3.1.4 5% Probability Test/10% Termination Test

In Rev. Rul. 77-374, the Service specified an additional test for charitable remainder annuity trusts payable for one or more lives. In addition to passing the Sec. 664(d)(1)(A) 5% minimum payout, 50% maximum payout and 10% minimum deduction interest tests, an annuity trust must pass a 5% probability test.

The 5% probability test is used to make certain that there will be a remainder for the charity. If the annuity, with adjustment for the payment frequency, exceeds the annualized earning rate under the Sec. 7520 applicable federal rate, then the trust corpus may diminish over time. In order to pay the annuity, the trustee will pay out all income and will also invade principal.

If the invasion of principal were to continue for a significant period of time, it could be possible to reduce the annuity principal to zero. The rate of principal reduction will increase each year, since the annuity earnings will decline as the principal declines and there will thus be greater invasions of principal each year to pay the annuity amount.

If under the 5% probability test with the selected applicable federal rate there is a 5% or greater probability that the donors will live long enough for the trust to be exhausted, unless the trust is reformed under Sec. 2055(e)the charitable deduction will be denied.

In Rev. Proc. 2016-42, the IRS created an alternative to the 5% Probability Test. A CRAT may avoid the 5% Probability Test if it is a qualified trust that includes specific IRS language. If the annuity trust document includes the Rev. Proc. 2016-42 language and trust corpus declines to 10% of the initial trust corpus (with discounting at the initial applicable federal rate), the annuity trust will terminate and the corpus must be distributed to the specified charities. This provision may be called the "10% Termination Test."

What is the impact of this change? Many CRATs will continue to use the 5% Probability Test for qualification. However, Rev. Proc. 2016-42 permits a donor to use the 10% Termination Test rather than the 5% Probability Test.

If use of the alternative 10% Termination Test is planned, donors should understand that the CRAT could terminate before the end of the life or lives in the trust agreement if the CRAT falls to an adjusted 10% of the initial trust corpus.

Qualified Contingency Provisions Under Rev. Proc. 2016-42

Section 5 of Rev. Proc. 2016-42 contains the sample language for the Sec. 664(f) qualified contingency that allows a CRAT to qualify under the 10% termination test. If the 10% termination test is selected, the annuity trust is valid even if it fails the 5% probability test. Section 5 contains the standard language for a qualified contingency in an inter vivos annuity trust and other variations for testamentary trusts, two life trusts and CRATs created using the IRS specified language.

Below is the qualified contingency language that must be included in a one life inter vivos annuity trust.

• "The first day of the annuity period shall be the date the property is transferred to the trust and the last day of the annuity period shall be the date of the Recipient's death or, if earlier, the date of the contingent termination. The date of the contingent termination is the date immediately preceding the payment date of any annuity payment if, after making that payment, the value of the trust corpus, when multiplied by the specified discount factor, would be less than 10 percent of the value of the initial trust corpus. The specified discount factor is equal to [1 / (1 + i)]^t, where t is the time from inception of the trust to the date of the annuity payment, expressed in years and fractions of a year, and i is the interest rate determined by the Internal Revenue Service for purposes of section 7520 of the Internal Revenue Code of 1986, as amended (section 7250 rate), that was used to determine the value of the charitable remainder at the inception of the trust. The section 7520 rate used to determine the value of the charitable remainder at the inception of the trust is the section 7520 rate in effect for [insert the month and year], which is [insert the applicable section 7520 rate]."

In a testamentary CRAT, the initial sentence phrase "the property is transferred to the trust" in this sample language must be replaced with "of my death." If the annuity is payable consecutively for two measuring lives, the initial sentence phrase "the Recipient's death" must be replaced with "the death of the survivor of the Initial Recipient and the Successor Recipient(s)."

Due to the risk that the AFR at death may be lower than when a will or trust is drafted, there is a risk that a testamentary annuity trust may fail the 5% probability test. Inclusion of the Rev. Proc. 2016-42 qualified contingency is now a solution for situations where a testamentary CRAT may otherwise fail the 5% probability test.

If the inter vivos or testamentary CRAT is created using the sample form provided in Rev. Proc. 2003-53, 2003-2 C.B. 230, or Rev. Proc. 2003-57, 2003-2 C.B. 257, respectively, the insertion of these sample provision in the appropriate locations will satisfy the requirements of a qualified contingency.

Rev. Rul. 77-374 Exact Test

In 1990 the President of Crescendo was discussing the methodology of Table S from IRS Pub. 1457 with IRS Chief Actuary Norman Greenberg. In describing the best actuarial method, Greenberg stated that optimum accuracy is obtained by completing all calculations and then rounding at the last step.

Rev. Rul. 77-374 was created before professional advisors had computers. In order to permit calculations using the existing printed IRS Publications, in the Rev. Rul. 77-374 example the ages and deferral periods were rounded to the nearest year for the funding year and the year "to completely exhaust" the fund. While this rounding method was helpful for manual calculations, it created inaccuracies in the calculation of the 5% test. Because of the initial rounding of ages and deferral period, this "approximate method" could be up to 1.3% off the exact probability. While all software vendors used this "approximate method" for over two decades and all probability calculations with this method were qualified, each company used conservative rounding methods to attempt to compensate for the potential divergence from the exact probability. In addition, with multiple lives the initial rounding error caused steadily greater divergence.

An improved and much more accurate method is now utilized in Crescendo software. It follows the actuarial rounding principle recommended by IRS Chief Actuary Greenberg and the probability formulas of Rev. Rul. 77-374. First, donor ages and years to exhaustion are calculated to eight-decimal accuracy. Using a standard actuarial interpolation method, the Lx mortality values are then determined for the exact ages for funding and potential "complete" exhaustion. The resulting 5% probability test percentage is accurate to within 0.01%. In actuarial terms, this "exact probability test" method has about 50 times greater accuracy than the prior approximate method.

The exact method permits higher confidence in calculations between 4% and 5% probability. Because a conservative probability value is still appropriate, the standard Crescendo limit is 4.5%. Therefore, the maximum annuity payout is limited to the amount that will pass the probability test at that threshold. The "Exact and Conservative" method combines exact calculation accuracy with the conservative 4.5% threshold.

The exact probability method has two benefits. First, most gift planners believe that they should offer the best possible service to donors. If a senior annuity trust donor requests the highest permitted rate, within reasonable limits the charitable trustee should offer that high rate. Second, some donors have approached two different charities and offered to create a very large remainder annuity trust with the charity paying the highest rate. When a trust funded with over one million dollars is available, the gift planner will want to know accurately the highest qualified payout rate.

Finally, some aggressive donors and their advisors may prefer a maximum payout rate with a probability between 4.5% and 4.95%. While this is clearly not recommended by Crescendo and the responsibility for selecting an aggressive rate is solely that of the donor and his or her professional advisors, the exact 5% probability method may permit an aggressive advisor and sophisticated donor to select a higher rate. If a sophisticated donor and advisor prefer to pay a higher amount from an annuity trust and move closer to the 5.0% probability limit, a staff attorney at Crescendo can explain how to do that calculation with the software.

Since Rev. Rul. 77-374 predates Sec. 7520, which became effective on May 1, 1989, using the AFR for the funding month or one of the two prior months may be an option under Sec. 7520 for purposes of the 5% probability test. Using this AFR as an annual earnings rate, the annuity payout, less the earnings rate, must be calculated each year to determine the number of years to exhaustion of the trust. Based on the applicable mortality tables for the gift date, it is then possible to determine the probability of survival to that exhaustion date.

An annuity trust may have existed for several years before the Tax Court finally determines the qualification issue. It may be possible to pass the 5% probability test if the trust indeed has sufficient earnings to show that it will not be exhausted within the time demonstrated under the tables. However, counsel should regard this as a backup strategy, since it is not a certain outcome.

The 5% test is particularly a problem for testamentary charitable remainder annuity trusts. The trust may qualify quite readily based on the AFR at the time a will or living trust with a testamentary CRAT is signed. However, if the AFR declines substantially, the annuity remainder trust may not qualify upon the death of the donor. While the Sec. 2055(e) reformation provisions do not specifically address the annuity trust 5% probability test, counsel may be able to persuade the probate court and the Service that a reformation should be permissible in this circumstance.