Monday, February 24, 2014

Detroit Bankruptcy Plan: A Lose-Lose for Retirees and a Math Problem

Detroit Bankruptcy Plan:  

A Lose-Lose for Retirees and a Math Problem

On February 21st, the city of Detroit emergency manager, Kevin Orr, presented a plan to U.S. District Court Judge Steven Rhodes to bring Detroit out of bankruptcy. One salient point of the plan is a reduction to the pensions of Detroit retirees. Police and fire retirees are poised for a 10% cut (4% if they accept right away), and general retirees are set to take a 34% reduction. This is on top of a giant increase in retiree health care costs, where the retirees will get a fraction of the amount necessary to keep health care.

It's unfair, and aggravating that retirees take a hit for a situation they didn't cause and a decision they didn't take part in. Bondholders and creditors made conscious decisions to lend money to Detroit, but a pensioner had the presumption that they were to receive a pension that was protected by the state constitution. But that's not all. I don't like the precedent; if you think about it, now the most distressed cities are having their risk increased, and that affects hiring to bonds. I also think that the private sector has at least a safety net of insurance, which apparently a state constitution cannot provide.

But beyond my moral and social objections, there's a potential problem with the math.

You see, pension math and bond math are different. A bond has a specific legal obligation that is written and enforceable. So if I lend the city of Detroit $1,000, they agree to pay me back my money at a maturity date and pay interest in the meantime. The liability is known, and measurable: the city owes me $1,000. In a pension, the city owes an annual obligation to fund the liabilities of the plan. The liabilities of the plan are calculated using a complicated formula. Basically, the liability of a pension plan is calculated based on:
a.       The amount promised the retirees (their monthly pension);
b.       An estimate of how long the plan has to pay it (the mortality 
          assumption), and
c.       The rate of return the plan will make to fund it (the interest rate 
Sound like a lot of moving parts? It is. Now, here's more of the problem. According to the EM, the unfunded liabilities of the pension system total $3.5 billion. In February, when the state of Michigan's team came in, the unfunded liabilities were $650 million. The city's actuary, which represents literally hundreds of municipal pension systems and applies a rigorous set of established actuarial standards, reported Detroit's general fund as 82.8% funded, and the police and fire pension as 99.9% funded.

The EM hired a different actuary, and that actuary recalculated the pension obligation using a different set of assumptions, which they called a, "very rough preliminary guestimate," and then recharacterized the pension obligations bonds (secured bonds to fund the pension) as unsecured bonds outside of the pension (not the normal practice). That new calculation adjusted the assets way down to 32% for the general fund and 50% for the police and fire fund. This put the pensions below the 80% threshold to allow them to be under EM control.

In other words, the liability might be $3.5 billion, or might be $650 million, or might be some other number, or nothing. In addition, the plans pay this liability over the lifetimes of the pensioners, not all at once, or even on a schedule like bonds. During that time frame, the plan's assets make money and retirees pass away. Depending on those factors, we can determine where the liability goes. If the plan makes more on its investments than the assumptions, the liability reduces, or even vanishes.

It reminds me of the quote attributed to Mark Twain: "Figures don't lie, but liars do figure."

Pension math is tricky, and I don't think people's financial lives should be at the stake of a single calculation. The average Detroit general retiree gets a pension of $19,000, which by my standard is very modest, and to cut that by $6,460, plus make them pay all but $125 of their retiree health care is devastating. I guess I can liken this to an appraisal on a house. We get one appraisal, then another one comes in way lower. Do you automatically accept the lower number (in the case of liabilities, the higher number?) To me, I'd like to see a real version of what the unfunded liabilities really are before the people who worked hard for the city get a haircut. Not only that, if the plan's assets are invested right, there may end up being no liability.

Let's prove the figures. The retirees deserve it.