Thanks for this very informative piece. I appreciate your efforts to educate readers as to the true state of retirement savings in this country. But good luck getting your voice heard above the noise. Today’s Barron’s cites the National Institute on Retirement Security as saying that “about 40% of older Americans rely on Social Security for their ENTIRE income in retirement”. Readers are sure to see that as a crisis.
This is what the facts about retirement are up against. In fact, so few retirees are 100% dependent on Social Security that no one really bothers measuring it anymore. The last decent study I can find is from SSA in 2008. At that time, 4.8% of seniors received all of their income from Social Security, roughly 1/10th the number that NIRS claims. Their study is so bad that you can prove the falseness -- falsity? -- of their claims from their own numbers. Even the retirees NIRS claims are 100% dependent on Social Security are shown to have income from non-Social Security sources. It's really bad. But quotable!
Definitely true. But at least the CPS is consistent! Americans aren't saving in retirement plans, the CPS would have you believe, which explains why they don't get much income from retirement plans...
It's really unfortunate that the leading source of income data is so poor in its handling of retirement savings.
Hi Andrew, not sure if you are aware of another study that compares CPS and HRS with matched data, although only CPS is linked to IRS data. It came out as an SSA working paper first, but it was later updated and published last year at JPEF, DOI: https://doi.org/10.1017/S1474747224000039
It does not change your story, just another data point (2016 CPS), although not more recent.
I want to continue with that work but data access is the main issue. If Census would share their IRS data with SSA researchers, I am sure we will have these statistics produced in no time.
Thanks for this very informative piece. I appreciate your efforts to educate readers as to the true state of retirement savings in this country. But good luck getting your voice heard above the noise. Today’s Barron’s cites the National Institute on Retirement Security as saying that “about 40% of older Americans rely on Social Security for their ENTIRE income in retirement”. Readers are sure to see that as a crisis.
This is what the facts about retirement are up against. In fact, so few retirees are 100% dependent on Social Security that no one really bothers measuring it anymore. The last decent study I can find is from SSA in 2008. At that time, 4.8% of seniors received all of their income from Social Security, roughly 1/10th the number that NIRS claims. Their study is so bad that you can prove the falseness -- falsity? -- of their claims from their own numbers. Even the retirees NIRS claims are 100% dependent on Social Security are shown to have income from non-Social Security sources. It's really bad. But quotable!
https://www.ssa.gov/policy/docs/ssb/v67n2/v67n2p65.html
To your point, the shortcomings (more recently, it's been noted that it also undercounts retirement plan participation - see https://www.ebri.org/content/retirement-plan-participation-and-the-current-population-survey-the-trends-from-the-retirement-account-questions) in the CPS have been well-known for some time (at least among academics) - to the point where you'd think any reference to it in retirement income assessments should come with a big *. But then, that doesn't feed the retirement crisis narrative...
Definitely true. But at least the CPS is consistent! Americans aren't saving in retirement plans, the CPS would have you believe, which explains why they don't get much income from retirement plans...
It's really unfortunate that the leading source of income data is so poor in its handling of retirement savings.
ha - so if you can't be right, be consistent? All the more reason for the *.
Thanks! Amazing!
Hi Andrew, not sure if you are aware of another study that compares CPS and HRS with matched data, although only CPS is linked to IRS data. It came out as an SSA working paper first, but it was later updated and published last year at JPEF, DOI: https://doi.org/10.1017/S1474747224000039
It does not change your story, just another data point (2016 CPS), although not more recent.
I want to continue with that work but data access is the main issue. If Census would share their IRS data with SSA researchers, I am sure we will have these statistics produced in no time.