The authors take issue with the practice of cross-testing defined contribution plans, and especially with new comparability plans. Cross-testing offers firms a method for making larger contributions for highly paid employees, so long as they are older than other employees participating in a plan. Variations on cross-testing permit permit some firms to provide lower contribution rates to older rank-and-file employees and to provide high contribution rates to all higher paid employees. Plans using such variations are often referred to as new comparability plans.

 

The authors argue that cross-testing should be rejected, if not universally, prohibited. They contend that weighting benefits for higher paid employees in traditional defined benefit plans can often be justified by various types of risk shifting provided by the defined benefit format. These risks include mortality, investment, and sometimes what the authors refer to as compensation-escalation risk, which is the risk that the employee will not be able to replace an adequate share of pre-retirement income if the employee receives late career compensation increases. Defined contribution plans do not generally offer risk shifting of these types.

 

The authors note that many small defined benefit plans also fail to shift risk of the types described but argue that this does not mean that cross-tested defined contribution plans should therefore be permitted. The authors also note that establishing a small defined benefit plan imposes costs on the employer and thus discourages some employers from adopting them. Moreover, there are other important distinctions between defined benefit and defined contribution plans. In addition, the authors argue that small defined benefit plans should be prohibited where they will not shift risk. The authors suggest that the qualification condition that a plan be permanent would screen out many non-risk shifting small defined benefit plans. The authors also note that permitting a firm to sponsor both a defined benefit plan and a cross-tested defined benefit plan can be viewed as inconsistent with the separate section 415 limits for defined contribution and defined benefit plans.

 

The authors defend the Department of Treasury's regulations limiting new comparability plans but suggest that the regulations do not go far enough in limiting cross-testing. On a more general level, the authors express skepticism that the nondiscrimination rules can bear the weight of the social good they are supposed to promote: retirement savings for lower and middle income people who otherwise would save inadequately for retirement. They thus agree with Professor Zelinsky that some type of safe-harbor approach might result in higher benefits for rank-and-file employees than the current nondiscrimination rules. They suggest a reverse match type of approach, in which the firm would make an initial contribution to the plan, which the employees could then match in some statutorily determined multiple.