COMMENTS & RESPONSES

VALUATION OF HYBRID RETIREMENT PLANS

 

Click here for the paper. 

 

 

I have actually received three major comments to this paper - one during the Q & A portion of the presentation proper. These are as follows:

 

1.  Citing FASB clarification (since IAS has no such clarification) - there is no need to break down the hybrid program into its DC & DB components. The hybrid program is considered a defined benefit program with the stream of benefits computed as the higher of the projected vested DC account value and the minimum defined benefit. It was also cited that one multi-national accounting firm considered the hybrid plan as either a DC plan or a DB plan but then reservation was made whether all accounting firms have the same view. (For purposes of this paper, let us call this Method 2 and the method presented in the paper Method 1.)

 

2. The preferred method of valuation for hybrid plans by IAS is simply taking the present value of the defined benefit obligation and comparing it to the DC Account Balance. Additional liability exists if the present value of the defined benefit obligation is greater than the DC Account Balance. It was further stated that a long-term return on the assets can not be made as it will give rise to actuarial gains or losses which is not supposed to be done under a Defined Contribution setup.  (For purposes of this paper, let us call this Method 3.)

  

3. The other method is simply to consider the DBO as the higher of the present value of the DB component (I suppose DBO of the DB component) and the projected DC account value.  (For purposes of this paper, let us call this Method 4.)

 

I decided to have a written response to the comments for the benefit of the members of the society who are interested in the paper. There was not much time to discuss them during the presentation. Also, it would be hard to present arguments without being backed up by illustrations.

 

But first, let me thank those who commented. It affirmed the relevance of the paper. It is not, after all, simple as it may seem.

 

Let me emphasize too that, as written, Method 1 is a suggested method. I do not claim that this is the only correct method of doing the valuation knowing too well that there are always many approaches or solutions possible to every problem. I did present in the paper one method that I think is "grossly" incorrect because it is being done by some peers with the hope they adopt an acceptable method.

 

Finally, I believe that there is no preferred method as far as the IASB is concerned. Actuaries (and auditors) may have preferences and based from the comments themselves, there are differing points of view. One thing is sure though – even if there is such thing as preferred method, it does not mean actuaries can not use other methods, otherwise it should instead be called the prescribed (not preferred) method. Or if IASB really has a preferred method, then let this paper be an appeal to consider (or reconsider) the method presented as a preferred method.  Members of the PSRC may also take a closer look and evaluate if this method violates its standards.

 

Now, why don't I prefer Methods 2, 3 or 4?

 

Method 2. The major reservation that I have with this method is that it does not preserve the original nature of the DC plan which is effectively the base plan. Under Method 1, the base cost is the contribution due (which is usually a percentage of salary) plus an additional amount to cover the expected shortfalls of the DC Account Balance from the defined minimum floor of benefit. The segregation of costs under Method 1 also allows for the application of PAS 19's DC rules on the DC component.    

 

To illustrate the difference, using the example in Part III.A of the paper, the benefit allocation for the benefit of P 56,275 and the current service costs would be:

 

Benefit / Cost Allocation				Current Service Cost
Method 1			Method 2
DC	DB	Total		Method 1	Method 2
7,200	2,933	10,193	11,255	9,570	8,915
7,416	2,933	10,409	11,255	9.929	9,450
7,638	2,933	10,631	11,255	10,302	10,017
7,868	2,933	10,860	11,255	10,691	10,618
8,104	2,933	11,096	11,255	11,096	11,255

 

The difference can be highlighted using an extreme example – suppose that at a certain valuation date, the projected DC account value is greater than the minimum benefit at all future durations (it could happen depending on prevailing assumptions as of each valuation date). Method 1 treats it exactly as a DC (DB component becomes nil). Method 2 still treats it as DB and does the benefit allocation.

 

I also have some reservation in considering the vested DC account value instead of the full DC account value. How is a pure DC plan with benefits subject to a vesting schedule classified? From the PAS 19 definition, it seems it still is a DC plan (note that it still satisfies the qualification that there is  no legal or constructive obligation to pay further contributions) and probably the proper expense to be recognized is the contribution for the period less forfeitures in the same period.

 

Now, between a pure DC plan and a hybrid plan whose DC component is exactly the same as the pure DC plan, which one should have a higher liability, if ever? Definitely it should be the hybrid plan because it provides more benefits. However, Method 2 would produce a liability for a hybrid plan lower than a pure DC plan in certain circumstances because the method takes into consideration future forfeitures.

 

Nevertheless, assuming future forfeitures should be considered, Method 1 should be able to easily take care of it under the DB component in the form of negative costs.

 

Method 3. This method prevents a false asset since the liability is taken as the higher of the DBO based on the defined benefit component and the DC account balance on a per-individual basis. However, it may also result in an understatement of the liability.

 

I also doubt that it maintains the actuarial equivalence principle (DBO + PV of Current Service Costs = PV of Future Benefits) at all times or if it does, I doubt whether the Currents Service Cost is computed using the Projected Unit Cost method or any semblance of it.

 

To illustrate:

 

A hybrid plans provide for a contribution rate of 10% of salary but with a minimum defined retirement benefit of P 100,000 upon reaching age 60.

 

The Company has two (2) employees with exactly the same profile:

*         Entry age: 55

*         Current age: 58

*         Current Annual Salary: 100,000

 

Let us assume the following:

*         Discount rate: 6.0% p.a.

*         Expected long-term return on plan asset: 6.0% p.a.

*         Salary increase rate: 0.0% p.a.

*         Retirement Rates: Prior to Age 60 - 50%; Age 60 - 100%

*         Death and disability rates: None

*         Contribution payment: Annual at end of year

*         Current DC account balance: 40,000 for each

 

Suppose, after another year, one of the employees separated and got his account balance of (40,000)(1.06) + 10,000 = 52,400.

 

Using this method, we determine the DBO as follows:

*         DBO (for the 2 employees) = (2) Max[(0.5)(100,000)(3/5)(1.06)^-2, 40,000] = 80,000

 

The understatement is quite obvious. With 40,000 in the account balance and remaining expected contribution of P 20,000, for the next two years, at a reasonably achievable interest rate, the account balance will not reach P 100,000 at expected retirement date.

 

Continuing with the DBO after one year when one employee has left:

*         DBO = (1) Max[(100,000)(4/5)(1.06)^-1, 52,400] = 75,472

 

I don't exactly know how the CSC is computed under this method. Most likely, it should follow the DBO formulation:

 

*         CSC (for the 2 employees) = (2) Max[(0.5)(100,000)(1/5)(1.06)^-1, 10,000] = 20,000

 

Given that there is no actuarial gain or loss because there is no change in assumption and actual experience exactly followed the assumptions, the CSC can be determined in another way based on the following formula: DBO_beg + CSC + Interest Cost - Benefit Payment = DBO_end.

 

The CSC will then be 75,472 (DBO_end.) - 80,000 (DBO_beg) - 4,800 (Interest Cost) + 52,400 (Benefit) = 43,072. This is different from the CSC of 20,000 earlier calculated.

 

In contrast, these concerns are not present in both Methods 1 and 2 where CSC can be computed using the PUC method and the movement of DBO serves to validate the soundness of the formula.

 

Under Method 1:

*         Expected DC Account Balance at Age 60: 40,000 (1.06)² + 10,000 (1.06)  + 10,000 = 65,544

*         Expected shortfall for each employee: 100,000 - 65,544 = 34,456

*         DBO_beg for the 2 employees: (2) [(0.5)(34,456) (3/5) (1.06)^-2  + 40,000] = 98,399

*         Current Service Cost for the 2 employees: (2) [(0.5)(34,456) (1/5) (1.06)^-1  + 10,000] = 26,501

*         DBO_end for the remaining employee: (1) [(34,456) (4/5) (1.06)^-1  + 52,400] = 78,405

 

The movement in the DBO serves to validate the formulation: 98,399 (DBO_beg) + 26,501 (CSC) + 5,904 (Interest Cost) - 52,400 (Benefit Payment) = 78,405 (DBO_end)

 

It should be true also for Method 2.

 

On the issue about assuming an expected return on the DC account balance that should give rise to actuarial gains and losses, it should be noted that under Method 1, the recognition of the expense corresponding to the DC component is simply the contribution due for the period and is not affected by any gain or loss. Any gain or loss is instead considered under the DB component.

 

Finally, why a different long-term return on asset may be assumed different from the discount rate? This is simply because there is always a possibility that the discount rate can not be achieved depending on the investment policy of the plan sponsor. That possibility is greater here in the country as there is not enough supply of long-term bonds. I have re-validated this in the Q&A portion in another paper presentation on investments during the convention.

 

Method 4. I can not exactly imagine how the method works. It should be noted however that it may not be appropriate to consider the two as independent benefits. The defined benefit may be higher than the DC account value in some duration and vice versa.

 

My reservations on Method 3 apply also to this method.