On August 4, 2016 trustees of the Bricklayers and Allied Craftsmen Local No. 5 Pension Plan out of Newburgh, NY became the seventh multiemployer (union) plan to file for benefit cuts under MPRA in an attempt to avoid insolvency.
From their latest 5500 form here is the plan’s relevant data:
Minneapolis was ranked by US News as the tenth most dangerous city in America. Minneapolis? Really? More dangerous than Newark, NJ?
Then earlier this month I was reading a book* that made the point that a major reason Minneapolis appeared so dangerous in the rankings was because they kept better crime statistics while other large cities were so out of control that many of their crimes went unreported.
So it seems with valuing liabilities for public pensions where a plan can have a low funded ratio in part because they are valuing their liabilities honestly while other plans look better because their liability numbers are understated. Has anyone ever questioned whether benefits are being costed properly for public plans? In the private sector it seems like PBGC checks every benefit calculation for every plan they cover but for public plans whatever the computer spits out is accepted as gospel.
But, since we went to the trouble of pulling off valuation data for 154 state plans, let’s take a stab at it.
Funded Ratio (FR):
Assets / Accrued Liabilities
is a faulty indicator of which state has the biggest pension problem since the liability numbers are surely fudged* so as to develop contributions as low as possible. But, since we did the grunt work of pulling off the data from state actuarial reports, it is easy enough to sort and subtotal all the data by state which show that, with an FR of 45.78% and total unfunded liabilities of over $135 billion (also the most), the worst funded state pension system is….
Depletion Ratio (DR):
(Deposits – Payouts) / Assets
is a good indicator of which state has the biggest pension problem since those numbers are difficult to fudge with.
Paring down our DR spreadsheet to only Illinois plans shows $11.6 billion being paid out of a fund that is supposed to have $114.5 billion with $9.8 billion coming in. That translates into a DR of -1.61% which would mean all the money is gone in 62 years.
Not exactly a crisis when compared to New Jersey which has a DR of -6.64% and 15 years left. What this tells me is that Illinois is making an effort to fund their pensions (even if they have to tax to do it) while many other states careen toward plan bankruptcies, benefit defaults, or some combination.
But, after sorting and subtotaling all the data, it turns out there is one state that has a lower combined DR for all their plans than New Jersey.
An aging pundit with Parkinson’s disease reflects on life, politics, journalism, and public pensions.
In 2012 Eileen Norcross wrote two pieces on how public plan actuaries manipulate assets higher and liabilities lower. The first on Asset Smoothing featured New Jersey prominently:
What are the effects of smoothing? It depends on the formula. Roman Hardgrave and I find in our 2011 paper that New Jersey’s smoothing formula allowed the AVA to remain far above the MVA for a decade. During the time, the assets looked larger than they were. Smoothing allowed an unpaid liability to accrue, pushing costs forward.
The second on Normal Cost Methods is an even more pernicious stratagem and, based on our recent review of actuarial valuations, universal.
In the private sector, since PPA, all funding in based on the Present Value of Benefits Accrued which seems logical: this is what the participants have accrued to some valuation date and what they would get (if the money were there) were they to leave. But, for public plans, why do what is logical when a Generally Accepted Actuarial Principle for public plans (not the one about getting paid up front if you do work for Illinois) allows you to low-ball your numbers through reporting the ‘Actuarial Liability’?
None of the valuation reports we reviewed noted the camouflage but the Arkansas Teachers Retirement System conveniently provided an exhibit detailing the method: