Funding Methods: Past, Present and Future

By Dan Johnson

Background, Context and Terminology:

 Back in the days when most Non-Qualified Benefit Plans (“NQBP”) were Defined Benefit in nature, Informal Funding of the plan with Corporate-Owned Life Insurance (“COLI”) involved projecting the benefit liabilities and then either buying enough COLI to take the cash retirement benefit out of the policies (“Cash Funding”), or buying enough life insurance such that the cash received upon the death of a participant was more than enough to offset the cash cost of the benefit liabilities (“Death Cost Recovery Funding”), or buying enough COLI such that the positive earnings generated by the COLI would more than offset the negative earnings due to the benefit liability accruals (“Earnings Offset Funding”) or buying enough COLI such that the net present value of the COLI cash flow just offset the net present value of the benefit-related cash flows (“NPV Funding”) or a few other schemes, including just buying as much COLI as you can (or are allowed to) because it’s after-tax return is excellent (“COLI as a Great Investment”), as is done in many BOLI situations still.  The existence of many variations of these methods, including after-tax versus before-tax variations, does not obviate the general statements that can be made about them.

As the NQBP world moved toward and became much more Defined Contribution (true Deferrals) in nature, funding the plan with COLI started to mean different things.  This was especially true as 401(k) Excess/Mirror Plans or other Share-Based Deferral Plans became popular, and stock market returns became high and variable life products became the dominant funding vehicle.

For the most part, fundingNQ Deferral Plans became a matter of matching the COLI premium to the deferral contribution amount (“Contribution Matching”), and then investing in variable funds that approximately matched the investment characteristics of whatever the deferral accounts (the liability side) were invested in (“Parallel Investment”).  As practice became better established, and COLI carriers lined up large numbers of relevant funds that paralleled the popular benefit account choices, Parallel Investment was simply assumed.

In the earlier days and generally (though not always) in simpler sales/admin situations, the funding was set up so that each individual’s contributions were matched by the same premium payments on their individual policies (“Individual Contribution Matching”).  While simple in concept, this created a number of problems, mostly of an underwriting nature, but also of a TAMRA nature and sometimes of an administrative nature.   A rather natural solution to some of these problems was afforded by aggregating the contributions of the whole group (and netting them out when appropriate) and then spreading these evenly across a group of policies (“Aggregate Funding”).  Actually, Aggregate Funding has been around a long time, and has often been used when COLI as a Great Investment was the funding method, at least as far back as the 1970’s.  When Aggregate Funding is used in the context of matching the aggregate net deferral contributions with aggregate premiums, we can call it Aggregate Contribution Matching.   Again, Parallel Investment is normally assumed to be the case, though the frequency of changing allocations or rebalancing portfolios has evolved to greater sophistication with the advent of good on-line deferral administration systems.

Before going deeper, let me continue with some breadth and pick up the range of funding techniques.  While Cash Funding on an individual policy basis (“Individual Cash Funding”) has diminished in frequency somewhat, it continues to be a standard technique particularly in the lower end of the market and when defined benefit plans are used.  Similar to the discussion above, to solve underwriting, TAMRA and administrative problems, Aggregate Cash Funding of benefits has become a somewhat common technique, though it‘s suitability can be questionable if there are significant benefit payments due upon the death of an individual participant (especially pre-retirement DBO Plans or Endorsement Split Dollar).  Liabilities contingent upon the death of a participant are very commonly used in conjunction with supplemental retirement income plans, especially when defined benefit in nature.

Some Problems with Contribution Matching

At the present time, however, Aggregate Contribution Matching (along with Parallel Investment) has become the most common funding method for reasonably sized plans.   There are, however, several problems with Aggregate Contribution Matching that the industry and practitioners need to address.   These derive from facts that are obvious to those of us in the industry, but need to be stated in order to properly understand these problems.   Most buyers of COLI-funded defined contribution NQBP’s have a 401(k) model in their heads as their starting point.  They tend to think of the non-qualified layer as a device to break through the caps and ceilings imposed on 401(k) Plans or other tax-qualified defined contribution plans.  At the risk of stating the obvious, when an individual participant defers money into a 401(k) Plan, the deferral equals the contribution naturally, and more important, the assets of the plan and the liabilities of the plan are the same (other than vesting considerations).  In fact, the whole concept of Funding is not an issue at all.  When the plan assets equal the plan liabilities by virtue of the nature of the plan itself, what do we mean by funding?  It’s not even a concept to talk about.

When 401(k) Mirror Plans (or their many variations in names, e.g. 401(k) Excess Plans, 401(k) Look-alike Plans, etc.) are sold with Contribution Matching as the general method, it seems to most buyers that the same kind of thing is happening.  In spite of what we might think as the repeated efforts to remind buyers that in non-qualified plans, the assets and the liabilities are most definitely NOT the same, I have been unpleasantly surprised by how often company administrators and individual plan participants do not understand this most basic of points.  Worse, they think they understand even when they’re wrong, or at least don’t know what they don’t know.  Whenever we have buyers in the industry who either don’t know what they don’t know, or are victims of “the things they know that just ain’t so”, we have the potential for real trouble.

One important reason why this issue can be poignant is that if we look carefully at the emerging assets and liabilities, in the great majority of cases, the assets lag the liabilities.  This is almost invariably true when the funding concept was Contribution Matching, regardless of whether that was done as Individual Contribution Matching or Aggregate Contribution Matching.  The reasons for this are not too hard to understand.  COLI has loads (for commissions, expenses, taxes, mortality, etc.) and the funds under a COLI policy that are used as parallel investments tend to have healthy loads as well.  In contrast to this, the benefit side of the equation (the liability side) is generally tracked without loads.  Typically, 100% of the participant deferrals go to the liability side, and the shares or units they are applied to generally have Net Asset Value (“NAV’s”) quoted either unloaded or lightly loaded.  But even if the NAV’s are the same, the loads mentioned above in the COLI product mean that the contributions or number of shares are different.  Unless plan designs were used to avoid this (not typical), it is quite hard for the assets to reach or exceed the liabilities.

Moreover, most sales illustrations using Contribution Matching (again, whether Individual or Aggregate), assumed level contributions at time of sale and reasonable (and consistent) positive investment returns.  While some plans are designed to force this to be the case, any true 401(k) Mirror Plan will experience lots of variations in reality.  Salaries and bonuses go up and down; we’ve seen a lot of downward movements in the last few years, especially compared to the roaring 90’s.  And fund values go up and down more severely than has usually been planned for.   While many good practitioners did Sensitivity Studies with respect to different fund returns on the liability and asset sides, the range of sensitivity studies was likely rather narrow.  While again, almost all of us are aware of the vicissitudes of the stock market and various funds, it is instructive to consider what happened to the main indices recently, as these usually are less volatile than individual funds.  The Dow Jones Industrial Average, after a long upward trend that culminated in a 25.2% gain in calendar year 1999, lost 6.2% in 2000, then lost 7.1% in 2001, then lost 16.8% in 2002, and then went up 25.3% in 2003.  The S&P 500 and Russell 2000 indices showed even greater volatility, e.g. the S&P 500 lost 23.4% in 2002, and the Russell 2000 gained 45.4% in 2003.  But as you many imagine, these paled compared to the NASDAQ Composite, which rose 85.6% in 1999, followed by a 39.3% decline in 2000, then a 21.0% decline in 2001, then a 31.5% decline in 2002 only to be followed by a 50.0% gain in 2003.   Again, individual funds tend to vary even more than the indices of their exchange, though “stable value” funds do exist and can be deployed.   I think we, as an industry, are somewhat fortunate that the net effect of all this volatility is, at the present moment, not nearly as dramatic as it would have been if measured over different ranges of years.  That is, the recovery in 2003 washed clean a lot of the problems from prior years.  But still, the point is that it is hard to see how the plan liabilities and assets would stay in sync given the realities of what can happen over time.

Getting to Better Funding Methods in this Context

The point of all this is that many corporate sponsors and participants think that plan assets and liabilities are the same (falsely), and many buyers who know better still expect the assets and liabilities to be approximately the same due to what was represented at time of sale (or since) and due to their starting point of the 401(k) world.  And actually, this starting point is not as unnatural as I may have made it sound.  Because matching assets and liabilities is, in fact, the natural and in most ways, the best conceptual method to informally fund a non-qualified benefit plan.  In contrast to the qualified plan trust world, where the plan assets and liabilities are moved off balance sheet and only the net difference is kept on balance sheet, in the non-qualified informal funding world, the entire assets and liabilities stay on the balance sheet in all normal circumstances.  Matching them up, called Asset-Liability Matching, is the core of what informal funding should be about, and when it is achieved, it is hard to be very critical of the funding method, unless the costs to do so are excessive in the short term.

Just as discussed above, Asset-Liability Matching can be done in several ways with COLI.  Certain plans can arrange the funding such that the assets of an individual COLI policy (normally the cash value at some point in time) just match the liabilities generated by the individual’s deferral accounts at the same point in time (“Individual Asset-Liability Matching”).  Due to underwriting,TAMRA and administrative issues, the assets of a group of COLI policies (the composite cash value) can be set to just match the liabilities generated by a group of deferral accounts (“Aggregate Asset-Liability Matching”).  But whether done on an individual or aggregate basis, Asset-Liability Matching can be considered the “Gold Standard” or strategic heights of funding non-qualified deferral plans.

Responses and Solutions for our Current Context

We at AFS have spent considerable effort to facilitate more precise, more convenient, and smarter technology to address these issues and the additional issues that are encountered when drilling down a layer or two into this world.  Let me explain this in somewhat more detail.

(1) First, we have made into the clear market leader for administering any kind of share-based non-qualified deferral plan.  Its participant web access functionality is better than people experience with online 401(k) plans, and the administrative functionality is now way out front.  This has taken many years of consistent technological development and millions of dollars of investment.  To provide the proper focus and business model, we have separated as a company and as a business line, and have given it the proper freedom to allow anyone with any plan, regardless of carrier funding, (or whether it is unfunded or mutual fund funded) very affordable access to this platform.  We have priced very aggressively, and allow co-branding options that provide nice marketing tie-ins.  The content is rich, the development engines are rolling rapidly, and we are experiencing rapid growth in enrollment of individuals and plans. 

(2) Second, we have designed and deployed Asset-Liability Matching, on both an Individual and Aggregate basis.  We have put in numerous new functions that make this easy to use, and dozens of new output columns to allow for assessing an individual or a case.  Of particular interest are functions that, upon drilling down a bit, are crucial.  To achieve asset-liability matching, either on an in-force basis or from time of issue, there is a solve that must be done to bring the two sides together.  Looking at the extra premium (over and above the basic contribution matching) needed to achieve asset-liability matching (or the withdrawal that can be taken to accomplish the same) is important.  The present value of that stream of extra premiums or withdrawals is an exceptionally good measure of the efficiency of the COLI being used.  Moreover, we recognize that some prefer to do the asset-liability matching on a “beginning of period” basis, while others prefer to do so on an “end of period” basis.  We have allowed for the benefit liability data to come into play in four different ways: (1) generated by AFS internal benefit logic [using our system to model and track the benefits, (2) generated by manual input [you can type in the streams, either on an individual or aggregate basis], (3) generated by pointing to an Excel spreadsheet range, and (4) generated by a Dynamic Link to an account being administered on  This last option allows for daily, real-time, automatic pick up of liability (and in some cases, asset) values.   Whether or not the elegant solution of a dynamic link with is available, we have also installed the new ability to correct deferral account balances on both a beginning or end of period basis, which allows the internal benefit logic at AFS to be much more powerful.

(3) Third, we have installed better De-MEC’ing solves, so that even in the presence of wild variations in deferral amounts or fund performance, policies can effectively be kept on the near side of the MEC border.  Our TAMRA solves, long the industry standard, have thus been pushed out even better.

(4) Fourth, we have significantly re-engineered our methods of selecting the right funding options, on both our desktop platform (the AFS Master System) and our online platform AFS-Link, located exclusively at   This re-engineering effort recognizes the important issues and principles discussed above, and changes the terminology/nomenclature.  We recognize the various funding methods as described above, and then make it much easier (and far more intuitive) to switch between them.  For example, it will now be possible to switch from Individual Asset-Liability Matching to Aggregate Asset-Liability Matching with great ease.  Or, it will also be easy to switch from Individual Contribution Matching to Aggregate Contribution Matching.  Similar ease will be rolled out soon for Individual Cash Funding and Aggregate Cash Funding.

(5) Fifth, within the set of aggregate funded methods, we will be offering carriers new and better options.  Beyond the usual distinction between equal premiums and equal face amounts, we will be offering equal net amount at risk and also “factored death benefits”, which have a strong potential for finding a better globally optimal funding solution.  This is more for “Phase II” of our new offerings but is an exciting technology to keep your eyes on.   We are also going to be offering carriers additional drill-down methods, such as distinguishing Level or Non-Level Death Benefits (within a policy) as well as Equal or Unequal Death Benefits (across the set of funding policies), new special solves, etc.

(6) Sixth, we have set up an excellent mapping architecture between and, such that in going from Individual Funding (one to one), one does not have to go all the way to “gross” Aggregate Funding (all to all).  We are allowing one to many, many to one, and many to many.  This is a much more “refined” and “granular” way of doing aggregate funding, and will be preferred in many cases and for a variety of client reasons.

(7) Seventh, the consideration of mortality within aggregate funding is an important issue.  We have already created some options for making mortality something other than a pure expense (e.g. partially offsetting the benefit liability need in Aggregate Cash Funding of Benefits).   We are incorporating this functionality into our Asset-Liability Matching logic as well, but because the basic funding method is different, mortality must be handled differently in these cases.  When you think about it, mortality proceeds from life insurance contracts (whether done in a partial mortality manner or a discrete mortality manner) are a different kind of asset than the cash value of contracts.  In other words, we have two asset classes to help offset the liability of a deferral plan.  The inclusion of this separate asset class makes life insurance look much better than mutual funds (than would be the case without this).  Our output and reports, as well as our solves, will reflect this.  This is an offering available now for our clients.

(8) Eighth, the use of partial mortality as described above is an expected value projection.  Though expected values are a highly valid technique in general, their reliability in predicting future cash flows is poor when the number of lives in the group is small.  In general, this reliability (properly measured), increases with the square root of the number of the number of participants.  In small cases, it is generally considered not reliable at all. However, there are techniques that can be employed to move from a discrete mortality model to partial mortality over time.

(9) Ninth, we have deployed new patent-pending technology for optimizing the structure of an insurance contract or set of contracts, based on the weighted set of goals and objectives input by the user.  This weighted scores method, which creates a “Product Suitability Score”, is a true revolution in the way of structuring financial products (“Just-In-Time Design”) based on client desires.  We will be publicly deploying this revolutionary technique, using a very rapid close algorithm our top development team has created, in the very near future, after extensive work with one of our carrier clients.  In the specific context of the funding of NQ Deferral Plans, this optimization technique really shines.

(10) Tenth, we have deployed new logic for doing Sensitivity Studies with our systems.  This is both a great time-saver, especially if one has to respond to well-designed RFP’s on a case, as well as being an important analytical tool in probing for what really is going on a case.  In essence, it allows a user of our technology to easily vary the range of input (for example, crediting rates, tax rates, discount rates, target cash values, benefit plan investment returns, or even the relative value one puts on various goals for the above mentioned optimization technique).  This sets up the input as an “Independent Variable”.  Hundreds of system inputs can be made into an independent variable, and the architecture has no theoretical limit (though there are practical limits) in the number of simultaneous independent variables allowed.  Also, you can choose what you are interested in seeing as a “Dependent Variables” (for example, the cash value at a particular time, the solved-for premium, the effect on earnings at any point in time, the status of asset-liability matching, the IRR on the case (measured in a wide variety of appropriate ways), etc.).  A user can choose from over 700 different potential dependent variables, each at any point in time.  Again, several of these can be chosen simultaneously.   The Sensitivity Study can then be set in motion, and will provide the summary data requested on each run, as well as a terrific assortment of graphs, including pseudo 3-D graphs measuring the impact of two independent variables simultaneously.   There is simply no better tool for really understanding what is going on in a case, and it can be a way of demonstrating this to a client, or just for internal/analytical use.  It also is highly synergistic with the product profitability logic described below.  It is available now.

(11) Eleveth, we are installing product profitability logic and output in our systems, so that a few appropriately-authorized people can view “the flip side of a case” in real time.  That is, while the system will solve for all the things needed for designing the funding properly from the buyer and broker points of view, it will simultaneously show the expected net present values of the durational components of profitability from a carrier point of view.  In combination with the other tools above, carrier actuaries would be able to probe for areas of greater or lesser profitability, and may find that this allows them to be either more flexible with respect to exception requests from brokers, or possibly less so, if the analytics show that they are already hard-pressed.  But since this can be done with everyone working off of all the same data (literally thousands of assumptions for any real case), carrier staff can enjoy far better confidence in their likely performance, and may find areas where they like certain case designs much better than others, etc.   This is important logic for anyone doing real cases.

(12) Twelfth, we have developed good case management functionality at, so that we now can collect all the data for printing policy applications electronically, actually print them (and re-print them if case design changes), as well as provide for a common location to track exactly where things stand on an individual and case basis during the new business process.  When things are all ready for issue, we can do an electronic “grand submit” to get all the information necessary to issue the policies, to the policy admin system.   This will be deployed in the near future.  Whether for small deferral cases or large, this can be a significant time saver.

(13) Thirteenth, we have created a dynamic connection to an online underwriting expert system, so that for simplified issue cases in the small-case market, we can often get a case issued quickly and with appropriate rates, without medical exams or waiting for APS material, etc.  We believe this will be an important step in making the small-case COLI market work as well as we all have hoped would be possible.

(14) Fourteenth, we have continued to hone our Presentation Designer utilities, so that any broker or carrier can design and use their own form of sales presentations, just the way they want it, while also retaining the full capacity to have compliance ledgers output for exactly the same numbers.   We have had this facility for a long time, but have been surprised to find that some of our clients don’t quite know what they have here.   For dealing with Asset-Liability Matching and other methods for funding deferral plans, you can design in an entire presentation, with intimate mixing of boiler plate material and actual dynamic output from the illustrations being run, with even graphs effortlessly reflecting the real output of illustrations.  One recent example is a template that makes inputting a BOLI illustration incredibly easy.

(15) Fifteenth, we are forming a number of productive partnerships with distribution channels.  In our modern context, this has a number of efficacious effects.  One of them is tighter integration with the specific sales strategies of various channels. Another good effect is that can create seamless logins with proper authorizations from a distribution site to reach just the right sales presentations and concepts.  Another example is a distribution partnership that puts all the pieces together for a highly-integrated small-case COLI effort done in a very efficient manner.

(16) Sixteenth, we have installed some rich content on our main websites. has been complimented hundreds of times for its clear and comprehensive content for participants, administrators and brokers.  Another example is that has a terrific section of accounting for deferred compensation, with good theoretical material and lots of examples.  We also have many sample plan documents and newly-engineered decision-trees for help in choosing the design of a case.  Wizards, meta-help, and conceptual help are all being deployed.


When I took a good look at the problems and opportunities in our business area last year, it was clear that technology could do many things to make our lives more productive and our business practices “smarter” if we applied ourselves.  We at AFS set about to make all this happen, and fast.  All of the above-mentioned responses have been primarily or exclusively developed in the past year.  I am very proud that our experienced technology teams could accomplish this much in such a short time, all while continuing to service and load products for our clients.  We are still proceeding at a very rapid pace; you can expect quite a number of new announcements, features and thoughts in the near future.

From the older context from which this essay started, to the present context described above, AFS has been the market share leader in technology for the sales and in-force administration of COLI cases for more than 15 years.  In short [sic], I hope you can see that we have been and continue to be busily at work, making our technology the best possible tools for the people and companies who actually and professionally deal with the complex issues surrounding the design, funding and administration of NQ Deferral Plans.

Your thoughts and comments are welcome.

-Dan Johnson

Founder and CEO,

American Financial Systems, Inc.




No comments yet... Be the first to leave a reply!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: