January 2019.
IAS 19 requires that an entity should not consider the effect of the asset ceiling when recognising and measuring any past service cost or, in this case, any gain or loss on settlement.
Entity A should first recognise the adjustment to the surplus through other comprehensive income in accordance with IAS 19, and then calculate the settlement loss of C35 according to IAS 19.
An entity need not distinguish between past service cost resulting from a plan amendment or a curtailment and a gain or loss on settlement if these transactions occur together. A plan amendment might occur before a settlement, if an entity changes the benefits under the plan and settles the amended benefits later, in which case the entity recognises past service cost before any gain or loss on settlement.
Plan closed to new employees Entity A operates a defined benefit plan (which is currently in deficit) for all employees that have completed three months’ service. Employee turnover is high, but the number of employees covered by the plan has been stable for a number of years.
Entity A decides to close the existing plan to new employees on 31 December 20X1 and to offer a new defined contribution plan instead. Employee turnover means that the number of employees in the defined benefit plan will diminish as time passes.
Entity A has not committed to reduce the number of employees currently in the defined benefit plan or to change the benefits payable under the plan. The benefit commitment is unaffected by the closure, and the reduction in employees will reflect staff turnover, rather than any action by the entity.
So, the action of closing the plan to new employees does not represent a plan amendment or curtailment. It does not impact the measurement of the defined benefit obligation and, therefore, there will be no accounting implications.
Benefit accrual ceased Assume the same facts as in previous FAQ. A year later, on 31 December 20X2, the entity changes the terms of the plan. Defined contribution arrangements will apply to all future service.
Hence, no further benefits based on final salary will be earned by employees, although the entity will retain its defined benefit obligation in connection with prior service. With effect from 1 January 20X3, the entity has no liability in respect of service after that date beyond its annual contributions to the defined contribution plan.
However, benefits earned in respect of this prior service will no longer be based on final salary, but will instead be based on salary as at 31 December 20X2.
The net pension liability was C10 million as at 31 December 20X2. The plan’s actuary has advised that the net pension liability after the plan amendment is C7 million.
Both the cessation of future benefit accrual and the change from final to current salary represent a negative past service cost. Hence, entity A should recognise a gain of C3 million when the terms of the plan are amended.
Plan wound up Assume the same facts as in previous FAQ. A year later, on 31 December 20X3, entity A decides to close its defined benefit pension plan. It proposes to make a one-off payment which will end its involvement in the plan. The payment will have two elements:
- An amount to cover the deficit in the plan (which remains at C7 million).
- A ‘risk premium’ of C1 million to an insurance company to take over the pension’s liability and risk, so that the plan can be wound up as fully paid.
The entity will have no further obligations in respect of the defined benefit plan after making the payments. Contributions to the defined contribution plan will not be related to the defined benefit plan.
There is no curtailment to account for in the current period, because the members of the plan were no longer earning any benefit.
However, since the payment of C8 million eliminates all further legal or constructive obligation for the benefits provided under the plan, it represents a settlement as defined by IAS 19.
The amount paid (C8 million) exceeds the deficit in the plan, so a settlement loss of C1 million will be recognised when the settlement occurs.
Plan wind up requiring approval Assume the same facts as in previous FAQ. The entity decides to close its defined benefit pension plan as of 31 December 20X3.
However, approval from a government agency is required to settle a defined benefit plan, which is not a formality. The entity has filed the appropriate paperwork to request approval. The approval process is expected to take up to 18 months.
IAS 19 is clear that the high-quality corporate bond rate should be used for discounting post-employment benefit obligations. It is also clear that the term of the corporate bonds should be consistent with the estimated term of the post-employment benefit obligation.
The term of the corporate bonds should be consistent with the estimated term of the post-employment benefit obligation; this reflects the length of the expected payment stream to employees, as currently required under the plan.
This is the case, even though the expected timing of benefit payment could be viewed as being in 18 months’ time (that is, when the entity expects to settle its obligation). The entity is not obligated to settle, even if approval is received. Therefore, settlement cannot be anticipated.
The entity also cannot accelerate the expense by using a rate for a corporate bond of 18 months’ duration, which is lower than the rate on a long-term corporate bond, to obtain a higher present value for the defined benefit obligation.
Furthermore, as noted above, although entity A anticipates a one-off payment of C8 million to settle its obligations, a settlement loss would not be recognised until the settlement occurs.
Plan replaced by a similar plan Entity A acquired another entity some years ago. Both entities have defined benefit plans, which pay similar pension benefits based on final salary and years of service. Entity A decided to simplify its pension administration by introducing a new pension plan to replace the two existing plans. Each employee’s pension entitlement transfers to the new plan, without any change of entitlement.
Does this have any impact on the entitlement of the employees?
The replacement of both plans does not amend the previous plans substantially. Employees retain their benefit entitlement. However, this will lead to one group plan replacing two separate plans of the individual entities and might impact the separate financial statements, if any, of the individual entities.
A settlement occurs together with a plan amendment and curtailment if a plan is terminated with the result that the obligation is settled and the plan ceases to exist. Determining the period in which a settlement or curtailment occurs can be difficult, especially in cases where there is both a curtailment and a settlement, but they are not recognised at the same time.
Further complications can arise when a defined benefit plan is in surplus, but the surplus is not recognised because the entity cannot gain economic benefit from it.