I finally got enough time to do a calculation.
I found a source for median income by education level for full-time, year-round workers so I used it for incomes. I graded to reflect some experience increases in the first 10 years. It was only available with a gender split. I got these numbers (expressed in 2016 dollars):
Males: taxes, $377,000 ; benefits, $374,000 ; IRR, 0.0%
Females: taxes, $261,000 ; benefits, $339,000 ; IRR, 0.9%
The women do better because they have lower incomes so more income is in the 90% and 32% bands, and they live 2 more years, and they have a steeper wage history and benefit a little more from the best-35 rule.
I also tried an “all full-time, year-round” workers wage history with blended mortality. That generated 0.7% IRR.
As you mentioned, I should adjust for DI. I backed out the taxes that explicitly go to the DI fund. My male IRR went up to 0.4% and the female to 1.3%.
(note that these IRRs are in excess of inflation)
There were comments that the boomers made out because they had so much income during lower tax years. Simply charging them 12.4% in every year reduced the IRRs to -0.2% and 0.7%. Not a dramatic difference to me.
(If I look at the OAS taxes only, the impact is even smaller. It turns out that the DI deduction grew in the later years when the total didn’t.)
All these numbers ignore the value of survivor and spousal benefits. And, I assume we’d get higher IRRs for workers who didn’t work full-time year-round for 44 years.
This is pretty much what I’d expect. We’ve had a growing labor force and that has kept IRRs about zero. The positive is less than critics would assume they could have earned if they had invested wisely themselves.
(earnings source is table P-24 here: Historical Income Tables: People )