Memorandum

DATE:March 10, 2010

TO:IASWG and interested parties

FROM:Rob Esson

RE:Risk Margins: An overview and a simplified method of calculation

SUMMARY RECOMMENDATION

  1. There is widespread agreement that the measurement approach for insurance contracts should be based on two initial specific key building blocks, with the details of any other building blocks still to be determined. The first building block would be unbiased probability weighted expected cash flows arising from the contract (the Current Estimate). The second building block is the time value of money (i.e. discounting). The final building block(s) is the risk margin or possibly the risk margin and residual margin, and this is the area where views diverge and is the subject of this paper.
  2. NAIC staff was asked to put forward a methodology utilizing one margin and requests your input on a method of calculating the margin as discussed below. This method utilizes an approach that maintains the entire risk margin in one building block, as opposed to splitting the margin into two separate building blocks (risk margin and residual margin). Under the alternative two margin method, the residual margin would be the amount needed in excess of the first three building blocks to eliminate profit on inception.
  3. There is certainly disquiet by some in the US regarding a split of margins into both a risk and a residual margin. This is for two basic reasons: firstly, problems over the ability to measure a split of margins with sufficient precision; and secondly a feeling that there may be a greater opportunity to manipulate income recognition inappropriately using two margins rather than one.
  4. The IAIS avoids the controversy over the naming of the margin by simply referring to it as a margin over current estimate or “MOCE”. This is because there is a lot of agreement that some form of margin is required but there are differing views on how best to describe it and more importantly on how to measure it. At minimum, it is agreed that the MOCE reflects a measurement of the uncertainty inherent in the calculation of the discounted cash flows.
  5. The method NAIC staff requests your input on is as follows. This method has elements of all the requisite factors desired by the Boards, and determines the margin by applying to the Current Estimate the margin implied by the charged premium.

NWP = Net Written Premium, and

PVt(cf) = present value of future cash flows at time t

then (MOCE) is determined by applying to PVt(cf) the ratio { NWP - PV0(cf) } / PV0(cf).

  1. An unearned premium reserve (UPR) method could still be utilized for the pre-claims liability.
  2. A minimum margin would likely need to apply to each line of business to prevent abuse. Regulators could specify this minimum margin.
  3. The advantages of this approach are:
  1. it is simple to apply and to audit
  2. it uses premium as an observable and objective measure for determining the MOCE
  3. it is consistent with the accounting standard setters’ wish to eliminate profit on inception
  4. it would run margins off over the life of the business. This is appropriate as, at derecognition, all liabilities will have been settled and no remaining uncertainty will exist
  1. Some disadvantages of this approach, and ways of addressing them, would be:

(i)Companies could lower margins across the life of the business by manipulating the cash flow calculations to increase the ratio PV0(cf) / NWP.

While the suggested answer is rules based, it is designed to prevent abuse. We believe that the fear of manipulation could be addressed by requiring some absolute minimum margins by line of business. It is likely that insurance regulators would require this anyway to prevent abuse.

(ii)Companies with higher premiums, possibly because they are better capitalised, will have higher margins.

This is an unavoidable consequence of the method, although any potential extra margin will be released over the lifetime of the contract. However, the counter-argument implies that there is sufficient precision in the measurement of the MOCE absent a calibration to premium as to enable a true comparison between companies, a proposition that lacks significant empirical evidence. Potential problems with this issue could be addressed by requiring separate disclosure of the current estimate and the margin, together with loss triangulation data.

(iii)Different companies reserve levels might not be comparable.

As noted in item (ii), the actual comparability between companies may not be nearly as precise as models and theory pretend. Requiring separate disclosure of the current estimate and the margin, together with loss triangulation data would alleviate issues.

(iv)Margins in premium will become public.

Some companies may be concerned that their MOCE will be used publicly to disadvantage them. While there may possibly be some truth to this concern, the alternative method of using a risk margin and separate residual margin is likely to cause even more problems if this were to be the case. Residual margins might well be used to make regulatory rate or tax decisions in various jurisdictions.

  1. Attached is an illustration, including a graph, of the risk margin and run-off for a short tail P&C policy where all liabilities have been run off by year 4. In particular, it shows that the margin increases proportionally when the estimated cash flows develop in an adverse manner, but that this continues to run off over time.
  2. It would be useful to obtain input from regulators and interested parties on the practicality of this approach. In particular, the time for long arguments about the pros and cons of complex theoretical alternatives is likely passed: it is now time to find – urgently – practical and pragmatic approaches capable of implementation.
  3. Failure to enunciate an alternative approach – if so desired – will likely result in the four building block method being adopted for IFRS, and maybe for US GAAP.

ILLUSTRATION OF RISK MARGIN CALCULATION & RUNOFF

Day 1 / YE 1 / YE 2 / YE 3 / YE 4
Net Written Premium / 100 / 100 / 100 / 100 / 100
PV of Future Cash Outflows / 80 / 90 / 85 / 82 / 82
Risk Margin Before Amortization / 20 / 22.5 / 21.25 / 20.5 / 20.5
Risk Margin % / 0.25
Expected Cash Flows by Years (Not PV for simplicity)
Year 1 (actuals in green) / 50 / 55 / 55 / 55 / 55
Year 2 / 15 / 17 / 14 / 14 / 14
Year 3 / 10 / 12 / 11 / 9 / 9
Year 4 / 5 / 6 / 5 / 4 / 4
Total / 80 / 90 / 85 / 82 / 82
Cumulative Amortization of Risk Margin / 13.75 / 17.25 / 19.5 / 20.5
Current Year Amortization / 13.75 / 3.5 / 2.25 / 1
Risk Margin After Amortization / 20 / 8.75 / 4 / 1 / 0
Insurance Liabilities / 100 / 43.75 / 20 / 5 / 0

BACKGROUND INFORMATION

  1. The author recognizes the help and input from Rick Shaw, the former Chief Actuary of the Bermuda Monetary Authority, in the drafting of this memo.
  2. The joint project by the IASB and the FASB on insurance contracts is approaching the stage at which an exposure draft will be issued. At present, the timing of the exposure draft is expected to be May 2010 although it may slip to June 2010.
  3. The insurance contracts project itself has been underway in one form or another at the IASB or its predecessor since 1997 – we are now in the 13thyear. If it were easy, it would have been completed a long time ago.
  4. There is widespread agreement among supervisors, industry, and the accounting standard setters that the measurement approach for insurance contracts should be based on building blocks. There is also widespread agreement that the first building block would be unbiased probability weighted expected cash flows arising from the contract. The second building block, being the time value of money, is also widely agreed although some would prefer an alternative approach. It is however the next building block or more accurately potentially the next two building blocks where views diverge the most.
  5. The next building block is either a composite margin, a risk margin, or a set of two building blocks being a risk margin and a residual margin. To the extent that it would exist, the residual margin would be the amount needed in excess of the first three building blocks to eliminate profit on inception.
  6. Discussions at the IASB and FASB have converged to consideration of two models, both of which will be exposed in the forthcoming Exposure Draft.
  7. At present, a narrow (8:7) majority of the IASB, and a remarkably narrow 3:2 majority of the FASB support the four building block approach using a residual margin. The FASB vote is remarkably narrow in that Chairman Bob Herz switched his vote in order to maintain a joint project rather than diverging from the IASB view.
  8. According to the IASB, the two margins within the four building blocks consist of:

•a risk adjustment for the effects of uncertainty about the amount and timing of future cash flows; and

•an amount that eliminates any gain at inception of the contract (a residual margin). The residual margin cannot be negative

  1. Additionally, “The risk adjustment should be the amount the insurer requires for bearing the uncertainty that arises from having to fulfil the net obligation arising from an insurance contract. The risk adjustment should be updated (remeasured) each reporting period.”
  2. The IASB has not, so far, defined how this risk adjustment (also known as risk margin) should be calculated, although it is clear that the thinking of many of the staff is influenced by the European Solvency II considerations. The cost of capital method seems to be a significant possibility.
  3. The IASB has also stated that “the residual margin should not be adjusted in subsequent reporting periods for changes in estimates. Staff will develop specific guidance on how the residual margin should be released to profit or loss over time.”
  4. At the time of the Discussion Paper, the subsequent measurement and run-off of margins was not addressed.
  5. The purpose of this paper is to suggest a simplification in the calculation and run-off of the MOCE, which, while not perfect, nonetheless would be practical and auditable.
  6. The IAIS first proposed a set of criteria against which a measurement of the MOCE should be judged. Subsequently, both the IAA and IASB have reiterated the same points. These were:

“Irrespective of the particular (calculation) methodology chosen, acceptable methods should reflect the inherent uncertainty in the expected future cash flows and would be expected to exhibit the following characteristics:

  1. The less that is known about the current estimate and its trend; the higher the risk margins should be
  2. Risks with low frequency and high severity will have higher risk margins than risks with high frequency and low severity
  3. For similar risks, contracts that persist over a longer timeframe will have higher risk margins than those of shorter duration
  4. Risks with a wide probability distribution will have higher risk margins than those risks with a narrower distribution
  5. To the extent that emerging experience reduces uncertainty, risk margins will decrease, and vice versa.”
  1. The suggested method will, perforce, defer some of these criteria into the original pricing of the policy, but this may not be that unreasonable: insurance companies usually do endeavour to price in a manner that exactly takes these risks into account. Substitute the words “premiums at inception or pricing” for risk margins in the five criteria, and they remain valid absent basic incompetence or error.
  2. The other reason that this may not be unreasonable is that the Boards are imposing a ‘no profit on inception’ constraint on the margin or margins, and this forces the initial margin to take into account the pricing of the policy.
  3. As noted above, there are currently a number of methods being used or proposed to determine the MOCE, including:

(i)percentile approach (Australia and Singapore non-life)

(ii) cost of capital approach (not in use, but proposed by Europe)

(iii) varying input assumptions to reflect a potential adverse scenario (Canada)

(iv) time value of money (everywhere else for non-life)

(v) Margin on Services (Australia life) - see below for more detail.

  1. There is much literature discussing the advantages and disadvantages of these methods, not the least of which is an IAA paper discussed below.
  2. About five years ago, I stated that in my view one of the most critical issues to solve in the insurance contracts project would be d(Margin)/dt – the time derivative of the margin, or put another way, the shape of the decay of a margin over time.
  3. If we were to calibrate a margin to premium on inception, then there are two things that we know about the margin. Firstly we know the amount of the margin at inception and secondly we know that the margin should be zero when all cash flows under the policy have been fulfilled. What happens between these two points is less clear, and indeed there is not unanimity that it is appropriate to calibrate the margin to the premium at inception.
  4. Consequently, it is possible that, if we do not calibrate the margin or margins to premium at inception, the only thing that we know about the margin for sure is that it should be zero when all policies liabilities have finally run off.
  5. In 2005 and 2006, the IAIS worked cooperatively with the International Actuarial Association (IAA) on certain input to the IASB, culminating in the realization that more work was needed on how to measure current estimates and risk margins. As a result the IAIS asked the IAA to undertake a study on these issues culminating in the publication of the IAA paper “Measurement of Liabilities for Insurance Contracts: Current Estimates and Risk Margins”, published in April 2009.

VIEWS ON FOUR BUILDING BLOCK APPROACH

  1. There is support at the IAIS for the four building block approach from a number of major jurisdictions. In particular, Australia has experience with an equivalent approach, although it allows remeasurement of residual margins in such a way that adjustments to current estimates and risk margins that do not relate to financial market observables are added back or subtracted from the residual margin, which is then amortized. This has the interesting effect of dampening potential management abuse and income manipulation: changes do not go through the profit and loss account immediately, but are absorbed by the residual margin such that only the current year’s amortization goes to profit and loss.

BASIS FOR CONCLUSION

  1. The NAIC has supported a limited letter to the IASB from the IAIS stating that “a residual margin arises in the accounting for an insurance contract when a reliable measurement of the expected cash flows and the risk margin results in a value less than the premium on the policy, and therefore represents what would otherwise be a profit on inception. However, based on the Board’s tentative decision, any change in expected future cash flows or a risk margin as a result of changes in (at least) future non-financial assumptions would result in a profit or loss being recognised even when there is a remaining residual margin. To the extent that the cash flows and risk margin are reliable at inception, an equally reliable adjustment of the cash flows and risk margin should adjust any remaining portion of the residual margin. For adverse adjustments, this should occur until the residual margin is exhausted, after which any further adverse adjustments should be recognised directly in income.” [Emphasis added]
  2. The NAIC position was concerned not to imply that cash flows and risk margins can in fact be measured to a sufficiently high degree of precision as to enable a reliable measurement of a residual margin. However, to the extent that it might exist, it was felt that the NAIC could support the IAIS letter. The letter also stated specifically that “it should not be interpreted as implying support by all IAIS jurisdictions” for the four building block method with residual margins.
  3. The alternative methodology considered by the Boards would not endeavor to split margins into components, calibrating the total margin to premium. This total margin is referred to as a composite margin.
  4. At the time of the decision by the IASB over the two models, the IASB staff announced in their papers that there was no driver for the remeasurement of a composite margin, and that it did not have an attribute. It seemed that, as a result of this statement, some Board members changed their likely vote due to discomfort with this assertion.
  5. Nonetheless, there remains considerable support for an approach that calibrates the margin in a manner that is computationally equivalent to the composite margin approach. This paper proposes a methodology equivalent to a remeasurement of a composite margin.
  6. As noted above, the IAA completed a comprehensive study of risk margins. Within that study, the IAA came to the conclusion that the most risk sensitive method of calculating risk margins is likely to be the cost of capital method. However the paper noted that this was also very challenging to implement in practice in a consistent manner.In particular, in order to calculate a risk margin using the cost of capital approach, one would need to identify the capital required at the policy level or more likely at the portfolio or line of business level. It is notable that the identification of the amount of capital required by line of business is fraught with difficulty. The actual amount tends to be quite uncertain and consequently whatever the cost rate may be (for example European Solvency II is likely to recommend a 6% rate), it is incontestable that the resultant margin will be at least as uncertain. This fundamental tenet of mathematics is sometimes glossed over by aficionados of the cost of capital method.
  7. One consequence of these uncertainties is that the resultant margin using a cost of capital methodology will be difficult to audit.
  8. It is for this reason – intrinsic uncertainty - that the US supported the idea of the rebuttable presumption approach to measurement of the risk margin at the time of the IASB & FASB Discussion Paper. While the response was couched in terms of current exit value, it was modified to allow entity specific portfolio evaluation and elimination of own credit standing. It was therefore quite close to what is now referred to as fulfilment value. In the response, it stated that US state insurance regulators believed that the appropriate calibration in practice would be effectively equivalent to the rebuttable presumption of calibrating the risk margin to the premium at inception.
  9. Computationally, this approach is the same at inception as current fulfilment value with a composite margin. However, there is a subtle and important difference: the approach takes the view that the amount that would be composite margin in the fulfilment methodis the risk margin. It does not deny that insurers intend to make a profit, nor that all margins, by whatever name they go, will ultimately become profit if the future works out exactly per the discounted expected cash flows. It does not seek to split the risk margin into components, and consequently, the suggestion by some that a composite margin cannot be remeasured becomes irrelevant.

Page 1 of 7