Issue 4

Pension versus Pension

The Pension Gap”, a recent Los Angeles Times feature, explores the growth of California’s enormous pension liability over time. One particular piece of interest was a comment on California Highway Patrol pensions changing dramatically in 1999. From the article:

Although all state employees benefited from SB 400, none hit the jackpot quite like the 6,500 sworn officers then on the California Highway Patrol. Previously, their pensions had been calculated by multiplying 2% of their salary times the number of years they worked. SB 400 raised that to 3%.

The article mentions that an officer who retired in 1999 after 30 years of work could collect $62,218 in pensions on average per year, but $96,270 per year if they retired after 1999. Because CalPERS data is accessible via Transparent California we can compare this to other public unions during the same time period.

Median total pension per year of work by retirement year (2014 $)
— California Highway Patrol
— Bay Area Rapid Transit
— Santa Clara County Schools

Taking two other large employers as a comparison it’s clear that the CPH pensions did grow remarkably after 1999. CPH median pension per year of service jumps from $1,917 in 1998 to $3,269 in 2009. An average employee that worked 30 years and retired in 1998 would take home $57,510 (1,917 × 30), but if they retired in 2009 they would receive $98,070 for approximately the same work. This roughly matches the figures in the LA Times article. The slight differences are explained in the methodology section below.

Because these are all 2014 payment figures from CalPERS they are directly comparable and don’t require inflation adjustment. The chart can be read as CalPERS liabilities changing over time or a charting of ideal retirement years by employer. The reader may want a time machine for the latter.

Taking all large Bay Area employers with at least 25 retirees per year in 1998 and 2014 we can calculate the difference in pensions between those years. Also included are the latest figures for 2014 pension payments. The employers with the largest changes aren’t necessarily the highest payers.

Normalizing the data such that employers and their pensions can be compared reveals all kinds of trends. Any particular line is the result of changing laws, contracts, job composition, and external forces. This post explores only a small subset of the total data, so there’s plenty more to investigate. The LA Times article discusses tax payer liability relative to this data and is well worth a read.

Thanks for reading. 🙏 to Brian Smith once again for his help with code, math, and words. See you in a couple weeks.

— Josh

Chart info & methodology

Why the weird unit? The raw data comes in something like this:

It’s impossible to compare these people because their pensions depend on three variables: their employer, their year of retirement, and the number of years they were employed. To normalize this data, pension is divided by years worked to get a single unit that can be compared.

Now we have a unit that can be compared directly for analysis. In our case retirees are separated by employer then a median is calculated for each year of retirement.

Source data that met any of the following were removed to simplify analysis:

  • Less than 1 year of service
  • Total pension included settlement payout
  • Retired before 1990
  • Employer not in Bay Area (except CHP)

In addition, employers were not included if there were less than 25 pensions to derive a median from.

46.87 is a regular exploration of San Francisco Bay Area data by Joshua Jenkins in partnership with Numeracy.