Change Request for Liabilities

Change Request to the Office of Management and Budget

OMB Control No. 0607-0585

Proposal

We propose to collect additional data by adding two questions to the “Liabilities for Defined Benefit Plans” part 6 of the 2007 F-12 Form (Survey of State Administered Public-Employee Retirement Systems). The new questions provide an estimate of the pension fund’s Covered Payroll and an estimate of the pension fund’s Annual Required Contribution (ARC).

We also propose to change the title, add instructions, and reword the existing five questions of the aforementioned section, so as to simplify them for respondents. The proposed new title is “Actuarial Information for Defined Benefit Plans”. The data collection will apply only to state government systems (a total of 221 systems).

History

The “Liabilities for Defined Benefit Plans” section was added to the F-12 Form for the 2007 survey year. Soon after the new form was mailed out, we received numerous calls from various respondents asking us for further information concerning these questions. There was a lot of perplexity and overall confusion as to what data the Census Bureau was attempting to collect. Once we analyzed the reported liabilities data, we confirmed that there was an array of interpretations to the questions – and therefore, faulty data.

Importance

Both the Bureau of Economic Analysis (BEA) and the Federal Reserve have expressed interest in actuarial estimates of pension liabilities and normal pension costs. These measures are necessary in order to estimate the adequacy of pension plan funding and would help examine the overall position of state governments and the household sector. Furthermore, these data would help to ensure national accounts be made consistent with international guidelines (which both agencies have stated as a long-term goal).

Operational Plan

We plan to collect the same liabilities data – but without confusing the respondents. The additional two questions are asking for figures that are easy to locate in the Comprehensive Annual Financial Report (CAFR). We are asking for figures that GASB already require the pension systems to publish. Also, we have included additional instructions after the title, which directs the respondent to the corresponding schedules in the CAFR. That is to say, where the answers to the survey questions can be obtained. The revised 2008 survey form would be posted where all the other survey forms are posted on the “Government” Website for the U.S. Census Bureau: http://www.census.gov/govs/www/surveyforms.html

Impact

Although the revised liabilities section of the form would have seven questions instead of five, the two additional questions would not increase the respondent burden significantly. The answer to the second question, “Provide an estimate of the pension fund’s Covered Payroll”, can be obtained from the Schedule of Funding Progress. This very same schedule provides the answer to the first question as well. Therefore, there is very little increase in how much time the second question requires from the respondent. The answer to the fourth question, “Provide an estimate of the pension fund’s Annual Required Contribution (ARC)”, can be obtained from the Schedule of Employer Contributions. The instructions on the form direct the respondent to these schedules.

There are a total of 221 state system respondents. The F‑12 survey form entails an average of two hours for the respondent to complete. The revised F-12 liability questions would entail an additional six minutes to complete. The estimate of total burden would increase from 442 hours to 464 hours; 22 hours more than what was previously approved by OMB for the F-12 survey form (clearance #0607-0585).

This proposed data collection would not only help the Census Bureau collect more consistent actuarial data, it would provide the BEA and the Federal Reserve with better data for their estimates of pension liabilities and normal pension costs. Finally, with our proposed additional instructions and rewording of the questions, survey respondents will have a much better understanding of the data that we are trying to collect from them.