When the Wall Street Journal praised Illinois for its pension reform (“In Praise of Illinois”, April 8), it was like praising a doctor who has been watching a patient bleed to death and then orders a band-aid. As shown in my previous blog post, the major Illinois pension funds run out of money in 2018, even under some pretty generous assumptions about investment returns and contributions. The bill signed yesterday by Governor Quinn will slow the growth of liabilities, but only for newly hired workers. Newly hired workers are generally decades away from retiring with a large pension. By the time this bill is likely to make any real difference in actual benefit cash flows, I predict an Illinois pension crisis will have forced more drastic measures.
On April 13, Bloomberg called Illinois a “poster child of debt crisis draining state services.” I agree with the assessment in that article, but even the figures stated therein are far too optimistic. In particular, the figures cited in the piece follow a simple procedure that has been employed in many studies and press reports, including the much-cited March 2010 study by the Pew Center: they simply subtract reported state pension assets from reported state pension liabilities and portray the difference as the unfunded liability.
In this post, I take an updated look at the three largest Illinois state pension funds: the Illinois Teachers Retirement System (IL TRS), the Illinois State University Retirement System (IL SURS) and the Illinois State Employees Retirement System (IL SERS). I show how discounting at economically more sensible rates changes the most recent figures.
Under what I would argue is the most sensible approach, the difference between assets and liabilities in these funds alone is $145 billion. That is around $11,000 per Illinois person, or $30,000 per Illinois household. It amounts to about 6 years of Illinois tax revenue, and is about 2.5 times the figures typically cited in the press.
I also estimate that when one also considers several other major Illinois plans that taxpayers are on the hook for at the municipal level, the difference between assets and liabilities rises to over $200 billion, which is over $15,000 per Illinois person, and around $42,000 per Illinois household. And the latest Illinois pension legislation does not change the calculations at all.
As Robert Novy-Marx and I have been arguing for a while, the liabilities published by state and local governments are calculated using inappropriate discount rates. From our 2009 paper* in the Journal of Economic Perspectives (JEP):
[G]overnment accounting standards require [states] to discount their liabilities at the expected return on their assets. This approach is analytically misguided: the magnitude of pension liabilities and how a pension’s funds are invested are two separate issues that should be considered independently. In practice, the accounting standard being used sets up a false equivalence between pension payments, which are extremely likely to be made, and the much less certain outcome of a risky investment portfolio.
The government accounting standards literally require states and local governments to violate the principles of financial economics. Discounting at expected rates of return on assets leads state governments to completely ignore risk when calculating the value of liabilities.
In deciding how to appropriately account for pension liabilities from a taxpayer perspective, one must first choose what liabilities should be recognized. A reasonable starting point is the Accumulated Benefit Obligation (ABO), which only counts benefits earned up until now under state benefit formulas. This is conservative because it ignores all benefits that employees will earn in the future unless the plan is completely frozen. In fact, the ABO recognizes fewer liabilities than state methods do. Illinois, for example, uses methods that incorporate projections of some future wage increases. Not only does the recent legislation not even touch the ABO, it doesn’t even affect future service accruals for existing workers.
The next choice is the discount rate. The JEP paper argues that the Treasury yield curve is appropriate given the relative safety of ABO benefits:
Consider payments that the states have promised employees for years of work already done […]. From the state’s point of view, these cash flows are extremely likely to be incurred. First, state constitutions in many cases provide explicit guarantees that public pension liabilities will be met in full (Brown and Wilcox, 2009). Second, state employees are a powerful constituency, making it hard to imagine that their already-promised benefits would be impaired. Third, the federal government might well bail out any state that threatened not to pay already promised pensions to state workers. In practice, Accumulated Benefit Obligation pension liabilities are probably the most senior of all unsecured state debt.
Treasury yield curves of course have some flaws for this purpose. The paper “Public Pension Promises: How Big are They and What are They Worth?” analyzes these issues. It also shows how these calculations change if one uses other discounting methods such as municipal bond rates, and other accrual methods that recognize future earnings or years of service. Treasury yields are not unambiguously too high or too low for discounting public pension promises. On the one hand, the fact that Treasury prices reflect a liquidity premium make yields lower than would be appropriate for discounting an illiquid stream of cash flows. On the other hand, the cost of living adjustments (COLAs) in public pensions offer some inflation protection, and non-indexed Treasury yields are higher than they would be if they also had inflation protection.
Of course, if Illinois ultimately were to get a federal bailout, the tab would be picked up by all US taxpayers, not just by taxpayers in the state of Illinois. So if you are thinking about a bailout, the valuations I present are really from the perspective of the US taxpayer in general.
This Table shows updated valuations of the three major Illinois state pension plans. The total liabilities stated in the latest reports reflecting June 2009 valuations amount to $124.6 billion. The state arrived at these valuations using discount rates of 8.5%. Re-discounting cash flows at the June 2009 Treasury curve yields total liabilities of $198.6 billion, or 1.59 times the stated liabilities.
This $198.6 billion in liabilities compares to $48.6 billion in the market value of net assets from the reports. As of September 2009, however, the funds reported a combined $53.8 billion in assets to Pensions & Investments(P&I) magazine. Using the more recent (and higher) asset values, the difference between assets and liabilities under the Treasury discounting is $144.8 billion. Spread over 12.9 million Illinois residents, that amounts to $11,221 per man, woman and child living in Illinois. Spread over 4.9 million households, it is around $30,000 per household.
There are number of Illinois plans that this analysis excludes.
- The Illinois Municipal Retirement Fund (IMRF), which covers employees of local Illinois governments and school districts with the exception of the City of Chicago and Cook County. Benefits are determined by the state, but the IMRF is funded at the municipal level (i.e. largely through property taxes). The latest IMRF report is from December 2008. Liabilities in that report were $27.1 billion. The IMRF itself discounts at 7.5% instead of 8.5% and also uses a broader actuarial accrual method than the other Illinois funds (Entry Age Normal instead of Projected Unit Credit). So moving to the ABO and redicsounting using the June 2009 Treasury curve only increases liabilities by 52%, less than the 76% increase at the major state funds. Compared to September 2009 Pensions & Investments assets, the gap is around $20 billion.
- The Chicago Teachers Pension Fund covers Chicago teachers, as they are not covered by the IL TRS. We have not done a complete valuation, but applying the very conservative 52% liability increase found at the IMRF to this fund (which uses a narrower accrual method than IMRF and an 8% discount rate instead of 7.5%) suggests an unfunded liability of $14.0 billion at the Chicago Teachers fund. This is based on the latest disclosures, which are June 2008 for liabilities and September 2009 for assets.
- The State of Illinois sponsors 2 other smaller plans: the Judges Retirement and the General Assembly Retirement System. Again based on the 52% liability increase assumption, total liabilities are around $2.4 billion for the Judges and $0.4 billion for the General Assembly. While the funding status of these plans is quite weak, together they only add about $2 billion to the unfunded liability.
Using these conservative procedures, I estimate that these additional plans have an unfunded liability of approximately $36 billion. Adding this to the $166 billion from the Table, the picture is not pretty. The difference between assets and liabilities is over $200 billion, which is over $15,000 per Illinois person, and around $42,000 per Illinois household.
I should note that Illinois municipalities run their own police and fire systems, which in many cases suffer from these same problems. An analysis would require collecting individual data from each city and town.
How can Illinois escape this mess? If the Illinois defined benefit systems are to be maintained, then benefit factors need to be reduced for all future service performed, including future service performed by existing workers. It is also time for Illinois and other states to consider freezing plan benefits entirely at their current levels. To compensate workers for these measures, the state could establish defined contribution plans for future service and bring more Illinois state workers into the Social Security system. Only when the growth of the massive off-balance-sheet debt has been stopped will the state be in a position to find ways of meeting its existing obligations.
* The JEP paper was summarized in Kellogg Insight.