Our method for pension system studies
Pension benefits enter the model as household income, not industry output. We start from benefit payments by retiree residence, usually ZIP code, and subtract federal and state income taxes and any health premiums deducted from the check to estimate disposable income. That income is assigned to IMPLAN household income groups, which set how much is spent and on what. Retirees who have moved out of the study region spend their benefits elsewhere, so those payments are modeled where the retiree lives or reported separately.
Because the shock is household spending, the results are the jobs and income supported by retiree purchases and the supply chains behind them. The system's own administrative and investment management spending is a separate operating stream. Employer contributions fund benefits paid later, so we take care not to count the same dollars once as contributions and again as benefits.
The results are a gross contribution. Without the plan, members would have had some other retirement income, and the study does not net that out unless the board asks for a defined alternative scenario. District results come from mapping retiree ZIP codes to legislative boundaries.
Typical inputs we ask for
- Annual benefit payments and payee counts by ZIP code of residence
- Payments to out-of-state retirees, by state
- Typical tax withholding and health premium deductions, or plan-level averages
- Administrative budget and payroll
- Investment management, custody and advisory fees paid, with firm locations
- The district maps you want results reported for
Questions clients ask
Can results be broken out by legislative district?
Yes. Our LACERA study reported results for districts within Los Angeles County, based on where retirees live. IMPLAN publishes congressional district models built from ZIP codes. For state senate and assembly districts we allocate ZIP-level results to district boundaries, and where a ZIP code is split between districts we apportion it and explain the method in the report.
Does the study claim the economy would lose all of this activity without the pension?
No. The study measures the economic activity that benefit spending supports today. If the plan did not exist, members would have had some other form of retirement income, and some of this spending would still happen. We state that clearly. If your board wants a net comparison, we define the alternative with you in writing before modeling it.
Do you need member-level data?
No. We can work from ZIP-level totals of payments and payee counts prepared by your staff, so names, member IDs and Social Security numbers never leave the system. Out-of-state payments can be summarized by state. Aggregated data is enough for the household spending model and for district mapping.
Related reading
Sources
- IMPLAN, Understanding Congressional District Data (congressional district models built from ZIP codes) Source
