The Appropriate Methods to compare various funding solutions both fairly and objectively

By Dan Johnson

.

Notes on Weighted Scores Methodology

In daily life, there are some decisions regarding product choice that are simple and straightforward, for example, when choosing between two products that are essentially the same but have different prices.  But in many important decisions, there are trade-offs necessary between multiple desirable criteria.  One such situation would be when one is buying a car, the buyer may want a powerful engine and great handling but also want excellent fuel economy and low price.  These criteria have inherent conflicts.   Another example comes to mind when creating lists of the most desirable cities to live in a country.  How much to we value warm weather versus low crime, or excellent schools versus low cost of living, or great access to medical care, or clean air versus excellent job opportunities, etc.  Any ranking system that attempts to figure out the “best” place to live has to assign weights to these various criteria, and also form some sort of objective numerical scale for the data found in various cities.  Almost every person would have a different set of weights (representing their personal values), and so it is extremely likely that people would prefer different cities even if they agreed on all the data.

Weighted Scores Methodology (“WSM”) is usually adopted when a fair, explicit and objective attempt is being made to find the best product or solution under complex criteria.  This method is used broadly whenever trying to make a single decision under multiple criteria and where trade-offs are necessary.  In such cases, one cannot have it all, and in general, by carefully assigning “weights” or relative values to the various measures or criteria, one can determine and document the “best” solution given those weights.

When trying to find the best financial product or solution in the context of optimally funding a benefit plan, similar trade-offs are necessary, and so WSM is applied. In the simplest of examples, it is often true that companies with the best performing products may not have the highest ratings possible for financial strength.   Or a product with the lowest cash flow requirements may have somewhat worse results with respect to its impact on corporate earnings.  Or there could be quite different results based on long-term versus short-term criteria.

Weighted Scores Methodology creates a “scaled” score for each product tested.   It is a method similar to creating scaled scores for the SAT or other standardized test, and also is very similar to what happens when a teacher grades students “on a curve,”  i.e. using relative scores.   To take the example further, a teacher could decide to create “absolute” scores by deciding ahead of time that anyone correctly answering 95 or more questions out of 100 will get an A, anyone getting 87-94 will get a B, etc.   Normally this works when a teacher knows ahead of time what the performance standards should be.  It would allow everyone to fail or everyone to get an A. In contrast to this, a teacher could decide to create “relative” scores (grading on a curve) by simply seeing how students do, and then assigning grades based on the actual results.   This would prevent everyone (even very good students) from getting bad grades on a difficult test, and prevent everyone from getting an A on an easy test.

When evaluating the funding performance of a financial product, it is quite impossible for the evaluator to know ahead of time what the right answer might be.  Among other things, this is because of the complexity involved and the dependence on dozens of important assumptions.  So there is really no substitute for simply seeing what answers come up, given fair and equivalent assumptions for each financial product.   Grading on a curve, or using relative scores, is really the only feasible method to employ.

More specifically, the method will look at the range of actual outcomes for the products being compared.  The more products compared, the “better” or “fairer” will be the results.  Since in any real case we are looking at only a subset of the possible products that could be used, the grading scale does not take into account the fact that there could be other product results not compared that could theoretically be better or worse than the results actually seen.  This means that scores may move around some as additional products are added to the comparison. Since WSM is truly a relative scoring system, a natural corollary is that the more products that are compared, particularly those whose characteristics are somewhat different, the more accurate and/or robust the assessment is likely to be.

For each evaluative measure/criterion, this method assesses the minimum and maximum actually achieved, and then creates a linear scale for all results that fall in-between the adjusted minimums and maximums.  These intermediate scores are then “weighted” based on the specific weights (relative importance) assigned to the criterion.   The sum of the weighted component pieces becomes the final weighted score assigned to the product.  In our specific use, we scale this to 1000 being the perfect score (best in all categories) and 0 resulting from being the worst product or solution in all categories.

The final score achieved is truly a relative score, and will vary if other products are added to the assessment.   This is much like what would occur in a classroom. If several more students take an exam that is graded on a curve, it could affect everyone else’s results. Clearly, adding geniuses or idiots would have a more skewed impact on other scores.

It is true that the presence (or absence) of variation in the results has an impact on the relative scores.  Again, it is generally preferable to have material variation in the actual results compared.

It is important to remember that the scores depend very much on the relative weights chosen for each measure.  These weights are meant to assign the relative priority of the various evaluative criteria.  In general, if the weights are changed, the method will produce different scores, and could produce a different product or solution having the best score.  Care should therefore be taken in the choice of weights.  In most financial product comparisons, the results are reasonably sensitive to the actual weights used.

In the end, WSM is a fair and objective method to compare and assess a financial product when trade-offs need to be made among the various desirable outcomes.  It is also self-documenting, in that it discloses to any reader of the results precisely what the basis for making the decision was.

Notes on Methods for Comparing Financial Products

There are a number of important principles for the fair comparison of financial products or solutions.  There are actually several distinctly different “Funding Methods” used, and the “Evaluative Criteria” used by Weighted Scores Methodology must adapt itself to the nature of the funding method.   This is not obvious at first consideration, so it will be useful to go over this important point in some detail.

Common Assumptions

The most obvious ones involve the use of the same, or equivalent, economic and/or actuarial assumptions when doing projections of the financial product or solution.  Any distortion in these assumptions would normally render the comparison meaningless and misleading.  These include tax rate assumptions, discount rate assumptions, gross investment return on equivalent investments, etc.

Beyond such basics, it is also very important to make sure that the products are given the same task, i.e., made to perform the same work.  There are three basic funding methods for the modern informal funding of non-qualified benefit plans (“NQBP”), and each involves somewhat different work.  By “modern” we are recognizing that the NQBP word has become much more oriented towards defined contribution plans, especially as 401(k) Excess/Mirror Plans or other Share-Based Deferral Plans became popular, and stock market returns were considered attractive, and investment choice on the part of the plan participants became desirable, and mutual funds and/or variable life products became the dominant funding vehicles.

With modern plans, the most common funding method became a matter of matching the funding amount to the deferral and/or corporate contribution amount (“Contribution Matching”).  This operates similarly to how a 401(k) plan would operate, except that it is generally necessary to have two sets of books (an NQBP is inherently a liability-side plan), both liabilities and assets.   With Contribution Matching, the plan sponsor’s cash flows are normally the same across the different products, at least in the early years, so that cash flow comparisons are not helpful unless they are for the entire duration of the projection.

Since NQBP plans (generally 409A plans) have somewhat different characteristics than qualified 401(k) plans, in particular that they do not need to be funded at all, and in general, the funds are separate and separable from the liabilities generated by the plan.  While this is quite obvious to practitioners, it is forgotten by plan sponsors with surprising frequency.  The asset side, of the funding products, normally have either some amount of taxes that are incurred along the way (mutual funds, for example, due to investment costs, and taxes on non-excluded dividends and realized capital gains) or have some amount of expense (mortality costs, for example) in order to get into funding products that are not normally subject to any current tax.  This means that even though the incoming cash flows are nicely matched (between the deferrals or contributions and the amounts invested in the funding products), the asset value results are highly likely to be somewhat different than the liabilities.  This is true even though the plan sponsor or rabbi trustee normally invests the assets in parallel accounts to the liabilities (“Parallel Investment”}.

A way to deal with this issue is to make a change in the funding method to “Asset-Liability Matching.”  In this funding method, the goal is to solve for the amount of money to invest in the asset-side funding products at the beginning of some period (most commonly, a year) such that the resulting assets just equal the projected liabilities at the end of some period, under reasonable assumptions.  Put simply, it is a balance sheet method that seeks to minimize any difference between the liabilities of the plan and any funding products used to informally fund the plan.

A third common funding method seeks to minimize the amount of cash required to put into any funding product over time while still being able to take all plan distributions out of the funding vehicle.   This is simply called “Cash Funding,” and is very common for dealing with defined benefit plans (where the only thing really defined is the amount of money coming out) or whenever cash flow into the funding vehicles is desired to be minimized.

These three fundamental funding methods are summarized below:

Comparative Summary of Funding Methods

Common Assumptions:

• Aggregate Funding Methods
• Parallel Investment
• Funding asset is managed in such a manner to either avoid creating asset-side taxes for the plan sponsor or if taxes are created, will compensate for such taxes using the asset itself.
• Common gross investment return assumptions (though a preference for different funds or fund families can be properly taken into account in the evaluation)

Contribution Matching

• Brief Description

1. The sum of the participant new deferrals and corporate contributions determine the precise amount of new funding.
2. When benefit payments are due to be made, withdrawals or loans from the asset are used to create the cash for either the pre-tax or after-tax payment (usually after-tax).

• Strengths

1. Easy to determine the asset-side investment amount.
2. Easy to explain to people not involved in the administration of the plan, since it seems to run precisely like a 401(k) plan would run.

• Weaknesses

1. Does not correct or adjust for under-funding due to loads, charges or taxes on the asset side.
2. Does not correct for over-funding due to death benefits received, loss of non-vested liabilities from terminating participants or any other reason.
3. Has no internal or natural method of lapse-prevention when COLI is the funding asset.

• Criteria to Distinguish different Funding Products when Evaluating

1. Since the early cash flows are essentially the same for all asset classes, the most common criteria are the comparative asset values projected at different durations.  These are normally after-tax values to create equivalence.  The durations should include longer durations, as different asset classes have quite different durational characteristics.
2. Over time, the cash flows actually do diverge significantly, as death benefits get paid for any COLI assets.  Even among COLI products, the type of DeMECing, the nature of the underwriting, the relative amount of death benefits versus cash value, etc. can differ materially from one product to another.  So measures of long-term cash flow, such as NPV of after-tax cash flows for the duration of the plan (at a suitable discount rate provided by the evaluator), and the IRR on long-term cash flows, are also important and suitable criteria to take into account.

Asset-Liability Matching

• Brief Description

1. The benefit plan is carefully projected with particular attention paid to benefit-side liabilities generated, and then iterative solves are done to determine the lowest amount of funding required each year (separately, year by year) that allows the funding product to internally deliver an asset value (usually after-tax) that precisely matches the benefit liability.
2. The benefit liability target can be either before-tax or after-tax, or in some cases, is set to be a pre-determined percentage of the before-tax liability.
3. The percentage of the liability target can vary year-by-year according to plan sponsor desires.
4. No specific action is taken to withdraw or borrow the actual benefit payments due, although this method naturally provides much to most of the cash needed by its simple operation.
5. For COLI products, death benefits are normally collected into an account and held as taxable cash to provide an additional amount of assets set aside to informally fund the liabilities.  Of course, these can be used to help pay benefits as they come due, but it is not inherent in the funding method to do so.

• Strengths

1. Naturally adjusts to the patterns of the benefit liability side, which can prevent both over-funding and under-funding.
2. The most natural and easiest to explain to a financial audience, though not necessarily to other audiences that are accustomed to think of the NQP as if it were a 401(k) plan.

•  Weaknesses

1. Normally costs more than the other methods, though that depends on specific choices made by the plan sponsor.
2. Is somewhat difficult for comparing products

•  Criteria to Distinguish different Funding Products when Evaluating

1. Since this method creates the same percentage of target liabilities for all products, there is no difference in any asset value measurement, so these should not be used.
2. The key difference is the lower or higher amount of cash needed to achieve the match.  However, this amount is very likely to be different between products year-by-year.
3. Therefore, a key criterion is the NPV of after-tax cash flows (at a suitable discount rate provided by the evaluator).
4. Since these NPVs are different depending on the durations of the cash flow included, it is recommended that the asset-side NPV through variations different durations should be used, probably including one through the projected end of the program on a closed block basis.
5. IRR on projected after-tax cash flows on the asset side is also a legitimate and important criterion in evaluating different product options.

Cash Funding of Benefit Payments

• Brief Description

1. The benefit payment obligations are carefully projected, and then complex iterative solves are done to determine the lowest possible “level” amount of funding required for a pre-determined duration that allows the funding product to internally deliver the cash required at the right times.  Normally the after-tax cash is used, since the funding asset either avoids or takes care of its own taxes.
2. If individual-level benefits are due upon the death of the participant that are higher or add to the living liability, this amount is indemnified or compensated for when using COLI products.
3. When benefit payments are due to be made, withdrawals or loans from the asset are used to create the cash for either the pre-tax or after-tax payment (usually after-tax).
4. The funding solve insures that a COLI product is protected from lapsation.
5. The pattern of funding can include other patterns, such as compound increases or decreases, tracking of the aggregate effective salary scale, etc. to create a more customized solution

•  Strengths

1. Can lower the cost of funding to the minimum possible while preserving all constraints, i.e. this can be the lowest cost funding solution while still doing the intended job.
2. Since a solve is used to determine the level investment amount, there is a natural ability to prevent lapses for COLI products
3. Naturally adjusts to all changes on the plan liability (benefit) side, as a new solve should be done for the funding level at anniversary reviews.
4. Very comprehensive; naturally adjusts to all events.
5. Easily measured and compared; all things revert to cash, allowing easy differentiation.
6. Allows the greatest degree of funding/financial optimization.

•  Weaknesses

1. Does not intentionally solve for balance sheet impact, though this is projected
2. Does not intentionally solve for impact on corporate earnings, though this is projected
3. The strength is also the weakness: the focus is really on cash.

•  Criteria to Distinguish different Funding Products when Evaluating

1. The cash flow solved for, even right at the beginning (the “solve point”), is a key distinguishing factor and should figure prominently
2. Other measures of comprehensive long-term cash flow, such as NPV of after-tax cash flows for the duration of the plan (at a suitable discount rate provided by the evaluator), and the IRR on long-term cash flows, are also important and suitable criteria to take into account.
3. Measure of product asset value and/or impact on earnings, despite being important, are problematic to use, since the best (most efficient) funding products will have lower cash flows and thus lower asset values and lower positive impact on earnings.  One can always create higher asset values and earnings by paying more money in, but this method seeks to reduce funding to its lowest amount possible.  Experience suggests it doesn’t work to use these criteria when within the Cash Funding method.