This is part two of the Jacksonville Police and Fire Plan saga.
Before getting into the math, let’s see another news item on the Jacksonville brou-ha-ha.
THAT’S SOME HICCUP
In the Jacksonville City Council Finance Committee Wednesday, Mike Weinstein, the city’s CFO, and Tim Johnson, the director of the Police and Fire Pension Fund, discussed a “waiver” in pension cost allocations that had been controversial in recent news cycles.
Their joint appearance was notable for a unified message, in which Weinstein said the city and the PFPF were on a “solution path” and that the money involved in the waiver was a “hiccup” compared to the larger unfunded liability.
However, countermessaging soon followed.
Shortly after this was originally posted on Wednesday afternoon, the city’s Chief Administrative Officer, Sam Mousa, took issue with our description of the unified messaging of Weinstein and Johnson as a “kumbaya moment” between the PFPF and the city.
“There is no kumbaya on this pension issue. We are still livid over the issue and we still are of the position we did not know and the fund did not follow state law. They misrepresented the fund value for years and we are not happy,” Mousa wrote in a text and said on the phone.
Weinstein and Johnson agreed on a more calming message: that so-called waiver, once expected to near an extra $45 million budgetary hit in the FY 17-18 budget, likely won’t be as dramatic or sudden a hit as expected … even as the city and the PFPF diverge on the legality of the waiver.
The two outlined potential workarounds for what was feared to be a budget-busting lump sum hit to the city’s general fund, including spreading out the cost over years and absorbing the impact into amortization after collective bargaining.
No matter how the waiver is mitigated, the city will be on the hook for at least $200 million in pension costs in the next fiscal year.
Oh lord, they plan to amortize the amortized cost.
Maybe I’m reading it incorrectly, but if they’re not recognizing the need for much higher payments just to reach the amortization….. and are going to amortize that amortization… ugh.
THE SIMPLE EXAMPLE: 30-YEAR PAYOFF OF LIABILITY
I’m setting up an extremely simple situation to try out the ideas of the impact of the payroll assumption. I have set my spreadsheet up so that there can be additional changes for further investigation, but for right now, this is what I’m modeling:
- Initial liability balance: $1 million
- Interest rate on liability: 7% compounded annually
- Payments made once a year, at the end of the year
- Amortization period: 30 years
Think of this as a 30-year loan with annual payments.
The twist is how we’re paying off this loan.
Let’s do it the “normal” way first: level payments for the 30 years.
The payments are the bars, at about $81K/year. The line shows the balance being paid off. With each payment, the balance decreases. This is just finance 101.
But let’s be clever and pay off the liability in a way that’s easier for us to stomach. After all, don’t all of us have steadily increasing wages in nominal terms, and it will be much easier in the future to meet a payment than now.
So instead of considering the payments level through the 30-year period, let’s assume our income increases at 3.25% per year, and we’ll instead pay a level percentage of that wage. This is similar to paying a level percent of payroll, and assuming payroll increases by 3.25% per year.
Let’s see what that looks like:
Okay, we see a steadily increasing payment amount over time, but that balance… well, that’s a bit odd. It’s increasing over the first few several years. Indeed, the balance increases from $1 million at the beginning to a peak of $1.1 million 9 years later, until it starts coming down again. Indeed, the balance doesn’t fall below $1 million (the original balance) until year 15.
But let’s compare the payment patterns of the two directly:
Oooh, a nice reduction in payments in the early years – we start out paying $57K with the growth assumption – of course, it’s quite a bit more by the end….
If I add the dollar amounts for each of these payment patterns, then you find that in the level payments situation, we paid a total of $2.4 million over the 30 years, and $2.8 million for the increasing payments.
To be sure, $400,000 paid much later may be easier to swing, so it’s not necessarily a bad idea.
WHAT IF YOUR ASSUMPTIONS ARE WRONG?
So let us think of something that will be analogous to the Jacksonville Police and Fire plan.
Let us suppose you figured out that you would pay off the liability as a level percent of payroll, and you’re assuming that the payroll is increasing 3.25% per year.
You’ve got your level percent set up, and you’ll just multiply that by your actual payroll each year to get the payment.
To put in the Jacksonville twist: you assumed the payroll would grow 3.25% per year… but it didn’t grow at all.
If that were the case, then you’d keep paying in that level percent… which would be inadequate.
So let’s say that you “discover” the discrepancy 10 years into the loan. You realize payroll isn’t growing at all, so you amortize the remaining 20 years with a level payroll assumption.
This is the result:
Notice the increasing balance through those ten years. This is important. The peak balance, at year 10, is $1.18 million — your principal increased by almost 20%.
Let’s do the direct comparison of payments – I changed the “assumption change” payment stream to a line to make it easier to compare:
So you start with a payment that is about 30% less than if you had been paying it in a level amount….and then end up making payments that are 40% more.
The increase you see from the first 10 years to the remaining 20 is almost a doubling.
YOU CAN’T WAIVE MATH
Now, these are very simplified examples — I’m not putting in complications such as additions to the liability year-to-year and how the liability value in general is affected by payroll growth, that the amortized amount is adjusted every year (though one still may have negative amortization), yadda yadda.
But here is the point: there is a danger of negative amortization when one amortizes as a percentage of payroll, when you’re assuming that the payroll increases at a steady rate. And then it doesn’t.
Negative amortization is what we see when the payments you’re making aren’t enough even to cover the interest charged. So that interest is added to the balance… which keeps getting higher as your payments continue to be inadequate.
You may remember some of the worst defaults on mortgages in the credit crisis were for negatively amortizing loans. These are very risky, but they were sold on similar underlying assumptions: that the base salary (or home value) would steadily grow, so it would be easy to afford later what one can’t afford now.
But the pension (and mortgage) math had the inexorably, reliably, accruing debts at stated interest.
And the assumed payroll/salary/property value growth did not emerge.
So the borrower kept falling behind, the balance kept growing, and then the lenders, who thought they were safe, will suffer the blow as well when the payments aren’t supportable.
There’s a reason prime mortgages are structured to be level payments, as opposed to inexorably increasing ones.
The reason is that negative amortization leaves both lender and borrower open to more risk. People have been behaving like pension funds are special sorts of creditors, so of course they can handle ways of “paying off” the unfunded liability that actually was guaranteed to cause it to increase in the “short run”.
Until the “short run” became 10 years.
And one finds one’s self in a much bigger hole than when one began.
THE OFFENDING HICCUP
Weinstein offered a response to this piece Thursday evening, and it is below in full.
“The chair [of the Finance Committee] asked for the conversation to move us forward not backwards. I mentioned numerous times the PFPF didn’t use the state calculation formula since 08 and each year it gave the city an incorrect result so our contribution was wrong. I also said numerous times when PFPF was asked by the division of retirement council members and administration why aren’t they using the state required formula their response was they have a waiver. I stated no such waivers can be given we find out. I also restated this is why we have to get of the defined benefit business. My opening comment was we have had decades of bad decisions that have brought us to this problem.”
“My position continues to be aligned with the mayor. Not following the state statutes resulted in a misrepresentation of the city’s financial obligation to the PFPF.”
“The hiccup was in reference comparing the 25 million to the 2.5 billion problem. The hiccup was not in reference to the ongoing misrepresentation,”
The $25 million is the portion that they are attributing to recognizing that payroll isn’t really increasing at all (a total increase of less than 2% over 10 years is pretty much 0.)
By all means, point out that $25 million is but 1% of $2.5 billion. That’s going to make people feel so much better about that pension hole.
Those waivers were a bad idea. Rather than helping to make the pensions more affordable, it just made the problem worse.
I have no issue with the assumption change finally being recognized, because that’s not the real problem. It’s that reality was ignored to begin with.
Trying to go back to the fantasy, or, dear lord, “refinancing” the refinancing by still pretending it will eventually get more affordable in the future… oh, some decades from now… will only make the problem worse again.
Kentucky Pension Liabilities: Trends in ERS, County, and Teachers Plans
Nevada Pensions: Liability Trends
Stupid Public Pension Trends: Divestment Expands