Pension Regulators Need Data

The following post reflects the frustration of me and my colleagues after having worked for three months to model the finances of a large private sector multiemployer pension plan.  The analysis was initiated in response to a contentious proposal that would enable plan sponsors and trustees to reduce accrued benefits for both retirees and actives in order to avoid their plans becoming insolvent.  One stipulation of the proposal is that these cuts would be sanctioned only if they allowed the plan to avoid insolvency.  In other words, plans would not be permitted to use this new tool if the cuts only prolonged the life of the plan for a few years.  If such a proposal were enacted, the relevant regulator – the Pension Benefit Guaranty Corporation (PBGC), the Department of Labor (DOL), or both – would need some mechanism to figure out whether cutting benefits would achieve the goal of long-term solvency.  The problem is that these regulators – like us – do not have access to the necessary data for this type of analysis.  

Sponsors of both single-employer defined benefit plans and multiemployer plans hold information regarding their plans close to their vest.  The plans are required to have regular actuarial assessments, but it is extremely hard to get a hold of the actuarial reports summarizing these assessments.  We recently tried to get actuarial reports for about 800 troubled multiemployer plans and were able to collect reports for only about ten.  We haven’t tried a similar exercise for single-employer plans, but I suspect it would be just as difficult. 

The plan we were trying to model did make its actuarial report available.  But data in the report alone were insufficient.  The plan had about seven different benefit options, and it was impossible to figure out which employees were entitled to what. 

What we needed were data for each individual in the plan — their age, tenure, and what benefit option they were entitled to.  Then we would not have had to spend weeks trying to match liabilities for active workers, separated workers, and retirees.  The plan data could be stripped of any identifying information such as Social Security number, so no question arises regarding invasion of privacy. 

The same need applies to public sector defined benefit plans.  Since the employers are public entities, the actuarial reports for their plans are readily available.  But it is virtually impossible to get data at the individual level, which is essential when trying to project the trajectory of future benefit payments. 

If this issue reflected simply the data needs of the Center for Retirement Research, it would not be worthy of even a blog.  But neither the DOL nor the PBGC has regular access to actuarial reports and never to plan data at the individual level.  These regulators have to rely on summary information provided in the DOL’s Form 5500.  This form provides a nice snapshot of the plan’s current funded status, but does not provide any of the material necessary to project where the plan is headed.  The firm’s actuaries are required under the Pension Protection Act of 2006 to model a plan’s future and classify the plan as endangered or critical if it faces impending insolvency.  But any forecast involves a myriad of small assumptions that can have a material effect on the outcome.  There is no substitute for doing the projections oneself.

This brings us back to the proposal to save severely troubled multiemployer plans by cutting accrued benefits.  Such a proposal clearly violates one of the basic tenets of pension law, but has the possibility to provide benefits to many workers who otherwise would receive nothing.  However the debate turns out, we know one thing for sure: any plan requesting permission to make such cuts should be required to turn over all its plan data to actuaries at the PBGC, DOL, or both.  In fact, these regulators should have regular access to actuarial reports and plan data.