IRS ISSUES PROPOSED REVENUE PROCEDURE FOR NONDISCRIMINATION TESTING.

Announcement 92-81

This announcement issues, in proposed form, a revenue procedure relating to the nondiscrimination provisions of Internal Revenue Code sections 401(a)(4), 410(b) and certain related provisions. This revenue procedure is being issued in proposed form to permit public comment on its provisions.

Announcement 92-29, released February 11, 1992, extended the effective dates applicable to the nondiscrimination regulations for qualified plans and requested comments or suggestions for facilitating compliance with these regulations.

The attached proposed revenue procedure is based on comments received and ideas suggested by employers and practitioners on data collection and related testing issues. It provides guidelines that simplify the process of substantiating compliance with the nondiscrimination rules. Specifically, the proposed revenue procedure simplifies the substantiation process in the following significant ways:

o QUALITY OF DATA. Employers that do not have precise data available at reasonable cost may substantiate compliance with the nondiscrimination requirements using a reliable substitute.

o SINGLE DAY "SNAPSHOT" TESTING. Employers may substantiate compliance with the nondiscrimination requirements using "snapshot" testing on a single representative day.

o SPECIAL RULE FOR IDENTIFYING HIGHLY COMPENSATED EMPLOYEES. Employers may use a simplified method for determining highly compensated employees to facilitate "snapshot" testing.

o THREE-YEAR TESTING CYCLE. Employers will not be required to test a plan more than once every three years provided there is no significant change.

Since publication of the nondiscrimination regulations, the Service and the Treasury have been actively considering the comments that have been received on potential areas for improvement or clarification and have been working with affected taxpayers and practitioners. The proposed revenue procedure issued today focuses on comments about data collection and substantiation which many employers and practitioners identified as one significant area of concern.

As part of the ongoing effort to reduce taxpayer burden, the Service and the Treasury intend to issue additional guidance relating to the nondiscrimination regulations as expeditiously as possible. It is anticipated that this guidance will focus on increased access to safe harbors and greater flexibility in general nondiscrimination testing. Among other things, it is anticipated that this guidance will modify the service crediting and compensation regulations to accommodate leaves of absence and employee transfers between plans and between employers (including situations involving mergers and acquisitions, participation in joint ventures, and transfers within an industry group).

Proposed Revenue Procedure

SECTION 1. PURPOSE

This revenue procedure provides guidelines for nondiscrimination testing that reduce the administrative burdens of substantiating compliance. Specifically, this revenue procedure simplifies the process of substantiating compliance with the nondiscrimination requirements in the following significant ways:

o QUALITY OF DATA. Employers that do not have precise data available at reasonable cost may substantiate compliance with the nondiscrimination requirements using a reliable substitute.

o SINGLE DAY "SNAPSHOT" TESTING. Employers may substantiate compliance with the nondiscrimination requirements using "snapshot" testing on a single representative day.

o SPECIAL RULE FOR IDENTIFYING HIGHLY COMPENSATED EMPLOYEES. Employers may use a simplified method for determining highly compensated employees to facilitate "snapshot" testing.

o THREE-YEAR TESTING CYCLE. Employers will not be required to test a plan more than once every three years provided there is no significant change.

SEC. 2. OVERVIEW

01 This revenue procedure provides practical and flexible guidance for purposes of substantiating that a plan is nondiscriminatory as to coverage under section 410(b) of the Internal Revenue Code, as to both the amount of contributions or benefits and the current availability of benefits, rights, and features under section 401(a)(4), and for purposes of determining whether a definition of compensation is nondiscriminatory under section 1.414(s)-1(d)(3) of the Income Tax Regulations (collectively referred to as the nondiscrimination requirements for purposes of this revenue procedure). Unless specifically provided otherwise, this revenue procedure does not apply to the special nondiscriminatory amounts tests under section 401(k)(3) or section 401(m)(2).

02 The guidance in this revenue procedure is designed to allow employers to use alternative methods for substantiating compliance with the nondiscrimination requirements. The Internal Revenue Service will treat a plan as satisfying the nondiscrimination requirements where an employer shows compliance under these alternative methods because these methods are intended to demonstrate whether there is a high likelihood that the plan satisfies the nondiscrimination requirements. The guidance in this revenue procedure should be interpreted in a reasonable manner consistent with this purpose.

03 Section 3 of this revenue procedure provides standards for determining the data that is appropriate for testing. Section 4 provides standards for identifying the population of employees who must be tested. Section 5 provides special rules for determining the highly compensated employees. Section 6 provides the standards under which an employer can rely on a prior year's test.

SEC. 3. QUALITY OF DATA

01 If an employer maintaining a plan does not have precise data available at reasonable expense, then the employer may substantiate that the plan satisfies the nondiscrimination requirements using less than precise data if (1) the data being used is the best data available for the plan year at reasonable expense and (2) the employer reasonably concludes that demonstrating compliance with the nondiscrimination requirements using this data establishes a high likelihood that the plan would satisfy the nondiscrimination requirements using precise data. Data that satisfies this section 3.01 is referred to in this revenue procedure as "substantiation quality data."

02 In a defined contribution plan, an employer usually obtains precise data for making allocations to accounts and will have this precise data available for the plan year. Thus, to the extent that the data obtained for allocations is relevant to testing for nondiscrimination, the substantiation quality data that the employer will have for the plan year to test a defined contribution plan generally will be precise data. However, data as to allocations generally is not relevant for testing a defined contribution plan that satisfies section 401(a)(4) of the Code using a design-based safe harbor under section 1.401(a)(4)-2(b)(3) or 1.401(a)(4)-8(b)(3) of the regulations. Thus, for example, in the case of such a design- based safe harbor plan, for purposes of demonstrating compliance with section 410(b), substantiation quality data need not include data obtained for allocations.

03 The following examples illustrate principles in this section. Assume for purposes of these examples that the plan uses a definition of compensation that satisfies section 414(s) of the Code.

EXAMPLE 1. Employer A sponsors a defined benefit plan for the employees in Divisions X, Y, and Z, which are located in different geographic areas. Each division separately maintains its own payroll and personnel records. Under Employer A's plan, the relevant components for determining the employees' accrual rates are compensation, age, and service. Because of the cost of gathering and merging precise data for all employees in each division, Employer A determines that for 1995 the best data that is available at reasonable expense is the data that is gathered as of the 1995 valuation date and used for preparing the Schedule B of the 1995 Form 5500. The valuation data includes each participant's rate of pay, date of birth, and date of hire. Using the valuation data and reasonable assumptions to construct compensation history, Employer A's plan passes the tests for determining whether the plan meets the nondiscrimination requirements. Because Employer A reasonably expects this data to approximate the relevant compensation, age, and service components, Employer A reasonably concludes that there is a high likelihood that the plan also would satisfy the nondiscrimination requirements if it were tested using precise data.

EXAMPLE 2. The facts are the same as in Example 1, except that the data gathered for Division Y is missing the date of hire for a group of Division Y employees that is a small percentage of the total employees in the plan. An estimate was made for these employees' years of service on the basis of the average number of years of service for other employees in Division Y of the same age. The estimation of the missing data does not preclude Employer A from reasonably concluding that there is a high likelihood that the plan would satisfy the nondiscrimination requirements using precise data. If, however, Employer A were missing relevant data on one or more of the most highly compensated employees in the plan, or on a group of employees subject to a benefit formula that generally was not available to the other employees, then Employer A would be required to undertake more scrutiny of the underlying information to support the reasonableness of its conclusion.

SEC. 4. SINGLE DAY "SNAPSHOT" TESTING

01 An employer may substantiate that a plan complies with the nondiscrimination requirements on the basis of the employees present on a single day during the plan year (snapshot day), provided that day is reasonably representative of the employer's workforce and the plan's coverage throughout the plan year. A snapshot day will not be treated as failing to be reasonably representative solely because of a significant change in the employer's workforce caused by an extraordinary event, such as a merger or acquisition. The snapshot day for a plan generally must be consistent from year to year.

02 Under this section 4, the employees to be tested for purposes of substantiating compliance are solely the individuals who are employees on the snapshot day (snapshot population).

03(1) The employer tests the amount of benefits or contributions and the current availability of benefits, rights, and features using substantiation quality data for the snapshot population. Generally, it will be practical for the employer to select a snapshot day that is the same day for which the employer has substantiation quality data. For example, if an employer determines that the valuation data for a defined benefit plan is substantiation quality data, it would be practical for the employer to select the valuation date as its snapshot day, provided that day is reasonably representative, so that in collecting its valuation data the employer simultaneously collects substantiation quality data for all the employees in the snapshot population.

(2) An employer need not, however, select a snapshot day that is the same day as of which substantiation quality data is available. For example, if an employer were using a January 1 snapshot day for its defined benefit plan, the employer might also select January 1 as the snapshot day for its defined contribution plan, provided that day is reasonably representative. Under section 3, if the defined contribution plan is a safe harbor plan under section 1. 401(a)(4)- 2(b)(3) or 1.401(a)(4)-8(b)(3) of the regulations and need only determine the number of employees who benefit under the plan, the employer's data as of the snapshot day may constitute substantiation quality data. In other cases, the employer generally will not have substantiation quality data for a defined contribution plan until the end of the plan year, when the employer calculates allocation amounts. In the latter case, the snapshot population could be established as of the beginning of the year for the defined contribution plan, but whether the plan meets the nondiscrimination requirements would be determined as of a date later than the snapshot day, when the substantiation quality data for the snapshot population is available to the employer.

04 If a defined benefit plan has a minimum service requirement that would preclude eligible employees from actually accruing benefits for a plan year, the application of section 410(b) of the Code to the snapshot population may overstate the plan's coverage under section 410(b). For example, where a defined benefit plan has a provision requiring that an employee have 1,000 hours of service to accrue benefits, and the snapshot day is early in the plan year, employees who terminate after the snapshot day may have been treated as benefiting who do not in fact benefit. Similarly, if the snapshot day is later in the plan year, employees who terminate prior to the snapshot day and who fail to benefit because of the service requirement will not be reflected. Therefore, if employees terminate in a defined benefit plan with a minimum service requirement and the substantiation quality data or snapshot population does not reflect the level of turnover, an adjustment must be made to the test that reasonably compensates for the difference between relative turnover rates among eligible highly compensated and nonhighly compensated employees The employer must account for the effect of the minimum service requirement by making an appropriate adjustment to the ratio percentage or nondiscriminatory classification percentage of section 410(b) to reflect terminations. For this purpose, an adjustment of 5 percent (i.e., 70 percent becomes 73.5 percent) for a 1,000 hour rule will be treated as a safe harbor.

05 Similarly, if a defined contribution plan has a minimum service requirement, such as a provision that limits allocations solely to those employees who are employed on the last day of the plan year, the application of section 410(b) of the Code to the snapshot population may overstate the plan's coverage. Where a defined contribution plan's substantiation quality data does not include actual allocation amounts as of the end of the year, however, the effect of a minimum service requirement and the need for any corresponding adjustment will depend upon whether the snapshot population reflects the level of turnover of eligible employees. If a defined contribution plan with a "last day" or similar service requirement has a snapshot day in the first quarter of the plan year and the plan's substantiation quality data includes actual allocation amounts, then the snapshot population will be treated as accurately reflecting the plan's coverage. In other cases where the snapshot population does not reflect the level of turnover, an adjustment must be made as described above in section 4.04. In addition, an adjustment to section 410(b) of 10 percent for a "last day" rule (i.e., 70 percent becomes 77 percent) will be treated as a safe harbor.

06 Under this section 4, it is intended that all plan provisions that are in effect during the plan year be tested on the basis of the snapshot population. In certain cases, however, a plan may include a provision that must be tested in that plan year but that primarily affects employees who were employed prior to the snapshot day and who are not in the snapshot population. For example, the snapshot population would not reasonably represent the employees for whom an early retirement window is available if the window was opened during the year and significant numbers of employees retired under the window prior to the snapshot day. In such case, the effect of the early retirement window is tested on the basis of a snapshot population that is reasonably representative of the employer's workforce when the window opens, such as the prior year's snapshot population if it is reasonably representative.

07 If an employer is substantiating whether a definition of compensation is nondiscriminatory, the employer tests the definition of compensation on the basis of substantiation quality data for the relevant employees in the snapshot population. Thus, for purposes of section 1.414(s)-1(d)(3) of the regulations, the employer compares the substantiation quality data for the definition of compensation being tested to the substantiation quality data for total compensation.

08 Although this revenue procedure does not apply for purposes of satisfying sections 401(k)(3) or 401(m)(2) of the Code, plans with salary reduction, matching, or after-tax features may be tested under this section 4 to substantiate compliance with the other nondiscrimination requirements (e.g., coverage or current availability of benefits, rights, and features).

SEC. 5. SIMPLIFIED IDENTIFICATION OF HIGHLY COMPENSATED EMPLOYEES FOR SINGLE DAY "SNAPSHOT" TESTING

01 In conjunction with substantiating that a plan satisfies the nondiscrimination requirements under this revenue procedure, an employer may determine, as of the snapshot day described in section 4, which employees in the snapshot population are highly compensated employees under section 414(q) of the Code. An employee in the snapshot population is a highly compensated employee if, as of the snapshot day: (1) the employee is a 5-percent owner; (2) the employee's compensation for the plan year exceeds the section 414(q)(1)(B) amount (indexed to $93,518 for 1992); (3) the employee's compensation exceeds the section 414(q)(1)(C) amount for the plan year (indexed to $62,345 for 1992) and the employee is in the top paid group of employees within the meaning of section 414(q)(4); or (4) the employee is an officer described in section 414(q)(1)(D).

02 The compensation used to determine whether an employee is a highly compensated employee under section 5.01 is compensation that reasonably approximates the employee's section 414(q)(7) compensation for the plan year. If the determination of who is a highly compensated employee is made earlier than the last day of the plan year, the employee's compensation that is used to determine an employee's status must be projected for the plan year under a reasonable method established by the employer. For purposes of item (3) in section 5.01, the number of employees in the "top paid" group of employees is determined on the basis of the employees in the snapshot population.

03 If an employer is substantiating that a plan satisfies the nondiscrimination requirements using this revenue procedure and the plan must also satisfy the "ADP" test in section 401(k)(3) of the Code or the "ACP" test in section 401(m)(2), there may be individuals not employed on the snapshot day that need to be categorized as either highly compensated or nonhighly compensated employees for purposes of these tests. In that case, the employer may use the option described in this section, subject to the following modifications. In addition to those employees who are determined to be highly compensated on the plan's snapshot day, as described above, the employer must treat as highly compensated any eligible employee for the plan year who:

(1) terminated prior to the snapshot day and was a highly compensated employee in the prior year;

(2) terminated prior to the snapshot day and had compensation for the plan year greater than or equal to the projected compensation of any employee who is treated as highly compensated on the snapshot day (except for employees who are highly compensated solely because they are 5-percent owners or officers); or

(3) becomes eligible subsequent to the snapshot day and has compensation for the plan year greater than or equal to the projected compensation of any employee who is treated as highly compensated on the snapshot day (except for employees who are highly compensated solely because they are 5-percent owners or officers).

04 In applying this section 5, section 1.414(q)-1T of the Temporary Income Tax Regulations applies to the extent that it is not inconsistent with the method provided above.

SEC. 6. THREE-YEAR TESTING CYCLE

An employer may rely for the two succeeding plan years on the tests substantiating that a plan complies with the nondiscrimination requirements for a plan year if the employer reasonably concludes that there are no significant changes subsequent to the test (e.g., changes in plan provisions, the employer's workforce, or compensation practices). For this purpose, whether a change is significant depends upon the relative margin by which the plan has satisfied the nondiscrimination requirements in the most recent year in which the plan was tested and the likelihood that the change would eliminate that margin. If there is a significant change in one plan provision, the effect of which can be isolated from the effect of other provisions, the employer may continue to rely on the prior test during the interim two years, provided that the employer can demonstrate that the effect of the amended plan provision is nondiscriminatory.

SEC. 7. EFFECTIVE DATE

This revenue procedure is issued in proposed form in order to permit comment. Employers may, however, rely on these substantiation guidelines while they are proposed. Employers using the three-year testing cycle described in section 6 for purposes of substantiating compliance should treat the year in which the final regulations under sections 401(a)(4) and 410(b) of the Code become effective as a year of significant change requiring actual testing.

COMMENTS

The Service invites comments with respect to this proposed guidance. Comments should be submitted in writing, referencing Announcement 92-81, and addressed to:

                                   Associate Chief Counsel