STUMP » Articles » Calpers: What Exactly is an "Average" Pension? » 30 November 2016, 07:26

I follow several major pension funds on twitter (like ya do), and I saw the following tweet from Calpers yesterday, and responded:

Are public #pension benefits excessive? Hear the truth from state #retirees via CapRadioNews</a> <a href="https://t.co/G2MKlYG8ZB">https://t.co/G2MKlYG8ZB</a></p>&mdash; CalPERS (CalPERS) November 29, 2016

The average CalPERS #pension is about $31,500 per year. Get the facts: https://t.co/B21xJZOC0j

— CalPERS (@CalPERS) November 29, 2016

CalPERS</a> Is that over all CalPERS recipients, or does it involve only those who worked 30+ years in the system?</p>&mdash; Mary Pat Campbell (meepbobeep) November 29, 2016

meepbobeep</a> That’s the average pension benefit payment of retirees, beneficiaries, and survivors across all years, as of June 30, 2015.</p>&mdash; CalPERS (CalPERS) November 29, 2016

PLAYING THE AVERAGES TO SKEW RESULTS

There’s a reason I asked that, obviously. I learned from the ancient book How to Lie with Statistics about just how deceptive averages can be.

Let me remind you of the following graphs I made of info on the Illinois State University Retirement System, and in particular this graph showing average pension by range of years of service:

My SURS data are here.

The red bars show the average pension for a particular range of years of service, and unsurprisingly, the average generally increases the more years of service one has. It drops after 40 years, but that’s mostly due to very few people having pensions with that many years of service… for all I know, that’s mainly janitors in that bucket, and the lower buckets are mainly professors and admins.

So the average pension, at 35 – 40 years of service (what I’d call a full career), is near $72,000. Pretty nice, given that the average income for households for those age 65 and older was about $56K in 2014, and the median income was about $37K. (Info from the Current Population Survey)

What was the average SURS pension in 2014 overall, by the way?

$34,411 in 2014.

Quite a bit less than $72K, eh?

Heck, the median is even lower than that, as over half of SURS pensioners/benefiaries had less than 20 years of service in the system, and the average at 20 years isn’t even $30K.

When you have a lot of people who are in a pension fund where it’s not their whole career, but only partly, then the overall average is going to be a lot less than the average for someone who works a full career in that particular system.

GETTING AT A MORE REPRESENTATIVE STATISTIC

Thank goodness, Transparent California stepped in and pointed me to some useful info:

meepbobeep</a> We have 30+ averages by employer here: <a href="https://t.co/UQaYU395KL">https://t.co/UQaYU395KL</a></p>&mdash; TransparentCal (Transparent_CA) November 29, 2016

Here is pension data, by employer, where Tansparent California broke out the 30+ year pensions vs. the overall.

And as they mention, the average for overall vs. the full career pensions are quite different:

meepbobeep</a> if you strip out beneficiaries, its $32,678 w/ 19.8 years of service. For &gt;=30 years of service its $65,545.</p>&mdash; TransparentCal (Transparent_CA) November 29, 2016

The full-career pension is about twice that of the whole system. I’m not claiming that a $66K average pension is excessive — heck, I wouldn’t even say a $100K+ pension is necessarily excessive — but I am saying that using the average that includes those who worked only a partial career in Calpers is being extremely misleading and Calpers knows it.

I highly recommend Calpers use something less deceptive if they want to claim Calpers pensions aren’t excessive. This is “How to Lie with Statistics 101”, and if you want to be persuasive, this is not the way to go about it.

Shame, Calpers, shame.

(Maybe y’all should stick to quoting replacement ratios… but I bet those are pretty high, too)

GRAPHING COMPARISONS OF AVERAGES

Since Transparent California was so nice to make a table showing enough info that I can do comparisons of average by employer, let’s do it!

If you look at my twitter feed from today, you’ll notice I went through several iterations of the graph I’m about to show you.

Here it is:

What I did with the underlying data:

I filtered out employers with fewer than 100 retirees/beneficiaries. The problem there is that too many of those have no full-career people, and besides, with such small sample sizes, the averages may not be very “stable”.

Then I did a linear fit, and forced the intercept to be zero, just to get a general rule-of-thumb.

Turns out, the average for full career pensions, no matter the employer, is a little less than twice the overall average for that employer.

So, maybe that is excessive or not. That’s not what I care about. I do care about looking at meaningful numbers when one is examining evidence.

Spreadsheet with the CalPERS info is here.