Often it will be obvious, from the structure of the transaction or the pricing of the interests sold, whether the transferor has transferred substantially all risks and rewards of ownership or retained substantially all risks and rewards, and there will be no need to perform any computations. In other cases, it will be necessary to compute and compare the entity’s exposure to the variability in the present value of the future net cash flows before and after the transfer. The computation and comparison is made using as the discount rate an appropriate current market interest rate. All reasonably possible variability in net cash flows is considered, with greater weight being given to those outcomes that are more likely to occur.
[IAS 39 - 22] This guidance is reminiscent of the principles in FASB Interpretation No. 46(R), Consolidation of Variable Interest Entities (revised December 2003) — an interpretation of ARB No. 51 (FIN 46R).
However, the IASB did not specify what statistic (e.g., standard deviation, variance, or mean absolute deviation as in FIN 46R) should be used to measure “variability.”
The IASB did provide a numerical example to illustrate the application of the continuing involvement approach in paragraph AG52 of IAS39. That example assumes, but does not demonstrate, the conclusion that the transferor has passed some significant risk and rewards of ownership of the transferred assets (e.g., prepayment risk) while continuing to retain some significant risks and rewards (e.g., credit risk).
The worksheet below shows one method, based on the standard deviation statistic,21 which might be used when calculations are required. The calculations require information about various possible future scenarios that is not included in the IAS 39 example, so hypothetical examples are used, while trying to fit as closely as possible to what information was given in AG52. In real life, you would probably need more than four scenarios, which should be based on real, supportable expectations and experience. With that said, the assumed scenarios are as follows:
SCENARIO 1: PREPAY IMMEDIATELY — All loans prepay all outstanding principal of $10,000 immediately after the closing of the securitization. There are no defaults and no interest has accrued since the closing date. Probability = 20%
SCENARIO 2: PREPAY IN ONE YEAR — All loans prepay all outstanding principal and accrued interest of
$1,000 on the coupon payment date, which is the first anniversary of the securitization closing. There are no defaults. Probability = 30%
SCENARIO 3: MATURE IN TWO YEARS — All loans run to their contractual maturity at the end of two years. They pay timely the $1,000 coupon interest payment due on the first anniversary of the
21 Other risk measures could also be used. For example, FIN 46R bases its risk measurements on a variation of the mean absolute deviation statistic. IAS 39 does not specify any particular statistical concept to measure risk. We chose standard deviation for this
securitization closing and all outstanding principal and accrued interest of $1,000 on the maturity date, which is the second anniversary of the securitization closing. There are no defaults. Probability = 30%
SCENARIO 4: DEFAULT IN ONE YEAR — All loans default on all outstanding principal and accrued interest on the coupon payment date, which is the first anniversary of the securitization closing. Through immediate foreclosure and sale of the collateral properties, a total of $10,746 is recovered and applied to repay principal and accrued interest, first to the senior bond, then the subordinated bond. No recovery proceeds are available to the IO strip. Probability = 20%
IAS 39 does not differentiate in regard to how the risk is allocated among the transferees. Similarly, IAS 39 does not care how or in how many tranches the transferor’s risk is retained. Therefore, the analysis can be simplified by dividing the asset cash flows for each scenario into two buckets — the total amount due to the transferor and everything else, which represents the net amount due to all the transferees taken as a group. That method is illustrated in the table below.
Probability weighted present value (using 8.5% risk-free discount rate)22
Scenario Probability Total loans Transferred-
senior Retained-
subordinate & IO
1 20% $ 2,000 $1,800 $ 200
2 30% 3,041 2,725 316
3 30% 3,079 2,747 332
4 20% 1,980 1,817 163
Total $10,100 $9,089 $1,011
Probability weighted squared deviations23
Scenario Probability Total loans
Transferred- senior
Retained- subordinate & IO
1 20% 2,000 1,584 24
2 30% 403 10 538
3 30% 8,003 1,374 2,746
4 20% 8,000 3 7,683
Variance 18,406 2,971 10,991
standard deviation (square root of Variance)
$135.67 $54.51 $104.84
Notice that the sum of the standard deviations of the transferred senior portion and the retained subordinated and IO portions add up to more than the standard deviation of the loans as a single portfolio. Try to visualize this effect with the following picture of a triangle. We use arrows for the sides of the triangle, because risk has a direction associated with it (e.g., a long or short position). The diagram
22 For example, under Scenario 2, the loan pays a total of $11,000 one year from today. Of that amount, $9,855 is paid to the senior interests and $1,145 is retained by the subordinated interests. The present values of those amounts, discounted for one year at 8.5 percent, are $10,138, $9,083, and $1,055, respectively. Weighting each by the 30 percent probability assigned to Scenario 2 gives us
$3,041, $2,725, and $316, respectively.
23 For example, the deviation of the total loan amount in Scenario 1 from the overall average is the difference between $10,000 and $10,100, which equals $100. Squaring that deviation gets us to 10,000 and weighting it by the 20 percent probability of Scenario1 yields 2,000.
shows that in total, we start at the same beginning spot and arrive at the same end point before and after the transaction. The paths are simply different.
One way to evaluate the amount of risk transferred would be to sum up the total risk exposure of the transferor after the transaction and the net risk exposure of the transferees as a group to create a new, larger denominator. Then you could divide the transferor’s total risk after the transaction by that sum.
But IAS 39 cares about the proportion of the risk of ownership of the assets that has been transferred or retained, not the absolute amount of risk exposure of the transferor or transferee.
This distinction is important, because risk does not add or subtract in simple, mathematical ways. The portfolio diversification effect of the loans as a
group reduced the total risk of the portfolio to something less than the arithmetic sum of the risks of each loan. The subordination structure of the transaction acts in a similar fashion, but in the opposite direction. Rather than reducing risk through diversification, it increases risk by introducing to both the transferor and transferees as a group an equal and offsetting new risk that is completely uncorrelated with the aggregate asset portfolio risk. In essence, the transferor and
transferee have made a small side bet using a portion of the asset portfolio as the amount of the wager.
Visualize this by breaking the previous diagram into two parts.
Again, each diagram shows the same beginning and end to represent the total risk to the transferor and the net risk to the transferees as a group. It also reveals the portion of the risk of the assets retained and transferred. If we knew how the asset portfolio risk correlated with the total risk to the transferor or the net risk to the transferees as a group, it would be a simple matter to measure the numerical amount of the asset risk retained and transferred. The calculations of all the required correlations (statistics) would quickly become tedious, however. Fortunately, some high school trigonometry allows us to use the previously calculated variance statistics to compute24 the percentage of risk transferred (28 percent) and the percentage of risk retained (72 percent).
IAS 39 does not establish any bright-line guidance on the cutoff levels for what represents “substantially all.” If the transferor had adopted something in the 80 percent to 90 percent range as their level for substantially all, then in this hypothetical case, the conclusion would be that the transferor had neither transferred substantially all of the portfolio risk nor retained substantially all of it.
Portfolio = $135.67
Transferor = $104.84 Transfer
ees = $54.51
Transferees net = $54.51 Side bet
24 The formulas work out to:
Risk Transfer % = 50% + [Var(Transferred) - Var(Retained)] ÷ [2×Var(Portfolio)]
= 50% + [2,971 - 10,991] ÷ [2×18,406] = .5 - (8,020 ÷ 36,812)
= 28%
Risk Retention % = 50% + [Var(Retained) - Var(Transferred)] ÷ [2×Var(Portfolio)]
= 50% + [10,991 - 2,971] ÷ [2×18,406] = .5 + (8,020 ÷ 36,812)
What changes are on the horizon for International Accounting Standards?
The IASB is working towards revamping guidance on both of the two major securitization accounting issues — consolidation and derecognition. These changes are likely to have a major impact on U.S.
securitization accounting as well. Even though the SEC has not published an updated roadmap for allowing U.S. registrants to use international accounting standards for securities filings, the FASB has committed to join with the IASB to develop standards that are as consistent as possible.
Currently, the IASB is finalizing its standard on consolidation. IASB’s approach is broadly similar to the FAS 167 approach by considering both control and financial interest elements. There are a couple of notable differences, however. The IASB standard will not limit itself to structured finance vehicles (or SPEs, VIEs, or any other name you might choose). Instead, it will address the full range of potential subsidiary company situations. The IASB also has a slightly different idea about what constitutes control, particularly when kick-out rights are involved.
The FASB is also expected to issue an Exposure Draft of an amendment to FAS 167 that will propose conforming the U.S. consolidation guidance to the IASB standard. Look for that as vacation reading for summer 2010.
The IASB plans to issue its final standard during the third quarter of 2010 after they have seen the public comments on the FASB draft. Smart money bets on an effective date in 2013.
The IASB had an interesting experience in its project on derecognition principles. In March 2009, the IASB issued an Exposure Draft suggesting changes to the current IAS 39 framework described above. In response, almost nobody (well, only 13 percent) of the comment writers favored their proposal. The status quo polled better at 23 percent. The majority of respondents favored an alternative approach contained in a short appendix explaining why five IASB members voted against the Exposure Draft to begin with.
The IASB cares more about the quality of the ideas expressed in comments rather than just the quantity of cards and letters it gets. In this case, the tide of constituent opinion carried with it many valid objections to the proposed approach and many thoughtful reasons supporting the alternative views. So the IASB decided to regroup and adopt the alternative view going forward.
Now the IASB plans to develop a derecognition standard based on control that differentiates between transfers where the transferor retains a proportional versus disproportional share of the transferred asset.
If the retained interest is disproportionate then the transferred asset is fully derecognized and a new asset is measured at fair value. If the retained interest is proportional then the transferor will treat the retained interest as part of the transferred asset and therefore will derecognize only the part not retained.
There are still details to be worked out by the IASB, including things like making sure everybody agrees on how to identify, describe, and value the new assets and liabilities. Some constituents may also not be happy with how this approach treats repo and securities lending transactions. For these reasons, it would not be surprising for the IASB to issue another Exposure Draft, even though it might not technically be required by their due process protocols. Another public comment process would also offer the FASB an opportunity to coordinate consideration of parallel changes to U.S. guidance. If done expeditiously, there would probably be time to produce a final derecognition standard that could be implemented in 2013 in conjunction with a