This Treasury Decision modifies existing regulations, including the regulations at §1.401(a)-1 which generally require a pension plan to be maintained primarily to provide systematically for the payment of definitely determinable benefits after retirement. These regulations provide two exceptions to this rule. First, they clarify that a pension plan is permitted to commence payment of retirement benefits to a participant after the participant has attained normal retirement age. The regulations also provide rules on how low a plan’s normal retirement age is permitted to be and include a related exception to the anti-cutback rules of section 411(d)(6) to allow conforming amendments during a transitional period. Second, the regulations reflect the provisions of new section 401(a)(36).

A. In General

These regulations adopt the rule of the proposed regulations under which a pension plan (a defined benefit plan or money purchase pension plan) is permitted to pay benefits upon an employee’s attainment of normal retirement age, even if the employee has not yet had a severance from employment with the employer maintaining the plan. Comments generally supported the inclusion of this rule as reflecting existing practice among some pension plans, based on an example in Rev. Rul. 71-24.

These regulations also include rules restricting a plan’s normal retirement age. The proposed regulations would have provided that a plan’s normal retirement age could not be set so low as to be a subterfuge to avoid the requirements of section 401(a), and, accordingly, normal retirement age could not be earlier than the earliest age that is reasonably representative of a typical retirement age for the covered workforce.^[2] Some comments expressed concern about the specifics of this rule, including concern about how it might be applied in various circumstances, and suggested that the regulations contain a safe harbor for which there would be no need for a demonstration of the typical retirement age for the covered workforce.

These final regulations modify the proposed regulations to replace the subterfuge standard with a requirement that the normal retirement age under a plan be an age that is not earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed. To address comments about the need for a safe harbor age, these regulations provide that a normal retirement age of at least age 62 is deemed to be not earlier than the typical retirement age for the industry in which the covered workforce is employed. Thus, a plan satisfies this safe harbor if its normal retirement age is age 62, or if its normal retirement age is the later of age 62 or another specified date, such as the later of age 62 or the fifth anniversary of plan participation. However, a plan that is subject to section 411 cannot provide for a normal retirement age that is later than the later of the time the participant attains age 65 or the fifth anniversary of the time the participant commenced participation in the plan. See section 411(a)(8)(B).

If a plan’s normal retirement age is earlier than age 62, the determination of whether the age is not earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed is based on all of the relevant facts and circumstances. If the normal retirement age is between ages 55 and 62, then it is generally expected that a good faith determination of the typical retirement age for the industry in which the covered workforce is employed that is made by the employer (or, in the case of a multiemployer plan, made by the trustees) will be given deference, assuming that the determination is reasonable under the facts and circumstances. However, a normal retirement age that is lower than age 55 is presumed to be earlier than the earliest age that is reasonably representative of the typical retirement age for the industry of the relevant covered workforce absent facts and circumstances that demonstrate otherwise to the Commissioner.

In the case of a plan where substantially all of the participants in the plan are qualified public safety employees (within the meaning of section 72(t)(10)(B), as added by section 828 of PPA ’06), a normal retirement age of age 50 or later is deemed not to be earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed. Under section 72(t)(10)(B), a qualified public safety employee means any employee of a State or political subdivision of a State who provides police protection, firefighting services, or emergency medical services for any area within the jurisdiction of such State or political subdivision.

B. Section 411(d)(6) Relief

These regulations include an amendment to the existing regulations under section 411(d)(6) to permit a plan to be amended during a transition period to conform to the rules concerning normal retirement age. Thus, a plan amendment that changes the normal retirement age under the plan to a later normal retirement age (pursuant to these regulations) does not violate section 411(d)(6) merely because the amendment eliminates a right to an in-service distribution prior to the amended normal retirement age. However, this rule does not provide any other relief. For example, this rule does not permit the amendment to reduce benefits in some other manner that fails to satisfy section 411(d)(6). Neither does the rule provide relief under section 411(a)(9) (requiring that the normal retirement benefit not be less than the greater of any early retirement benefit payable under the plan or the benefit under the plan commencing at normal retirement age), section 411(a)(10) (if the amendment changes the plan’s vesting rules), or section 4980F (or section 204(h), the parallel provision of ERISA) (relating to amendments that reduce the rate of future benefit accrual). See also Rev. Rul. 81-210, 1981-2 C.B. 89 (see §601.601(d)(2)(ii)(b)). An example is included to illustrate this rule.

* * * * *

(b) * * *

(1) * * *

(i) In order for a pension plan to be a qualified plan under section 401(a), the plan must be established and maintained by an employer primarily to provide systematically for the payment of definitely determinable benefits to its employees over a period of years, usually for life, after retirement or attainment of normal retirement age (subject to paragraph (b)(2) of this section). A plan does not fail to satisfy this paragraph (b)(1)(i) merely because the plan provides, in accordance with section 401(a)(36), that a distribution may be made from the plan to an employee who has attained age 62 and who is not separated from employment at the time of such distribution.

* * * * *

(2) Normal retirement age—(i) General rule. The normal retirement age under a plan must be an age that is not earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed.

(ii) Age 62 safe harbor. A normal retirement age under a plan that is age 62 or later is deemed to be not earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed.

(iii) Age 55 to age 62. In the case of a normal retirement age that is not earlier than age 55 and is earlier than age 62, whether the age is not earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed is based on all of the relevant facts and circumstances.

(iv) Under age 55. A normal retirement age that is lower than age 55 is presumed to be earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed, unless the Commissioner determines that under the facts and circumstances the normal retirement age is not earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed.

(v) Age 50 safe harbor for qualified public safety employees. A normal retirement age under a plan that is age 50 or later is deemed to be not earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed if substantially all of the participants in the plan are qualified public safety employees (within the meaning of section 72(t)(10)(B)).

(3) Benefit distribution prior to retirement. For purposes of paragraph (b)(1)(i) of this section, retirement does not include a mere reduction in the number of hours that an employee works. Accordingly, benefits may not be distributed prior to normal retirement age solely due to a reduction in the number of hours that an employee works.

(4) Effective date. Except as otherwise provided in this paragraph (b)(4), paragraphs (b)(2) and (3) of this section are effective May 22, 2007. In the case of a governmental plan (as defined in section 414(d)), paragraphs (b)(2) and (3) of this section are effective for plan years beginning on or after January 1, 2009. In the case of a plan maintained pursuant to one or more collective bargaining agreements that have been ratified and are in effect on May 22, 2007, paragraphs (b)(2) and (3) of this section do not apply before the first plan year that begins after the last of such agreements terminate determined without regard to any extension thereof (or, if earlier, May 22, 2010). See §1.411(d)-4, A-12, for a special transition rule in the case of a plan amendment that increases a plan’s normal retirement age pursuant to paragraph (b)(2) of this section.

Par. 3. Section 1.411(d)-4 is amended by adding Q&A-12 as follows:

* * * * *

Q-12. Is there a transition period during which a plan is permitted to eliminate a right to in-service distributions in connection with an amendment to ensure that the plan’s normal retirement age satisfies the requirements of §1.401(a)-1(b)(2)?

A-12. (a) In general. A plan amendment that changes the normal retirement age under the plan to a later normal retirement age pursuant to §1.401(a)-1(b)(2) does not violate section 411(d)(6) merely because it eliminates a right to an in-service distribution prior to the amended normal retirement age. However, this paragraph does not provide relief from any other applicable requirements; for example, this relief does not permit the amendment to violate section 411(a)(9) (requiring that the normal retirement benefit not be less than the greater of any early retirement benefit payable under the plan or the benefit under the plan commencing at normal retirement age), section 411(a)(10) (if the amendment changes the plan’s vesting rules), section 411(d)(6) (other than elimination of the right to an in-service distribution prior to the amended normal retirement age), or section 4980F (relating to an amendment that reduces the rate of future benefit accrual). This paragraph only applies to a plan amendment that is adopted after May 22, 2007, and on or before the last day of the applicable remedial amendment period under §1.401(b)-1 with respect to the requirements of §1.401(a)-1(b)(2) and (3).

(b) Example. The following example illustrates the application of this section:

(i) Facts. (A) Plan A is a defined benefit plan intended to be qualified under section 401(a). Plan A is maintained by a calendar year taxpayer and has a normal retirement age that is age 45. For employees who cease employment before normal retirement age with a vested benefit, Plan A permits benefits to commence at any date after the attainment of normal retirement age through attainment of age 70¹/₂ and provides for benefits to be actuarially increased to the extent they commence after normal retirement age. For employees who continue employment after attainment of normal retirement age, Plan A provides for benefits to continue to accrue and permits benefits to commence at any time, with an actuarial increase in benefits to apply to the extent benefits do not commence after normal retirement age. Age 45 is an age that is earlier than the earliest age that is reasonably representative of the typical retirement age for the industry in which the covered workforce is employed.

(B) On February 18, 2008, Plan A is amended, effective May 22, 2007, to change its normal retirement age to the later of age 65 or the fifth anniversary of participation in the plan. The amendment provides full vesting for any participating employee who is employed on May 21, 2007, and who terminates employment on or after attaining age 45. The amendment provides employees who cease employment before the revised normal retirement age and who are entitled to a vested benefit with the right to be able to commence benefits at any date from age 45 to age 70¹/₂. The plan amendment also revises the plan’s benefit accrual formula so that the benefit for prior service (payable commencing at the revised normal retirement age or any other age after age 45) is not less than would have applied under the plan’s formula before the amendment (also payable commencing at the corresponding dates), based on the benefit accrued on May 21, 2007, and provides for service thereafter to have the same rate of future benefit accrual. Thus, for any participant employed on May 21, 2007, with respect to benefits accrued for service after May 21, 2007, the amount payable under the plan (as amended) at any benefit commencement date after age 45 is the same amount that would have been payable at that benefit commencement date under the plan prior to amendment. The plan amendment also eliminates the right to an in-service distribution between age 45 and the revised normal retirement age. Plan A has been operated since May 22, 2007, in conformity with the amendment adopted on February 18, 2008.

(ii) Conclusion. The plan amendment does not violate section 411(d)(6). Although the amendment eliminates the right to commence benefits in-service between age 45 and the revised normal retirement age, the amendment is made before the last day of the remedial amendment period applicable to the plan under §1.401(b)-1 with respect to the requirements of §1.401(a)-1(b)(2) and (3), and therefore the amendment is permitted under paragraph (a) of this A-12. Further, the amendment does not result in a reduction in any benefit for service after May 22, 2007. Thus, the amendment does not result in a reduction in any benefit for future service, and advance notice of a significant reduction in the rate of future benefit accrual is not required under section 4980F.

Approved May 9, 2007.

Under the proposed regulations, the custodial parent is the parent with whom the child resides for the greater number of nights during the calendar year. The noncustodial parent is the parent who is not the custodial parent. The proposed regulations further provide that, if a child is temporarily absent from a parent’s home for a night, the child is treated as residing with the parent with whom the child would have resided for the night. However, if the child resides with neither parent for a night, for example because another party is entitled to custody of the child for that night, the child is treated as not residing with either parent for that night. Comments are requested specifically on alternative methods of allocating nights when a child resides with neither parent and whether nights residing with neither parent should not be allocated to either parent. The proposed regulations provide a tie-breaking rule that, if a child resides with each parent for an equal number of nights during the calendar year, the parent with the higher adjusted gross income for the calendar year is treated as the custodial parent. Cf. section 152(c)(4)(B).

Sections 151 and 152, not state law, determine whether a divorced or separated parent may claim an exemption for a child for Federal income tax purposes. A state court order or decree does not operate to allocate the federal exemption between parents.

Section 152(e)(2) provides that a custodial parent may release a claim to an exemption for a child by signing a written declaration (in such form and manner as the Secretary may prescribe by regulations) that he or she will not claim the child as a dependent. The noncustodial parent must attach the written declaration to the tax return to claim a dependency exemption for the child. Section 1.152-4T, Q&A-3, states that the written declaration may be made on a form developed by the IRS. Form 8332, Release of Claim to Exemption for Child of Divorced or Separated Parents, currently is used for this purpose. The temporary regulations further provide that any declaration not made on that form must conform to the substance of the form. Section 1.152-4T, Q&A-4, states that a claim to an exemption may be released for a single year, for a number of years, or for all future years, as specified in the declaration.

The proposed regulations incorporate these rules and further provide that a written declaration must include an unconditional statement that the custodial parent will not claim the child as a dependent for the specified year or years. A statement is unconditional if it does not expressly condition the custodial parent’s waiver of the right to claim the child as a dependent on the noncustodial parent’s meeting of an obligation such as the payment of support. The written declaration must specify the year or years for which the release is effective. A written declaration that does not specify a year or years has no effect. A written declaration that specifies all future years is treated as specifying the first taxable year after the taxable year the release is executed and all subsequent taxable years. A court order or decree may not serve as the written declaration required by section 152(e)(2).

The proposed regulations provide that a custodial parent who released the right to claim a child may revoke the release for future taxable years by providing written notice of the revocation to the other parent. The revocation may be made on a form designated by the IRS, such as Form 8332, which may be revised for this purpose, or by a written declaration that conforms to the substance of that form, whether or not the release was made on the form, and must specify the year or years for which the revocation is effective. A revocation that does not specify a year or years has no effect. A revocation that specifies all future years is treated as specifying the first taxable year after the taxable year the revocation is executed and all subsequent taxable years. The revocation may be effective no earlier than the taxable year that begins in the first calendar year after the calendar year in which the parent revoking the release provides notice of the revocation to the other parent. The parent revoking the release must attach the original or a copy of the revocation to the parent’s tax return for any taxable year the parent claims the exemption as a result of the revocation, and keep a copy of the revocation and evidence of delivery of written notice of revocation to the noncustodial parent.

In King v. Commissioner, 121 T.C. 24 (2003), the United States Tax Court decided that section 152(e) applies to parents who had never married each other. The parents lived apart for the years at issue, and each had claimed a dependency deduction for the same child. In concluding that the Form 8332 executed by the custodial parent released her claim to the deduction, the court determined that section 152(e)(1)(A)(iii), which refers to parents who “live apart at all times during the last 6 months of the calendar year,” encompasses both married parents and parents who never married each other. The proposed regulations follow the decision in King v. Commissioner.

The regulations are proposed to apply to taxable years beginning after the date the regulations are published as final regulations in the Federal Register.

(a) In general. A taxpayer may claim a dependency deduction for a child (as defined in section 152(f)(1)) only if the child is the qualifying child of the taxpayer under section 152(c) or the qualifying relative of the taxpayer under section 152(d). Section 152(c)(4)(B) provides that a child who is claimed as a qualifying child by parents who do not file a joint return together is the qualifying child of the parent with whom the child resided for a longer period of time during the taxable year or, if the child resided with both parents for an equal period of time, of the parent with the higher adjusted gross income. However, a child is treated as the qualifying child or qualifying relative of the noncustodial parent if the custodial parent releases the claim to the exemption under section 152(e) and this section.

(b) Release of claim by custodial parent—(1) In general. Under section 152(e)(1), notwithstanding section 152(c)(1)(B), (c)(4)(B), and (d)(1)(C), a child is treated as the qualifying child or qualifying relative of the noncustodial parent (as defined in paragraph (c) of this section) if the requirements of paragraphs (b)(2) and (b)(3) of this section are met.

(2) Support, custody, and parental status. The requirements of this paragraph (b)(2) are met if the parents of the child provide over one-half of the child’s support for the calendar year, the child is in the custody of one or both parents for more than one-half of the calendar year, and the parents—

(i) Are divorced or legally separated under a decree of divorce or separate maintenance;

(ii) Are separated under a written separation agreement; or

(iii) Live apart at all times during the last 6 months of the calendar year whether or not they are or were married.

(3) Release of claim to child. The requirements of this paragraph (b)(3) are met if—

(i) The custodial parent signs a written declaration that the custodial parent will not claim the child as a dependent for the taxable year beginning in that calendar year and the noncustodial parent attaches the declaration to the noncustodial parent’s return for the taxable year; or

(ii) A qualified pre-1985 instrument, as defined in section 152(e)(3)(B), effective for the taxable year beginning in that calendar year, provides that the noncustodial parent is entitled to the dependency exemption for the child and the noncustodial parent provides at least $600 for the support of the child during the calendar year.

(c) Custodial parent—(1) In general. The custodial parent is the parent with whom the child resides for a greater number of nights during the calendar year, and the noncustodial parent is the parent who is not the custodial parent.

(2) Absences. For purposes of this paragraph (c), when a child resides with neither parent for a night, the child is treated as residing with the parent with whom the child would have resided for the night but for the absence. However, if the child would have resided with neither parent for a night during an absence (for example, because a court awarded custody of the child to a third party for the period of absence), the child is treated as residing with neither parent for the night of the absence.

(3) Special rule for equal number of nights. If a child is in the custody of one or both parents for more than one-half of the calendar year and the child resides with each parent for an equal number of nights during the calendar year, the parent with the higher adjusted gross income for the calendar year is treated as the custodial parent.

(d) Written declaration—(1) Form of declaration—(i) In general. The written declaration under paragraph (b)(3)(i) of this section must constitute the custodial parent’s unconditional release of the parent’s claim to the child as a dependent for the year or years for which the declaration is effective. A declaration is unconditional if it does not expressly condition the custodial parent’s release of the right to claim the child as a dependent on the noncustodial parent’s meeting of an obligation such as the payment of support. A written declaration must name the noncustodial parent to whom the exemption is released. A written declaration must specify the year or years for which it is effective. A written declaration that does not specify a year or years has no effect. A written declaration that specifies all future years is treated as specifying the first taxable year after the taxable year of execution and all subsequent taxable years. A court order or decree may not serve as the written declaration.

(ii) Form designated by IRS. A written declaration may be made on a form designated by the IRS (currently Form 8332, Release of Claim to Exemption for Child of Divorced or Separated Parents). A written declaration not on the form designated by the IRS must conform to the substance of that form.

(2) Attachment to return. A noncustodial parent must attach the original written declaration to the parent’s return for the taxable year in which the child is claimed as a dependent. If a release of a claim to a child is for more than one year, the noncustodial parent must attach the original written declaration to the parent’s return for the first taxable year for which the release is effective. The noncustodial parent must attach a copy of the written declaration to the parent’s return for each subsequent taxable year for which the noncustodial parent claims the child as a dependent.

(3) Revocation of written declaration—(i) In general. A written declaration described in paragraph (d)(1) of this section may be revoked by providing written notice of the revocation to the other parent. The revocation may be effective no earlier than the taxable year that begins in the first calendar year after the calendar year in which the parent revoking the written declaration provides the written notice.

(ii) Form of revocation. The revocation may be made on a form designated by the IRS whether or not the written declaration was made on a form designated by the IRS. A revocation not on that form must conform to the substance of the form. The revocation must specify the year or years for which the revocation is effective. A revocation that does not specify a year or years has no effect. A revocation that specifies all future years is treated as specifying the first taxable year after the taxable year the revocation is executed and all subsequent taxable years.

(iii) Attachment to return. The custodial parent must attach the original revocation to the parent’s return for the taxable year for which the custodial parent claims a child as a dependent. If a revocation is for more than one year, the custodial parent must attach the original revocation to the parent’s return for the first taxable year for which the revocation is effective and a copy of the revocation to the parent’s return for each subsequent taxable year for which the custodial parent claims the child as a dependent. The custodial parent must keep a copy of the revocation and evidence of delivery of written notice of the revocation to the noncustodial parent.

(e) Coordination with other sections. A child who is treated as the qualifying child or qualifying relative of the noncustodial parent under section 152(e) and this section is treated as a dependent of both parents for purposes of sections 105(b), 132(h)(2)(B), and 213(d)(5).

(f) Examples. The provisions of this section are illustrated by the following examples which assume that each taxpayer’s taxable year is the calendar year, one or both of the child’s parents provide over one-half of the child’s support for the calendar year, the child is in the custody of one or both parents for more than one-half of the calendar year, and the child otherwise meets the requirements of a qualifying child under section 152(c). In addition, in each of the examples, there is no qualified pre-1985 instrument in effect. The examples are as follows:

Example 1. (i) B and C, the parents of Child, are divorced. In 2007, Child resides with B for 7 months and with C for 5 months. B signs a Form 8332 for 2007 allowing C to claim Child as a dependent for that year.

(ii) Under paragraph (c) of this section, B is the custodial parent of Child in 2007 because B is the parent with whom Child resides for the greater number of nights in 2007. Because B signs a Form 8332, under paragraph (b) of this section, Child is treated as the qualifying child of C if C attaches the Form 8332 to C’s 2007 return.

Example 2. (i) D and E, the parents of Child, are divorced. In 2007, Child resides with D for 7 months and with E for 5 months. D, the custodial parent, does not execute a Form 8332 or similar declaration for 2007.

(ii) Because D does not execute a Form 8332 or similar declaration for 2007, section 152(e) and this section do not apply to determine whether Child is treated as the qualifying child of D or E. Instead, whether Child is the qualifying child of D or E is determined under section 152(c).

Example 3. F and G, who never married, are the parents of Child. In 2007, Child spends alternate weeks residing with F and G. During a week when Child is residing with F, F gives Child permission to spend a night at the home of a friend. Under paragraph (c)(2) of this section, the night Child spends at the friend’s home is treated as a night in which Child resides with F for purposes of determining whether Child is residing with F or G for the greater number of nights in the calendar year.

Example 4. J and K are the divorced parents of Child. In 2007, Child spends alternate periods residing with J or K. In August of 2007, J and Child spend 10 nights together in a hotel while on vacation. Under paragraph (c) of this section, the 10 nights when J and Child are on vacation are treated as nights in which Child resides with J for purposes of determining whether Child is residing with J or K for the greater number of nights in the calendar year.

Example 5. (i) In 2006, L and M, the parents of Child, execute a written separation agreement. The agreement provides that Child will live with L and that M will make monthly child support payments to L. The agreement further provides that L will not claim Child as a dependent in 2007 and in subsequent alternate years. The agreement does not expressly condition L’s agreement not to claim Child as a dependent on M’s payment of child support or any other condition. The agreement contains all the other information requested on Form 8332. M attaches the agreement to M’s tax returns for 2007 and 2009.

(ii) In 2008, M fails to provide child support for Child, and L signs a Form 8332 revoking the release of L’s right to claim Child as a dependent for 2009 and delivers a copy of the Form 8332 to M. L attaches the Form 8332 revoking the release to L’s tax return for 2009 and keeps a copy of the revocation and evidence of delivery of written notice to M.

(iii) M may claim Child as a qualifying child for 2007 because L releases the right to claim Child as a dependent under paragraph (b)(3) of this section by executing the separation agreement, and M attaches the separation agreement to M’s tax return in accordance with paragraphs (d)(1) and (d)(2) of this section. The separation agreement qualifies as a written declaration under paragraph (d)(1) of this section because L’s agreement not to claim Child as a dependent is not conditioned on M’s payment of support or meeting of any other obligation, and the agreement otherwise conforms to the substance of Form 8332. For 2009, only L may claim Child as a qualifying child because in 2008 L revokes the release of the claim in accordance with paragraph (d)(3) of this section, and the revocation takes effect in 2009, the taxable year that begins in the first calendar year after L provides written notice of the revocation to M.

Example 6. The facts are the same as Example 5, except that the agreement expressly states that L agrees not to claim Child as a dependent only if M is current in the payment of support for Child at the end of the calendar year. The separation agreement does not qualify as a written declaration under paragraph (d)(1) of this section because L’s agreement not to claim Child as a dependent is conditioned on M’s payment of support. Therefore, M may not claim Child as a qualifying child in 2007 or 2009.

Example 7. (i) N and P are the divorced parents of Child. Child resides with N for ten months and with P for two months in each year 2007 through 2009. In 2007, N provides a written statement to P that provides that N will not claim Child as a dependent but does not specify a year or years. P attaches the statement to P’s returns for 2007 through 2009.

(ii) Because the written statement provided by N does not specify the year or years for which P may claim Child as a qualifying child, under paragraph (d)(1) of this section, the written statement is not a written declaration that conforms to the substance of Form 8332. Therefore, P may not claim Child as a qualifying child in 2007 through 2009.

Example 8. (i) R and S are the divorced parents of Child. Child resides solely with R. The divorce decree requires S to pay child support to R and requires R to execute a Form 8332 to release the right to claim Child as a qualifying child to S. R fails to sign a Form 8332 for 2007, and S attaches an unsigned Form 8332 to S’s return for 2007.

(ii) Child is the qualifying child of R for 2007. The order in the divorce decree requiring R to execute a Form 8332 is ineffective to allocate the right to claim Child as a qualifying child to S. Furthermore, under paragraph (d)(1) of this section, the unsigned Form 8332 does not conform to the substance of Form 8332. Therefore, S may not claim Child as a qualifying child in 2007.

(iii) If, however, R executes a Form 8332 for 2007 and S attaches the Form 8332 to S’s return, then S may claim Child as a qualifying child for 2007 under paragraph (d)(1) of this section.

(g) Effective date. This section applies to taxable years beginning after the date these regulations are published as final regulations in the Federal Register.

These proposed regulations set forth the methodology the IRS and the Treasury Department would use to establish mortality tables as provided under section 430(h)(3)(A) to be used for participants and beneficiaries to determine present value or make any computation under section 430. These mortality tables would apply as well for purposes of determining the current liability of a multiemployer plan pursuant to section 431(c)(6)(D)(iv)(II). In addition, pursuant to §1.412(l)(7)-1(a), these proposed regulations would apply for purposes of determining the current liability of a plan for which application of the PPA changes to section 412 is delayed (see sections 104 through 106 of PPA). Under the proposed regulations, mortality tables to be used with respect to disabled individuals would be provided in guidance published in the Internal Revenue Bulletin (IRB).

The new mortality tables under section 430(h)(3)(A) would be based on the tables contained in the RP-2000 Mortality Tables Report. In response to Notice 2003-62, commentators generally recommended that the RP-2000 mortality tables be the basis for the mortality tables used under section 412(l)(7)(C)(ii). The IRS and the Treasury Department reviewed the RP-2000 mortality tables and the accompanying report published by the Society of Actuaries, and determined to use the RP-2000 mortality tables as the basis for final regulations under section 412(l)(7)(C)(ii) because the RP-2000 mortality tables form the best available basis for predicting mortality of pension plan participants and beneficiaries (other than disabled individuals) based on pension plan experience, including expected trends. Because section 430 applies this same standard, the mortality tables set forth in these proposed regulations under section 430 are also based on the RP-2000 mortality tables. Like the mortality tables provided in the final section 412(l) regulations, the mortality tables set forth in these proposed regulations are gender-distinct because of significant differences between expected male mortality and expected female mortality. See §601.601(d)(2)(ii)(b) of this chapter.

The mortality tables set forth in these proposed regulations would provide separate mortality rates for annuitants and nonannuitants. This distinction has been made because the RP-2000 Mortality Tables Report indicates that these two groups have significantly different mortality experience. This is particularly true at typical ages for early retirees, where the number of health-induced early retirements results in a population that has higher mortality rates than the population of currently employed individuals. While the use of separate mortality rates for these groups of individuals will likely entail changes in programming of actuarial software, the IRS and the Treasury Department believe that the improvement in accuracy resulting from the use of separate mortality tables for annuitants and nonannuitants more than offsets the added complexity.

Under these proposed regulations, the annuitant mortality tables would be applied to determine the present value of benefits for annuitants. The annuitant mortality tables are also used for nonannuitants (active employees and terminated vested participants) for the periods beginning when the nonannuitants are projected to commence receiving benefits, while the nonannuitant mortality tables are applied for the periods before nonannuitants are projected to commence receiving benefits. For any period in which an annuitant is projected to be receiving benefits, the mortality table applicable to any beneficiary of that annuitant is the annuitant mortality table.

The RP-2000 Mortality Tables Report sets forth mortality tables that reflect expected mortality as of 2000, along with projection factors that are used to reflect the impact of expected improvements in mortality. Similarly, the mortality tables set forth in the proposed regulations are based on expected mortality as of 2000 and reflect the impact of expected improvements in mortality. Commentators to prior guidance generally stated that the projection of mortality improvement is desirable because it reflects expected mortality more accurately than using mortality tables that do not reflect such projection. The IRS and the Treasury Department agree with these comments, and believe that failing to project mortality improvement in determining the funding target would tend to result in underfunding. The proposed regulations permit plan sponsors to apply the projection of mortality improvement in either of two ways: through use of static tables that are updated annually to reflect expected improvements in mortality, or through use of generational tables.

The proposed regulations set forth base tables for annuitants and nonannuitants, as well as a set of projection factors. The base tables set forth in the proposed regulations generally provide the same rates as the RP-2000 mortality tables, except that they have been extended so that the annuitant and nonannuitant tables have mortality rates available at each age. The RP-2000 Mortality Tables Report did not develop annuitant rates before age 50 or nonannuitant rates after age 70. The extended nonannuitant tables in these proposed regulations were created by (1) using nonannuitant rates through age 70, (2) using annuitant rates for ages over 80, and (3) blending the rates to produce a smooth transition between the two tables, using increasing fractions. The total difference between the rates at ages 70 and 80 is divided by 55; the rate at age 71 is set equal to the rate at age 70 plus 1/55 of the total difference, the age 72 rate is equal to the rate at age 71 plus 2/55 of the total difference, etc.

A similar approach was used to develop the base tables for annuitants. For male annuitants, annuitant rates from the RP-2000 Mortality Tables Report were used for ages 50 and over, nonannuitant rates from the RP-2000 Mortality Tables Report were used through age 40, and rates between ages 41 and 49 were smoothed to create a smooth transition using the same methodology as was used for the nonannuitant tables. For female annuitants, annuitant rates from the RP-2000 Mortality Tables Report were used for ages 50 and over. However, to avoid anomalous results, female nonannuitant rates were used through age 46 (rather than age 40) and, accordingly, rates were smoothed between ages 47 and 49. The smoothing methodology for the female annuitant tables was the same as that used for the male tables but, because a shorter transition period was used, the difference between the age 46 and the age 50 mortality rates was smoothed using a denominator of 10 instead of 55.

For a plan sponsor that chooses to use the generational mortality tables, the mortality rate for each particular age would be projected for each individual participant to reflect projected improvement for the period of time until the participant reaches the particular age using the applicable base table along with the projection factors provided under the proposed regulations. These projection factors are from Mortality Projection Scale AA, which was recommended for use in the UP-94 Study and in the RP-2000 Mortality Tables Report. For example, to obtain the age 54 mortality rate for a male annuitant born in 1974 using the generational mortality tables, the age 54 male annuitant table rate is projected 28 years using the age 54 male Projection Scale AA rate set forth in the proposed regulations. The projection period is 28 years because a participant born in 1974 would attain age 54 in 2028, 28 years after the base year of the tables set forth in the proposed regulations. In this instance, because the male age 54 annuitant rate is .005797 under the base table, and the male age 54 Projection Scale AA rate set forth in paragraph (d) of §1.430(h)(3)-1 is .020, the age 54 male annuitant rate for participants born in 1974 is .003293 (.005797 * (1-.020)²⁸).

The static mortality tables that would be permitted to be used under the proposed regulations are constructed from the base table used for purposes of the generational mortality tables. The static mortality tables are projected from the base table for the year 2000 through the year of valuation with further projection to reflect the approximate expected duration of liabilities. The static mortality tables for annuitants under the proposed regulations reflect projection through the year of valuation with a further projection period of 7 years, and the static mortality tables for nonannuitants under the proposed regulations reflect projection through the year of valuation with a further projection period of 15 years. These projection periods were selected as the expected average duration of liabilities and are consistent with projection periods suggested by commentators. To be consistent with the original construction of the RP-2000 mortality tables, both the static annuitant and nonannuitant tables use the rates from the projected annuitant table for ages 80 and over and from the projected nonannuitant table for ages 40 and younger (ages 44 and younger for females). For a smooth transition between the different projection periods for annuitants versus nonannuitants, the rates for ages 71 through 79 and for ages 41 through 49 (ages 45 through 49 for females) were smoothed using the same technique as that used in constructing the base tables.

The static mortality tables that would apply with respect to valuation dates occurring during 2008 are set forth in the proposed regulations. The mortality tables to be used for valuation dates in subsequent years would be published in the IRB. Comments are requested regarding whether it would be desirable to publish a series of tables for each of a number of years (such as five years) along with final regulations, with tables for subsequent years to be published in the IRB.

As an example of the use of the static tables for the 2008 calendar plan year, with respect to a 45-year-old active participant who is projected to commence receiving an annuity at age 55, the funding target would be determined using the applicable nonannuitant mortality table for the period before the participant attains age 55 (so that the probability of an active male participant living from age 45 to the age of 55 using the mortality table that would apply in 2008 is 98.61%) and the applicable annuitant mortality table after the participant attains age 55. Similarly, if a 45-year-old terminated vested participant is projected to commence an annuity at age 65, the funding target would be determined using the applicable nonannuitant mortality table for the period before the participant attains age 65 and the applicable annuitant mortality table for ages 65 and above.

These proposed regulations would provide an option for smaller plans (plans where the total of active and inactive participants is less than 500) that choose to use static mortality tables to use a single blended static table for all participants — in lieu of the separate tables for annuitants and nonannuitants — in order to simplify the actuarial valuation for these plans. This blended table would be constructed from the separate nonannuitant and annuitant tables using the nonannuitant/annuitant weighting factors published in the RP-2000 Mortality Tables Report. However, because the RP-2000 Mortality Tables Report does not provide weighting factors before age 51 or after age 69, the IRS and the Treasury Department would extend the table of weighting factors (using straight-line interpolation) for ages 41 through 50 (ages 45-50 for females) and for ages 70 through 79 in order to develop the blended table.

These proposed regulations would set forth the framework for the development and use of substitute mortality tables in connection with present value determinations and other computations under section 430(h)(3)(C). The provision generally provides for the use of substitute mortality tables by a plan that is subject to section 430, in lieu of the mortality tables provided under section 430(h)(3)(A) and §1.430(h)(3)-1, upon written request of the plan sponsor and approval of the Commissioner.

Substitute mortality tables must reflect the actual mortality experience of the pension plan maintained by the plan sponsor for which the tables are to be used and that mortality experience must be credible. Separate mortality tables must be established for each gender under the plan, and a substitute mortality table is permitted to be established for a gender only if the plan has credible mortality experience with respect to that gender. If the mortality experience for one gender is credible but the mortality experience for the other gender is not credible, the substitute mortality tables are used for the gender that has credible mortality experience, and the mortality tables under §1.430(h)(3)-1 are used for the gender that does not have credible mortality experience. If separate mortality tables under section 430(h)(3)(D) are used for certain disabled individuals under a plan, then those individuals are disregarded for all purposes with respect to substitute mortality tables under section 430(h)(3)(C). Thus, if the mortality tables under section 430(h)(3)(D) are used for certain disabled individuals under a plan, mortality experience with respect to those individuals must be excluded in determining mortality rates for substitute mortality tables with respect to a plan.

Under the proposed regulations, a substitute mortality table is based on credible mortality experience for a gender within a plan if and only if the mortality experience is based on at least 1,000 deaths within that gender over the period covered by the experience study. The experience study must be based on mortality experience data over a 2, 3, or 4-consecutive year period, the last day of which must be less than 3 years before the first day of the first plan year for which the substitute mortality tables are to apply. The 1,000 deaths threshold is set at a level so that there is a high degree of confidence that the plan’s past mortality experience will be predictive of its future mortality, and is consistent with relevant actuarial literature (see, for example, Thomas N. Herzog, Introduction to Credibility Theory (1999); Stuart A. Klugman, et al., Loss Models: From Data to Decisions (2004)).

Development of a substitute mortality table under the proposed regulations requires creation of a base table and identification of a base year, which are then used to determine a substitute mortality table. The base table would be developed from a study of the mortality experience of the plan using amounts-weighted data. The proposed regulations set forth rules regarding development of amounts-weighted mortality rates for an age and the determination of the base year. The proposed regulations provide that amounts-weighted mortality rates may be derived from amounts-weighted mortality rates for age groups. Guidance issued by the Commissioner may specify grouping rules (for example, 5-year age groups, except for extreme ages) and methods for developing amounts-weighted mortality rates for individual ages from amounts-weighted mortality rates initially determined for each age group. In addition, the proposed regulations would provide that base tables may be constructed either directly through graduation of amounts-weighted mortality rates or indirectly by applying a level percentage to tables prescribed by section 430(h)(3)(A), provided that the resulting tables sufficiently reflect the plan’s mortality experience. The Commissioner may permit the construction of base tables through application of a level percentage to other recognized mortality tables, applying similar standards to ensure that the resulting tables are sufficiently reflective of the plan’s mortality experience.

In general, substitute mortality tables are permitted to be used for a plan only if the use of substitute mortality tables is approved for each other pension plan subject to the requirements of section 430 that is maintained by the plan sponsor or by a member of the sponsor’s controlled group. However, under the proposed regulations, the use of substitute mortality tables for one plan would not be prohibited merely because another plan subject to section 430 that is maintained by the plan sponsor (or by a member of the plan sponsor’s controlled group) cannot use substitute mortality tables because neither the males nor the females under that plan have credible mortality experience for a plan year. Thus, if a sponsor’s controlled group contains two pension plans subject to section 430, each of which has credible mortality experience for at least one gender, either both plans must obtain approval from the Commissioner to use substitute mortality tables or neither plan may use substitute mortality tables. By contrast, if for one of those plans neither males nor females have credible mortality experience, then the plan without credible mortality experience will not interfere with the ability of the plan with credible mortality experience to use substitute mortality tables.

Under the proposed regulations, the requirement that the plan sponsor demonstrate the lack of credible mortality experience for both the male and female populations in other plans maintained by the plan sponsor (and by members of the plan sponsor’s controlled group) must be satisfied annually. For each plan year in which a plan uses substitute mortality tables, the demonstration that both genders of another plan maintained by the plan sponsor do not have credible mortality experience (that is, there are less than 1,000 deaths within each gender) must be made using a 4-year period for mortality experience that ends less than 3 years before the first day of that plan year.

For example, a plan sponsor that requests to use substitute mortality tables for a plan for the plan year that begins January 1, 2008, would have to show, as part of its submission to the Commissioner, that both the male and female populations in all other defined benefit plans of the plan sponsor (and in the plan sponsor’s controlled group) that are subject to section 430 and that do not use substitute mortality tables do not have credible mortality experience using a 4-year period that ends no earlier than January 2, 2005 (that is, each gender in those plans did not experience 1,000 deaths during that 4-year period). If the plan sponsor chooses to use the 4-year period from January 1, 2003, through December 31, 2006, to demonstrate the lack of credible mortality experience for the other plans, then the plan can rely on this same data to demonstrate the lack of credible mortality experience for 2009 as well because the less-than-3-years requirement is still met with respect to the 2009 plan year. However, the plan would not be able to use this same data to demonstrate lack of credibility for the 2010 plan year because the last day of the experience study used for the demonstration (the January 1, 2003 — December 31, 2006 period) is too distant in time (3 or more years) from the first day of the plan year (January 1, 2010).

Although the proposed regulations permit a plan sponsor to use an experience study to demonstrate a lack of credible mortality experience for a plan population for multiple years, plan sponsors are encouraged to update experience studies annually as new mortality data become available for the plan population. In such a case, if an updated test reveals 1,000 or more deaths for the more recent 4-year period, the plan sponsor nonetheless will be able to continue to use substitute mortality tables for one plan year by demonstrating that the other plans in the controlled group do not have credible mortality experience based on the earlier experience study. This will give the plan sponsor sufficient time to develop substitute mortality tables for the plan population with newly credible mortality experience and to obtain the Commissioner’s approval to use those tables prior to the first year substitute mortality tables are to be used for that population.

Under the proposed regulations, a plan’s substitute mortality tables must be generational mortality tables. Substitute mortality tables are determined using the base mortality tables developed from the experience study and the projection factors provided in Projection Scale AA, as set forth in §1.430(h)(3)-1(d). Under the generational mortality tables, the probability of an individual’s death at a particular age is determined as the individual’s base mortality rate (that is, the applicable base mortality rate from the base mortality table for the age for which the probability of death is being determined) multiplied by the mortality improvement factor. The mortality improvement factor is equal to (1 - projection factor for that age)ⁿ, where n is equal to the projection period (that is, the number of years between the base year for the base mortality table and the year for which the probability of death is being determined).

The proposed regulations would require separate tables to be established for males and females under a plan. Under the proposed regulations, separate substitute mortality tables would be permitted (but not required) to be established for separate populations within a gender, such as annuitants and nonannuitants or hourly and salaried individuals. The proposed regulations would provide that separate substitute mortality tables are permitted to be used for a separate population within a gender under a plan only if all individuals of that gender in the plan are divided into separate populations, each separate population has credible mortality experience (determined in the same manner as determining whether a gender has credible mortality experience), and the separate substitute mortality table for each separate population is developed using mortality experience data for that population. For example, in the case of a plan that has credible mortality experience data for both its male hourly and male salaried populations, separate substitute mortality tables could be used for those two separate populations. However, if the plan does not have credible mortality experience for its male salaried population, it would not be permissible to use substitute mortality tables for its male hourly population and the standard mortality tables described in §1.430(h)(3)-1 for its male salaried population.

The requirement that each separate population have credible mortality experience does not apply in the case of separate mortality tables that are developed for annuitant and nonannuitant populations within a gender. Thus, the proposed regulations would provide that substitute mortality tables for separate annuitant and nonannuitant populations may be used within a gender even if only one of those separate populations has credible mortality experience. Similarly, if separate populations with credible mortality experience are established within a gender, then any of those populations may be further subdivided into separate annuitant and nonannuitant subpopulations, provided that at least one of the two resulting subpopulations has credible mortality experience. In such a case, the standard mortality tables under §1.430(h)(3)-1 must be used for a resulting subpopulation that does not have credible mortality experience. For example, in the case of a plan that has credible mortality experience for both its male hourly and salaried individuals, if the male salaried annuitant population has credible mortality experience, it may use substitute mortality tables with respect to that population even if the male salaried nonannuitant population uses the standard mortality tables under §1.430(h)(3)-1 (because that nonannuitant population does not have credible mortality experience). For purposes of demonstrating that an annuitant or nonannuitant population within a gender or within a separate population does not have credible mortality experience, the demonstration of lack of credible mortality experience is made on the same basis as for purposes of demonstrating a lack of credible mortality experience for a gender.

The proposed regulations would provide a limited time period during which a newly acquired plan that does not use substitute mortality tables does not prevent another plan from using substitute mortality tables. Under the proposed regulations, the use of substitute mortality tables for a plan is not prohibited merely because a newly acquired plan does not use substitute mortality tables, but only through the last day of the plan year of the plan using substitute mortality tables that contains the end of the period described in section 410(b)(6)(C). For the following plan year, the mortality tables prescribed under §1.430(h)(3)-1 would apply with respect to the plan (and all other plans within the plan sponsor’s controlled group, including the acquired plan) unless approval to use substitute mortality tables has been obtained with respect to the acquired plan, or the acquired plan cannot use substitute mortality tables because neither the males nor the females under the plan have credible mortality experience. For example, if the employer acquires a plan in September 2009 that does not use substitute mortality tables and that has a plan year that ends June 30, the acquisition of that plan will not impair the continued use of substitute mortality tables by a pre-existing calendar year plan of the employer through the end of the 2011 calendar year. This is because the section 410(b)(6)(C) transition period for the newly acquired plan will end on June 30, 2011. Under the proposed regulations, a plan is treated as a newly acquired plan if it is acquired or otherwise becomes maintained by the plan sponsor (or by a member of the plan sponsor’s controlled group) in connection with a merger, acquisition, or similar transaction described in §1.410(b)-2(f). The proposed regulations would provide that a plan is also treated as a newly acquired plan if it is established in connection with a transfer in accordance with section 414(l) of assets and liabilities from another employer’s plan in connection with a merger, acquisition, or similar transaction described in §1.410(b)-2(f).

In the case of a newly acquired plan, the demonstration of whether credible mortality experience exists for the plan may be made by either including or excluding mortality experience data for the period prior to the date the plan becomes maintained by a member of the new plan sponsor’s controlled group. If a plan sponsor excludes mortality experience data prior to the date the plan became maintained within the new plan sponsor’s controlled group, the exclusion must apply for all populations within the plan. For example, it is impermissible to include the data for hourly individuals for the pre-acquisition period but exclude the data for salaried individuals for that same period.

In order to demonstrate a lack of credible mortality experience with respect to a gender for a plan year, a special rule applies if the plan’s mortality experience demonstration for a plan year is made by excluding mortality experience for the period prior to the date the newly acquired plan becomes maintained within the new plan sponsor’s controlled group. In such a case, an employer is permitted to demonstrate a plan’s lack of credible mortality experience using an experience study period of less than four years, provided that the experience study period begins with the date the plan becomes maintained within the employer’s controlled group and ends not more than one year and one day before the first day of the plan year with respect to which the lack of credible mortality experience demonstration is made.

The proposed regulations would provide rules for aggregating plans for purposes of using substitute mortality tables. Under the proposed regulations, in order to use a set of substitute mortality tables for two or more plans, the applicable rules are applied by treating those plans as a single plan. In such a case, the substitute mortality tables must be used for all such plans and must be based on data collected with respect to all such plans. Although plans generally are not required to be aggregated, the proposed regulations would require a plan to be aggregated with any plan that was previously spun off from that plan if one purpose of the spinoff was to avoid the use of substitute mortality tables for any of the plans involved in the spinoff.

Under the proposed regulations, in order to use substitute mortality tables with respect to a plan, a plan sponsor must submit a written request to the Commissioner that demonstrates that those substitute mortality tables comply with applicable requirements. A request to use substitute mortality tables must state the first plan year and the term of years (not more than 10) that the tables are requested to be used. In general, substitute mortality tables cannot be used for a plan year unless the plan sponsor submits the written request to use substitute mortality tables at least 7 months prior to the first day of the first plan year for which the substitute mortality tables are to apply. However, the timing of the written request to use substitute mortality tables does not prevent a plan from using substitute mortality tables for a plan year if the written request is submitted no later than October 1, 2007. This special rule allows plan sponsors sufficient time to review the proposed regulations and other guidance and prepare requests to use substitute mortality tables for use in 2008.

Under the proposed regulations, experience data cannot be used to develop a base table if the number of individuals in the population covered by the table (for example, the male annuitants) as of the last day of the plan year before the year the request to use substitute mortality tables is made (or a reasonable estimate of that number), compared to the average number of individuals in that population over the years covered by the experience study on which the substitute mortality tables are based, reflects a difference of 20 percent or more, unless it is demonstrated to the satisfaction of the Commissioner that the experience data is accurately predictive of future mortality of that plan population (taking into account the effect of the change in individuals) after appropriate adjustments to the data are made (for example, excluding data from individuals with respect to a spun-off portion of the plan).

Under the proposed regulations, the Commissioner may, in revenue rulings and procedures, notices and other guidance published in the IRB (see §601.601(d)(2)(ii)(b) of this chapter), provide additional guidance regarding the approval and use of substitute mortality tables under section 430(h)(3)(C) and related matters. The IRS will shortly issue a revenue procedure that will set forth the requirements related to requests to use substitute mortality tables.

In general, the Commissioner has a 180-day period to review a request for the use of substitute mortality tables. If the Commissioner does not issue a denial within this 180-day period, the request is deemed to have been approved unless the Commissioner and the plan sponsor have agreed to extend that period. The Commissioner may request additional information with respect to a submission. Failure to provide that information on a timely basis is grounds for denial of the plan sponsor’s request. In addition, the Commissioner will deny a request if the request fails to meet the requirements to use substitute mortality tables or if the Commissioner determines that a substitute mortality table does not sufficiently reflect the mortality experience of the applicable plan population.

The proposed regulations would provide rules regarding the duration of use of substitute mortality tables. Under the proposed regulations, substitute mortality tables generally are used with respect to a plan for the term of consecutive plan years specified in the plan sponsor’s written request to use such tables and approved by the Commissioner, or such shorter period prescribed by the Commissioner in the approval to use substitute mortality tables. If the term of use of a substitute mortality table ends for any reason, the mortality tables specified in §1.430(h)(3)-1 will apply with respect to the plan unless the plan sponsor has obtained approval to use substitute mortality tables for a further term. The proposed regulations would provide that a plan’s substitute mortality tables cannot be used as of the earliest of the following: the second plan year following the plan year in which there is a significant change in the population covered by the substitute mortality table (generally, a change of at least 20% from the average number of individuals included in the experience study); or the plan year following the plan year in which a substitute mortality table for a plan population is no longer accurately predictive of future mortality of that population, as determined by the Commissioner or as certified by the plan’s actuary to the satisfaction of the Commissioner. In addition, the proposed regulations would provide that a plan’s substitute mortality tables cannot be used after the date specified in guidance published in the IRB (see §601.601(d)(2)(ii)(b) of this chapter) pursuant to a replacement of mortality tables specified under section 430(h)(3)(A) (other than annual updates to the static mortality tables).

(a) Basis for mortality tables—(1) In general. This section sets forth rules for the mortality tables to be used in determining present value or making any computation under section 430. Generally applicable mortality tables for participants and beneficiaries are set forth in this section pursuant to section 430(h)(3)(A). In lieu of using the mortality tables provided under this section with respect to participants and beneficiaries, plan-specific substitute mortality tables are permitted to be used for this purpose pursuant to section 430(h)(3)(C) provided that the requirements of §1.430(h)(3)-2 are satisfied. Mortality tables that may be used with respect to disabled individuals are to be provided in guidance published in the Internal Revenue Bulletin. See §601.601(d)(2)(ii)(b) of this chapter.

(2) Static tables or generational tables permitted. The generally applicable mortality tables provided under section 430(h)(3)(A) are the static tables described in paragraph (a)(3) of this section and the generational mortality tables described in paragraph (a)(4) of this section. A plan is permitted to use either of those sets of mortality tables with respect to participants and beneficiaries pursuant to this section.

(3) Static tables. The static mortality tables that are permitted to be used pursuant to paragraph (a)(2) of this section are updated annually to reflect expected improvements in mortality experience as described in paragraph (c)(2) of this section. Static mortality tables that are to be used with respect to valuation dates occurring during 2008 are provided in paragraph (e) of this section. The mortality tables to be used with respect to valuation dates occurring in later years are to be provided in guidance published in the Internal Revenue Bulletin. See §601.601(d)(2)(ii)(b) of this chapter.

(4) Generational mortality tables—(i) In general. The generational mortality tables that are permitted to be used pursuant to paragraph (a)(2) of this section are determined pursuant to this paragraph (a)(4) using the base mortality tables and projection factors set forth in paragraph (d) of this section. Under the generational mortality tables, the probability of an individual’s death at a particular age is determined as the individual’s base mortality rate (that is, the applicable mortality rate from the table set forth in paragraph (d) of this section for the age for which the probability of death is being determined) multiplied by the mortality improvement factor. The mortality improvement factor is equal to (1 - projection factor for that age)ⁿ, where n is equal to the projection period. For this purpose, the projection period is the number of years between 2000 and the year for which the probability of death is being determined.

(ii) Examples of calculation. As an example of the use of generational mortality tables under paragraph (a)(4)(i) of this section, for purposes of determining the probability of death at age 54 for a male annuitant born in 1974, the base mortality rate is .005797, the projection factor is .020, and the projection period (the period from the year 2000 until the year the participant will attain age 54) is 28 years, so that the mortality improvement factor is .567976, and the probability of death at age 54 is .003293. Similarly, under these generational mortality tables, the probability of death at age 55 for the same male annuitant would be determined by using the base mortality rate and projection factor at age 55, and a projection period of 29 years (the period from the year 2000 until the year the participant will attain age 55). Thus, the base mortality rate is .005905, the projection factor is .019, so that the mortality improvement factor is .573325 ((1-.019)²⁹), and the probability of death at age 55 is .003385 (.573325 times .005905). Because these generational mortality tables reflect expected improvements in mortality experience, no periodic updates are needed.

(b) Use of the tables—(1) Separate tables for annuitants and nonannuitants—(i) In general. Separate tables are provided for use for annuitants and nonannuitants. The nonannuitant mortality table is applied to determine the probability of survival for a nonannuitant for the period before the nonannuitant is projected to commence receiving benefits. The annuitant mortality table is applied to determine the present value of benefits for each annuitant, and for each nonannuitant for the period beginning when the nonannuitant is projected to commence receiving benefits. For purposes of this section, an annuitant means a plan participant who has commenced receiving benefits and a nonannuitant means a plan participant who has not yet commenced receiving benefits (for example, an active employee or a terminated vested participant). A participant whose benefit has partially commenced is treated as an annuitant with respect to the portion of the benefit which has commenced and a nonannuitant with respect to the balance of the benefit. In addition, for any period in which an annuitant is projected to be receiving benefits, any beneficiary with respect to that annuitant is also treated as an annuitant for purposes of this paragraph (b)(1).

(ii) Examples of calculation. As an example of the use of separate annuitant and nonannuitant tables under paragraph (b)(1)(i) of this section, with respect to a 45-year-old active participant who is projected to commence receiving an annuity at age 55, the funding target would be determined using the nonannuitant mortality table for the period before the participant attains age 55 (so that, if the static mortality tables are used pursuant to paragraph (a)(3) of this section, the probability of an active male participant living from age 45 to age 55 using the table that applies for a plan year beginning in 2008 is 98.61%) and the annuitant mortality table for the period ages 55 and above. Similarly, if a 45-year-old terminated vested participant is projected to commence an annuity at age 65, the funding target would be determined using the nonannuitant mortality table for the period before the participant attains age 65 and the annuitant mortality table for ages 65 and above.

(2) Small plan tables. If static mortality tables are used pursuant to paragraph (a)(3) of this section, as an alternative to the separate static tables specified for annuitants and nonannuitants pursuant to paragraph (b)(1) of this section, a combined static table that applies the same mortality rates to both annuitants and nonannuitants is permitted to be used for a small plan. For this purpose, a small plan is defined as a plan with fewer than 500 participants (including both active and inactive participants).

(c) Construction of static tables—(1) Source of basic rates. The static mortality tables that are used pursuant to paragraph (a)(3) of this section are based on the base mortality tables set forth in paragraph (d) of this section.

(2) Projected mortality improvements. The mortality rates under the base mortality tables are projected to improve using the projection factors provided in Projection Scale AA, as set forth in paragraph (d) of this section. Using these projection factors, the mortality rate for an individual at each age is determined as the individual’s base mortality rate (that is, the applicable base mortality rate from the table set forth in paragraph (d) of this section for the individual at that age) multiplied by the mortality improvement factor. The mortality improvement factor is equal to (1 - projection factor for that age)ⁿ, where n is equal to the projection period. The annuitant mortality rates for a plan year are determined using a projection period that runs from the calendar year 2000 until 7 years after the calendar year that contains the valuation date for the plan year. The nonannuitant mortality rates for a plan year are determined using a projection period that runs from the calendar year 2000 until 15 years after the calendar year that contains the valuation date for the plan year. Thus, for example, for a plan year with a January 1, 2012, valuation date, the annuitant mortality rates are determined using a projection period that runs from 2000 until 2019 (19 years) and the nonannuitant mortality rates are determined using a projection period that runs from 2000 until 2027 (27 years).

(3) Construction of combined tables for small plans. The combined mortality tables that are permitted to be used for small plans pursuant to paragraph (b)(2) of this section are constructed from the separate nonannuitant and annuitant tables using the weighting factors for small plans that are set forth in paragraph (d) of this section. The weighting factors are applied to develop these mortality tables using the following equation: Combined mortality rate = [nonannuitant rate * (1- weighting factor)] + [annuitant rate * weighting factor].

(d) Base mortality tables and projection factors. The following base mortality tables and projection factors are used to determine generational mortality tables for purposes of determining present value or making any computation under section 430 as set forth in paragraph (a)(4) of this section. In addition, the following base mortality tables and projection factors are used to determine the static mortality tables that are used for purposes of determining present value or making any computation under section 430 as set forth in paragraphs (a)(3) and (c) of this section. See §1.430(h)(3)-2(c)(3) for rules regarding the required use of the projection factors set forth in this paragraph (d) in connection with a plan-specific substitute mortality table.

(e) Static mortality tables with respect to valuation dates occurring during 2008. The following static mortality tables are used pursuant to paragraph (a)(3) of this section for determining present value or making any computation under section 430 with respect to valuation dates occurring during 2008.

(f) Applicability date. This section applies for plan years beginning on or after January 1, 2008.

Par. 3. Section 1.430(h)(3)-2 is added to read as follows:

(a) In general. This section sets forth rules for the use of substitute mortality tables under section 430(h)(3)(C) in determining any present value or making any computation under section 430 in accordance with §1.430(h)(3)-1(a)(1). In order to use substitute mortality tables, a plan sponsor must obtain approval to use substitute mortality tables for the plan in accordance with the procedures set forth in paragraph (b) of this section. Paragraph (c) of this section sets forth rules for the development of substitute mortality tables, including guidelines for determining whether a plan has sufficient credible mortality experience to use substitute mortality tables. Paragraph (d) of this section sets forth special rules regarding the use of substitute mortality tables. The Commissioner may, in revenue rulings and procedures, notices and other guidance published in the Internal Revenue Bulletin (see §601.601(d)(2)(ii)(b) of this chapter), provide additional guidance regarding approval and use of substitute mortality tables under section 430(h)(3)(C) and related matters.

(b) Procedures for obtaining approval to use substitute mortality tables—(1) Written request to use substitute mortality tables—(i) General requirements. In order to use substitute mortality tables, a plan sponsor must submit a written request to the Commissioner that demonstrates that those substitute mortality tables meet the requirements of section 430(h)(3)(C) and this section. This request must state the first plan year and the term of years (not more than 10) that the tables are requested to be used.

(ii) Time for written request—(A) In general. Except as provided in paragraph (b)(1)(ii)(B) of this section, substitute mortality tables cannot be used for a plan year unless the plan sponsor submits the written request described in paragraph (b)(1)(i) of this section at least 7 months prior to the first day of the first plan year for which the substitute mortality tables are to apply.

(B) Special rule for requests submitted on or before October 1, 2007. Notwithstanding the rule of paragraph (b)(1)(ii)(A) of this section, the timing of the written request described in paragraph (b)(1)(i) of this section does not prevent a plan from using substitute mortality tables for a plan year provided that the written request is submitted no later than October 1, 2007.

(2) Commissioner’s review of request—(i) In general. During the 180-day period that begins on the date the plan sponsor submits a request to use substitute mortality tables for a plan pursuant to this section, the Commissioner will determine whether the request to use substitute mortality tables satisfies the requirements of this section (including any published guidance issued pursuant to paragraph (a) of this section), and will either approve or deny the request. The Commissioner will deny a request if the request fails to meet the requirements of this section or if the Commissioner determines that a substitute mortality table does not sufficiently reflect the mortality experience of the applicable plan population.

(ii) Request for additional information. The Commissioner may request additional information with respect to the submission. Failure to provide that information on a timely basis constitutes grounds for denial of the request.

(iii) Deemed approval. Except as provided in paragraph (b)(2)(iv) of this section, if the Commissioner does not issue a denial within the 180-day review period, the request is deemed to have been approved.

(iv) Extension of time permitted. The Commissioner and a plan sponsor may, before the expiration of the 180-day review period, agree in writing to extend that period, provided that any such agreement also specifies any revisions in the plan sponsor’s request, including any change in the requested term of use of the substitute mortality tables.

(c) Development of substitute mortality tables—(1) Mortality experience requirements—(i) In general. Substitute mortality tables must reflect the actual mortality experience of the pension plan maintained by the plan sponsor for which the tables are to be used and that mortality experience must be credible mortality experience as described in paragraph (c)(1)(ii) of this section. Separate mortality tables must be established for each gender under the plan, and a substitute mortality table is permitted to be established for a gender only if the plan has credible mortality experience with respect to that gender.

(ii) Credible mortality experience. There is credible mortality experience for a gender within a plan if and only if, over the period covered by the experience study described in paragraph (c)(2)(ii) of this section, there are at least 1,000 deaths within that gender.

(iii) Gender without credible mortality experience—(A) In general. If, for the first year for which a plan uses substitute mortality tables, one gender has credible mortality experience but the other gender does not have credible mortality experience, the substitute mortality tables are used for the gender that does have credible mortality experience and the mortality tables under §1.430(h)(3)-1 are used for the gender that does not have credible mortality experience. For a subsequent plan year, the plan sponsor may continue to use substitute mortality tables for the gender with credible mortality experience without using substitute mortality tables for the other gender only if the other gender continues to lack credible mortality experience for that subsequent plan year.

(B) Demonstration of lack of credible mortality experience for a gender. In order to demonstrate that a gender within a plan does not have credible mortality experience for a plan year, the mortality experience of that population must be analyzed using a 4-year experience study that ends less than 3 years before the first day of that plan year. For example, if a plan uses substitute mortality tables based on credible mortality experience for its male population and the standard mortality tables under §1.430(h)(3)-1 for its female population, there must be an experience study which shows that the plan’s female population does not have at least 1,000 deaths in a 4-year period that ends less than 3 years before the first day of that plan year.

(iv) Disabled individuals. Under section 430(h)(3)(D), separate mortality tables are permitted to be used for certain disabled individuals. If such separate mortality tables are used for those disabled individuals, then those individuals are disregarded for all purposes under this section. Thus, if the mortality tables under section 430(h)(3)(D) are used for disabled individuals under a plan, mortality experience with respect to those individuals must be excluded in developing mortality rates for substitute mortality tables under this section.

(2) Base table and base year—(i) In general. Development of a substitute mortality table under this section requires creation of a base table and identification of a base year under this paragraph (c)(2). The base table and base year are then used to determine a substitute mortality table under paragraph (c)(3) of this section.

(ii) Experience study and base table requirements—(A) In general. The base table for a plan population must be developed from an experience study of the mortality experience of that plan population that generates amounts-weighted mortality rates based on experience data for the plan over 2, 3, or 4 consecutive years. The last day of the final year reflected in the experience data must be less than 3 years before the first day of the first plan year for which the substitute mortality tables are to apply. For example, if July 1, 2008, is the first day of the first plan year for which the substitute mortality tables will be used, then an experience study using calendar year data must include data collected for a period that ends no earlier than December 31, 2005.

(B) Amounts-weighted mortality rates. The amounts-weighted mortality rate for an age is equal to the quotient determined by dividing the sum of the accrued benefits (or payable benefits, in the case of individuals in pay status) for all individuals at that age at the beginning of the year who died during the year, by the sum of the accrued benefits (or payable benefits, in the case of individuals in pay status) for all individuals at that age at the beginning of the year, with appropriate adjustments for individuals who left the relevant plan population during the year for reasons other than death. Because amounts-weighted mortality rates for a plan cannot be determined without accrued (or payable) benefits, the mortality experience study used to develop a base table cannot include periods before the plan was established.

(C) Grouping of ages. Amounts-weighted mortality rates may be derived from amounts-weighted mortality rates for age groups. The Commissioner, in revenue rulings and procedures, notices, and other guidance, may specify grouping rules (for example, 5-year age groups, except for extreme ages such as ages above 100 or below 20) and methods for developing amounts-weighted mortality rates for individual ages from amounts-weighted mortality rates initially determined for each age group.

(D) Base table construction. The base tables must be constructed from the amounts-weighted mortality rates determined in paragraph (c)(2)(ii)(B) of this section. The base tables must be constructed either directly through graduation of the amounts-weighted mortality rates or indirectly by applying a level percentage to the applicable mortality table set forth in §1.430(h)(3)-1, provided that the adjusted table sufficiently reflects the mortality experience of the plan. The Commissioner also may permit the use of other recognized mortality tables in the construction of base tables, applying a similar mortality experience standard.

(iii) Base year requirements. Where there are 2 years of experience data, the base year is the calendar year in which the first year of the experience data begins. Where there are 3 or 4 years of experience data, the base year is the calendar year in which the second year of the experience data begins. If the base table is constructed by applying a level percentage to a table set forth in §1.430(h)(3)-1, then the percentage must be applied to the table under §1.430(h)(3)-1 after it has been projected to the base year using Projection Scale AA, as set forth in §1.430(h)(3)-1(d). Thus, for example, if the base year of the mortality experience study is 2004, the applicable base (year 2000) mortality rates must be projected four years prior to determining the level percentage to be applied to the applicable projected base (year 2000) mortality rates.

(iv) Change in number of individuals covered by table. Experience data cannot be used to develop a base table if the number of individuals in the population covered by the table (for example, the male annuitant population) as of the last day of the plan year before the year the request to use substitute mortality tables is made, compared to the average number of individuals in that population over the years covered by the experience study on which the substitute mortality tables are based, reflects a difference of 20 percent or more, unless it is demonstrated to the satisfaction of the Commissioner that the experience data is accurately predictive of future mortality of that plan population (taking into account the effect of the change in individuals) after appropriate adjustments to the data are made (for example, excluding data from individuals with respect to a spun-off portion of the plan). For this purpose, a reasonable estimate of the number of individuals in the population covered by the table may be used, such as the estimated number of participants and beneficiaries used for purposes of the PBGC Form 1-ES.

(3) Determination of substitute mortality tables—(i) In general. A plan’s substitute mortality tables must be generational mortality tables. Substitute mortality tables are determined using the base mortality tables developed pursuant to paragraph (c)(2) of this section and the projection factors provided in Projection Scale AA, as set forth in §1.430(h)(3)-1(d). Under the generational mortality tables, the probability of an individual’s death at a particular age is determined as the individual’s base mortality rate (that is, the applicable mortality rate from the base mortality table for the age for which the probability of death is being determined) multiplied by the mortality improvement factor. The mortality improvement factor is equal to (1 - projection factor for that age)ⁿ, where n is equal to the projection period (the number of years between the base year for the base mortality table and the calendar year in which the individual attains the age for which the probability of death is being determined).

(ii) Example of calculation. As an example of the use of generational mortality tables under paragraph (c)(3)(i) of this section, if approved substitute mortality tables are based on data collected during 2005 and 2006, the base year would be the first year of experience (2005) because the substitute tables are based on two years of experience data. If the tables show a base mortality rate of .006000 for male annuitants at age 54, the probability of death at age 54 for a male annuitant born in 1974 would be determined using the base mortality rate of .006000, the age-54 projection factor of .020 (pursuant to the Scale AA Projection Factors set forth in §1.430(h)(3)-1(d)) and a projection period of 23 years. The projection period is the number of years between the base year of 2005 and the calendar year in which the individual reaches age 54. Accordingly, the mortality improvement factor would be .628347 and the probability of death at age 54 would be .003770.

(4) Separate tables for specified populations—(i) In general. Except as provided in this paragraph (c)(4), separate substitute mortality tables are permitted to be used for separate populations within a gender under a plan only if—

(A) All individuals of that gender in the plan are divided into separate populations;

(B) Each separate population has credible mortality experience as provided in paragraph (c)(4)(iii) of this section; and

(C) The separate substitute mortality table for each separate population is developed using mortality experience data for that population.

(ii) Annuitant and nonannuitant separate populations. Notwithstanding paragraph (c)(4)(i)(B) of this section, substitute mortality tables for separate populations of annuitants and nonannuitants within a gender may be used even if only one of those separate populations has credible mortality experience. Similarly, if separate populations that satisfy paragraph (c)(4)(i)(B) of this section are established, then any of those populations may be further subdivided into separate annuitant and nonannuitant subpopulations, provided that at least one of the two resulting subpopulations has credible mortality experience. The standard mortality tables under §1.430(h)(3)-1 are used for a resulting subpopulation that does not have credible mortality experience. For example, in the case of a plan that has credible mortality experience for both its male hourly and salaried individuals, if the male salaried annuitant population has credible mortality experience, it may use substitute mortality tables with respect to that population even if the male salaried nonannuitant population uses the standard mortality tables under §1.430(h)(3)-1 (because that nonannuitant population does not have credible mortality experience).

(iii) Credible mortality experience for separate populations. In determining whether a separate population within a gender has credible mortality experience, the requirements of paragraph (c)(1)(ii) of this section must be satisfied but, in applying that paragraph, the separate population should be substituted for the particular gender. In demonstrating that an annuitant or nonannuitant population within a gender or within a separate population does not have credible mortality experience, the requirements of paragraph (c)(1)(iii) of this section must be satisfied but, in applying that paragraph, the annuitant (or nonannuitant) population should be substituted for the particular gender.

(d) Special rules—(1) All plans in controlled group must use substitute mortality tables—(i) In general. Except as otherwise provided in this paragraph (d)(1), substitute mortality tables are permitted to be used for a plan only if the use of substitute mortality tables is approved under this section for each other pension plan subject to the requirements of section 430 that is maintained by the sponsor and by each member of the plan sponsor’s controlled group. For purposes of this section, the term controlled group means any group treated as a single employer under subsection (b), (c), (m), or (o) of section 414.

(ii) Plans without credible experience—(A) In general. For the first year for which a plan uses substitute mortality tables, the use of substitute mortality tables for the plan is not prohibited merely because another plan described in paragraph (d)(1)(i) of this section cannot use substitute mortality tables because neither the males nor the females under that other plan have credible mortality experience for a plan year. For each subsequent plan year, the plan sponsor may continue to use substitute mortality tables for the plan with credible mortality experience without using substitute mortality tables for the other plan only if neither the males nor the females under that other plan have credible mortality experience for that subsequent plan year.

(B) Analysis of mortality experience. For each plan year in which a plan uses substitute mortality tables, in order to demonstrate that the male and female populations of another plan maintained by the plan sponsor (or by a member of the plan sponsor’s controlled group) do not have credible mortality experience, the requirements of paragraph (c)(1)(iii)(B) of this section must be satisfied for that plan year. Thus, a plan is not prohibited from using substitute mortality tables for a plan year merely because another plan in the controlled group of the plan sponsor does not have at least 1,000 male deaths and does not have at least 1,000 female deaths in a 4-year period that ends less than 3 years before the first day of that plan year.

(iii) Newly acquired plans not using substitute mortality tables—(A) In general. The use of substitute mortality tables for a plan is not prohibited merely because a newly acquired plan does not use substitute mortality tables, but only through the last day of the plan year of the plan using substitute mortality tables that contains the end of the period described in section 410(b)(6)(C). Thus, for the following plan year, the mortality tables prescribed under §1.430(h)(3)-1 apply with respect to the plan (and all other plans within the plan sponsor’s controlled group, including the acquired plan) unless—

(1) Approval to use substitute mortality tables has been obtained with respect to the acquired plan pursuant to paragraph (b)(1) of this section; or

(2) The acquired plan cannot use substitute mortality tables because neither the males nor the females under the plan have credible mortality experience as described in paragraph (c)(1)(ii) of this section (as determined in accordance with the rules of paragraph (d)(1)(iv) of this section).

(B) Definition of newly acquired plan. For purposes of this section, a plan is treated as a newly acquired plan if it becomes maintained by the plan sponsor (or by a member of the plan sponsor’s controlled group) in connection with a merger, acquisition, or similar transaction described in §1.410(b)-2(f). In addition, a plan also is treated as a newly acquired plan for purposes of this section if a plan is established in connection with a transfer in accordance with section 414(l) of assets and liabilities from another employer’s plan in connection with a merger, acquisition, or similar transaction described in §1.410(b)-2(f).

(iv) Demonstration of credible mortality experience for newly acquired plan—(A) In general. In general, in the case of a newly acquired plan described in paragraph (d)(1)(iii) of this section, the demonstration of whether credible mortality experience exists for the plan for a plan year may be made by either including or excluding mortality experience data for the period prior to the date the plan becomes maintained by a member of the new plan sponsor’s controlled group. If a plan sponsor excludes mortality experience data for the period prior to the date the plan becomes maintained within the new plan sponsor’s controlled group, the exclusion must apply for all populations within the plan.

(B) Demonstration of credible mortality experience. Regardless of whether mortality experience data for the period prior to the date a newly acquired plan becomes maintained within the new plan sponsor’s controlled group is included or excluded for a plan year, the provisions of this section, including the demonstration of credible mortality experience in accordance with paragraph (c)(1)(ii) of this section, must be satisfied before substitute mortality tables may be used with respect to the plan. Thus, for example, the plan must meet the rule in paragraph (c)(2)(ii)(A) of this section that the base table be based on mortality experience data for the plan over a 2, 3, or 4-consecutive year period that ends less than 3 years before the first day of the plan year for which substitute mortality tables will be used.

(C) Demonstration of lack of credible mortality experience. In the case of a newly acquired plan described in paragraph (d)(1)(iii) of this section, in order to demonstrate a lack of credible mortality experience with respect to a gender for a plan year, the rules of paragraph (c)(1)(iii)(B) of this section generally will apply. However, a special rule applies if the plan’s mortality experience demonstration for a plan year is made by excluding mortality experience for the period prior to the date the plan becomes maintained by a member of the new plan sponsor’s controlled group. In such a case, an employer is permitted to demonstrate a plan’s lack of credible mortality experience using an experience study period of less than four years, provided that the experience study period begins with the date the plan becomes maintained within the sponsor’s controlled group and ends not more than one year and one day before the first day of the plan year with respect to which the lack of credible mortality experience demonstration is made. Thus, if the transaction occurred on July 1, 2010, in order to demonstrate a lack of credible mortality experience for males and females for the plan year beginning January 1, 2012 (the first day of the plan year following the section 410(b)(6)(C) transition period), there must be an experience study which shows that the plan’s male and female populations each do not have 1,000 deaths during the period from July 1, 2010 - December 31, 2010. Similarly, in order to perform the demonstration to show a lack of credible mortality experience for the plan year beginning January 1, 2013, there must be an experience study which shows that the plan’s male and female populations each do not have 1,000 deaths during the period from July 1, 2010 - December 31, 2011.

(2) Duration of use of tables. Except as provided in paragraph (d)(4) of this section, substitute mortality tables are used with respect to a plan for the term of consecutive plan years specified in the plan sponsor’s written request to use such tables under paragraph (b)(1) of this section and approved by the Commissioner, or such shorter period prescribed by the Commissioner in the approval to use substitute mortality tables. Following the end of such term of use, or following any early termination of use described in paragraph (d)(4) of this section, the mortality tables specified in §1.430(h)(3)-1 apply with respect to the plan unless approval under paragraph (b)(1) of this section has been received by the plan sponsor to use substitute mortality tables for a further term.

(3) Aggregation—(i) Permissive aggregation of plans. In order for a plan sponsor to use a set of substitute mortality tables with respect to two or more plans, the rules of this section are applied by treating those plans as a single plan. In such a case, the substitute mortality tables must be used for the aggregated plans and must be based on data collected with respect to those aggregated plans.

(ii) Required aggregation of plans. In general, plans are not required to be aggregated for purposes of applying the rules of this section. However, for purposes of this section, a plan is required to be aggregated with any plan that was previously spun off from that plan for purposes of this section if the Commissioner determines that one purpose of the spinoff is to avoid the use of substitute mortality tables for any of the plans that were involved in the spinoff.

(4) Early termination of use of tables—(i) General rule. A plan’s substitute mortality tables cannot be used as of the earliest of—

(A) The plan year in which the plan fails to satisfy the requirements of paragraph (c)(1) of this section (regarding credible mortality experience requirements and demonstrations);

(B) The plan year in which the plan fails to satisfy the requirements of paragraph (d)(1) of this section (regarding use of substitute mortality tables by controlled group members);

(C) The second plan year following the plan year in which there is a significant change in individuals covered by the plan as described in paragraph (d)(4)(ii) of this section;

(D) The plan year following the plan year in which a substitute mortality table used for a plan population is no longer accurately predictive of future mortality of that population, as determined by the Commissioner or as certified by the plan’s actuary to the satisfaction of the Commissioner; or

(E) The date specified in guidance published in the Internal Revenue Bulletin (see §601.601(d)(2)(ii)(b) of this chapter) pursuant to a replacement of mortality tables specified under section 430(h)(3)(A) and §1.430(h)(3)-1 (other than annual updates to the static mortality tables issued pursuant to §1.430(h)-1(a)(3)).

(ii) Significant change in coverage—(A) Change in coverage from time of experience study. For purposes of applying the rules of paragraph (d)(4)(i)(C) of this section, a significant change in the individuals covered by a substitute mortality table occurs if there is an increase or decrease in the number of individuals of at least 20 percent compared to the average number of individuals in that population over the years covered by the experience study on which the substitute mortality tables are based. However, a change in coverage is not treated as significant if the plan’s actuary certifies in writing to the satisfaction of the Commissioner that the substitute mortality tables used for the plan population continue to be accurately predictive of future mortality of that population (taking into account the effect of the change in the population).

(B) Change in coverage from time of certification. For purposes of applying the rules of paragraph (d)(4)(i)(C) of this section, a significant change in the individuals covered by a substitute mortality table occurs if there is an increase or decrease in the number of individuals covered by a substitute mortality table of at least 20 percent compared to the number of individuals in a plan year for which a certification described in paragraph (d)(4)(ii)(A) of this section was made on account of a prior change in coverage. However, a change in coverage is not treated as significant if the plan’s actuary certifies in writing to the satisfaction of the Commissioner that the substitute mortality tables used by the plan with respect to the covered population continue to be accurately predictive of future mortality of that population (taking into account the effect of the change in the plan population).

(e) Applicability date. This section applies for plan years beginning on or after January 1, 2008.

Par. 4. Section 1.431(c)(6)-1 is added to read as follows:

The mortality assumptions that apply to a defined benefit plan for the plan year pursuant to section 430(h)(3)(A) and §1.430(h)(3)-1(a)(2) are used to determine a multiemployer plan’s current liability for purposes of applying the rules of section 431(c)(6). A multiemployer plan is permitted to apply either the static mortality tables used pursuant to §1.430(h)(3)-1(a)(3) or generational mortality tables used pursuant to §1.430(h)(3)-1(a)(4) for this purpose. However, for this purpose, a multiemployer plan is not permitted to use substitute mortality tables under §1.430(h)(3)-2.