Nice post, though I wish you (and Equable and others) would point out that not only are reported asset values almost certainly overstated, but the liabilities are most definitely significantly understated - probably a bigger problem than the former. 85% is probably more like 65-70% when one stops pretending that fundamental principles of finance somehow don't apply to the valuation of pension liabilities.
Other than the assets and liabilities, though, everything's fine.
When one sees "think tanks" that probably know better continue to ignore this issue in what appear to be comprehensive (or at least very long) analyses, one must wonder what source of funding or influence or other benefit to them would be at risk if they addressed the issue honestly instead of utterly ignoring it. They're not doing enough thinking in their pension finance tank.
Kudos to Oliver Giesecke and Hoover for being virtually alone in honestly examining this aspect of public pension funded status: https://publicpension.stanford.edu/.
I had a whole section about calculating the funded ratio, but I deleted it, because this post was getting to be too long in the first place (and it really needs to be its own standalone piece). I’ve been writing on these things online in semi-permanent places since 2014, so I don’t feel like cramming all the relevant points in a single post.
I do want to do something with Stanford's new dashboard/tools, because they're really nice -- and yes, as you know, I'd rather use a different valuation rate altogether.
But let me get back to the original point: if you use their standard approaches, how useful is it if the funded ratios can (and do) drop 30 percentage points in 2 years? What does that tell you about the pension systems? How good does that make one feel about 85% funded ratios?
Well, actually, without using a meaningful liability, funded status and year-to-year changes are misleading, close to meaningless. One should feel bad about 85% not because it can go down, but because it's wrong and the real number is much lower (and, secondarily, might go down further).
In calendar year 2022, when market interest rates increased substantially, causing market liabilities and bonds to decrease, and risky assets like stocks also decreased, reported funded status for public plans got much worse, but private sector funded status based on quasi-market liability measurements actually improved slightly because they reflected the impact of market interest rates on liabilities. (Check out the Milliman indices for each during calendar year 2022).
If you’re looking at measures of funded status and comparing reported non-market liabilities to market assets (though a good chunk are mismarked, to your point), why would you think the results have meaning? The reported assets and reported liabilities are real(ish) apples and wax oranges.
Yes, the fact that maybe 30% of assets are overmarked by probably 10-20% is a big problem, but isn't the fact that 100% of liabilities (a bigger number) are understated by 20-30% a much bigger problem? Focusing on only the assets is like a news outlet spending all its efforts reporting on a single murder and ignoring a mass shooting. Yet that's what everyone does when it comes to public pensions.
Continuing to dutifully report and analyze actuarial numbers undermines the credibility of CRR's Public Plans Database, Equable, and others, no matter how pretty the graphs, how extensive the (bad) data, how much work goes into compilation, or how lengthy and sophisticated-seeming the reports. It's astrology claiming to be astronomy.
This issue needs to be called out, not reinforced.
Sorry for ranting. It's just very disappointing to see these analyses reported on like there's nothing wrong with them. I hope you do get around to a separate post on that topic. You're an important voice. Stanford/Hoover would be a good starting point - not because their graphics are conceived and executed well, but because the numbers are realistic.
I don't mind you going on the rant -- after all, I agree with you re: valuation...
HOWEVER
This is a "hobby" for me (in my copious free time), so I'm not going to be spending a lot of time coming up with new numbers. If other databases are developed with the equivalent of LDROM for all the years... yay! I'll use that info. I'm not recalculating all this stuff myself.
If the Stanford numbers are downloadable, I can do stuff with that. [And yes, the -- "what's wrong with the funded ratio?" would be a good followup]
ALL THAT SAID, there is a reason I came up with my extremely simplistic cash flow model (which I need to update): all these balance sheet shenanigans, whether asset- or liability-side, are very abstract. The ultimate goal is cash flows to pension benefits.
So let's just do that.... (and see the money run out... or not) -- I know the actuarial models that project using various assumption sets could be used to do far more sophisticated stress tests/sensitivity analyses.... and perhaps I'll get Gemini to mock something up for me.... But that's for another time ;)
Funded ratios are just a crude point-in-time metric, whereas looking at cash flow simulations and failure (money runs out), sustainability (under certain conditions), etc. is a more general risk management perspective. I like looking at failure modes -- what would it take to break the system? Some of the plans are obviously fragile right now, but what about all those 85%ers? I can think of all sorts of "fun" scenarios....
Thanks! I understand and would never expect you to recalculate numbers. And as I said, you're an important voice on this topic - your hobby is productive and a public service.
The big problem I see is that "pension debt" as reported understates the real debt, under any meaningful and understood definition of that term, by $2-3 trillion. All the analyses regarding debt, funded status, and the sources of change thereof, are worth very little if they don't reflect market liabilities.
Stanford/Hoover being the exception, the reported data, while copious and accessible, are highly misleading because they are based on what actuaries and accountants misleadingly label "liability." Why do these groups fail to acknowledge this problem? CRR, for one example, is likely conflicted by its affiliations with NASRA and GFOA, both of which endorsed and may have contributed to (along with their actuarial collaborators) the ASOP 4 toolkit (https://www.ncpers.org/asop-4-toolkit-pensions-ldrom), which, to be frank, is unconscionably misleading. I worry that those types of collaborations mean that many widely cited groups will never fully report what they should if they want to further a meaningful discussion.
Writing about and analyzing and publicizing all the bad data because that's what's available -- without noting how misleading they are in the context of a discussion on pension funding -- is, in my opinion, an example of the "Streetlight Effect" (https://en.wikipedia.org/wiki/Streetlight_effect), where the drunk who lost his keys in the park across the street looks for them under the streetlamp because the light's better. It contributes to an illusory truth effect around public pension funding by "flooding the zone." I think it's very important to call this out in reporting on these numbers.
Nice post, though I wish you (and Equable and others) would point out that not only are reported asset values almost certainly overstated, but the liabilities are most definitely significantly understated - probably a bigger problem than the former. 85% is probably more like 65-70% when one stops pretending that fundamental principles of finance somehow don't apply to the valuation of pension liabilities.
Other than the assets and liabilities, though, everything's fine.
When one sees "think tanks" that probably know better continue to ignore this issue in what appear to be comprehensive (or at least very long) analyses, one must wonder what source of funding or influence or other benefit to them would be at risk if they addressed the issue honestly instead of utterly ignoring it. They're not doing enough thinking in their pension finance tank.
Kudos to Oliver Giesecke and Hoover for being virtually alone in honestly examining this aspect of public pension funded status: https://publicpension.stanford.edu/.
I had a whole section about calculating the funded ratio, but I deleted it, because this post was getting to be too long in the first place (and it really needs to be its own standalone piece). I’ve been writing on these things online in semi-permanent places since 2014, so I don’t feel like cramming all the relevant points in a single post.
I do want to do something with Stanford's new dashboard/tools, because they're really nice -- and yes, as you know, I'd rather use a different valuation rate altogether.
But let me get back to the original point: if you use their standard approaches, how useful is it if the funded ratios can (and do) drop 30 percentage points in 2 years? What does that tell you about the pension systems? How good does that make one feel about 85% funded ratios?
Well, actually, without using a meaningful liability, funded status and year-to-year changes are misleading, close to meaningless. One should feel bad about 85% not because it can go down, but because it's wrong and the real number is much lower (and, secondarily, might go down further).
In calendar year 2022, when market interest rates increased substantially, causing market liabilities and bonds to decrease, and risky assets like stocks also decreased, reported funded status for public plans got much worse, but private sector funded status based on quasi-market liability measurements actually improved slightly because they reflected the impact of market interest rates on liabilities. (Check out the Milliman indices for each during calendar year 2022).
If you’re looking at measures of funded status and comparing reported non-market liabilities to market assets (though a good chunk are mismarked, to your point), why would you think the results have meaning? The reported assets and reported liabilities are real(ish) apples and wax oranges.
Yes, the fact that maybe 30% of assets are overmarked by probably 10-20% is a big problem, but isn't the fact that 100% of liabilities (a bigger number) are understated by 20-30% a much bigger problem? Focusing on only the assets is like a news outlet spending all its efforts reporting on a single murder and ignoring a mass shooting. Yet that's what everyone does when it comes to public pensions.
Continuing to dutifully report and analyze actuarial numbers undermines the credibility of CRR's Public Plans Database, Equable, and others, no matter how pretty the graphs, how extensive the (bad) data, how much work goes into compilation, or how lengthy and sophisticated-seeming the reports. It's astrology claiming to be astronomy.
This issue needs to be called out, not reinforced.
Sorry for ranting. It's just very disappointing to see these analyses reported on like there's nothing wrong with them. I hope you do get around to a separate post on that topic. You're an important voice. Stanford/Hoover would be a good starting point - not because their graphics are conceived and executed well, but because the numbers are realistic.
I don't mind you going on the rant -- after all, I agree with you re: valuation...
HOWEVER
This is a "hobby" for me (in my copious free time), so I'm not going to be spending a lot of time coming up with new numbers. If other databases are developed with the equivalent of LDROM for all the years... yay! I'll use that info. I'm not recalculating all this stuff myself.
If the Stanford numbers are downloadable, I can do stuff with that. [And yes, the -- "what's wrong with the funded ratio?" would be a good followup]
ALL THAT SAID, there is a reason I came up with my extremely simplistic cash flow model (which I need to update): all these balance sheet shenanigans, whether asset- or liability-side, are very abstract. The ultimate goal is cash flows to pension benefits.
So let's just do that.... (and see the money run out... or not) -- I know the actuarial models that project using various assumption sets could be used to do far more sophisticated stress tests/sensitivity analyses.... and perhaps I'll get Gemini to mock something up for me.... But that's for another time ;)
Funded ratios are just a crude point-in-time metric, whereas looking at cash flow simulations and failure (money runs out), sustainability (under certain conditions), etc. is a more general risk management perspective. I like looking at failure modes -- what would it take to break the system? Some of the plans are obviously fragile right now, but what about all those 85%ers? I can think of all sorts of "fun" scenarios....
Thanks! I understand and would never expect you to recalculate numbers. And as I said, you're an important voice on this topic - your hobby is productive and a public service.
The big problem I see is that "pension debt" as reported understates the real debt, under any meaningful and understood definition of that term, by $2-3 trillion. All the analyses regarding debt, funded status, and the sources of change thereof, are worth very little if they don't reflect market liabilities.
Stanford/Hoover being the exception, the reported data, while copious and accessible, are highly misleading because they are based on what actuaries and accountants misleadingly label "liability." Why do these groups fail to acknowledge this problem? CRR, for one example, is likely conflicted by its affiliations with NASRA and GFOA, both of which endorsed and may have contributed to (along with their actuarial collaborators) the ASOP 4 toolkit (https://www.ncpers.org/asop-4-toolkit-pensions-ldrom), which, to be frank, is unconscionably misleading. I worry that those types of collaborations mean that many widely cited groups will never fully report what they should if they want to further a meaningful discussion.
Writing about and analyzing and publicizing all the bad data because that's what's available -- without noting how misleading they are in the context of a discussion on pension funding -- is, in my opinion, an example of the "Streetlight Effect" (https://en.wikipedia.org/wiki/Streetlight_effect), where the drunk who lost his keys in the park across the street looks for them under the streetlamp because the light's better. It contributes to an illusory truth effect around public pension funding by "flooding the zone." I think it's very important to call this out in reporting on these numbers.
We need to write more stuff using the LDROM numbers…. Hmmm, I could pick some juicy targets
Amen. Plans like Chicago’s are basket cases under any measure. Looking at “well-funded” plans might be more interesting?
LDROM isn’t perfect but it’s 1000x more reasonable than the widely reported numbers.
Seem like good choices to me