Investment Losses and Public Pensions

Earlier this week, San Diego City Employees Retirement System CEO Mark Hovey penned an editorial in the Union-Tribune providing some very useful background on the city’s defined-benefit retirement system for public employees. There is much to like in Hovey’s commentary — except his analysis of the role of investment losses in creating the city’s current unfunded pension liability. The main thrust of Hovey’s argument is that investment losses are not a major cause of the liability. His analysis is flawed and misleading, and his conclusion is wrong on three key points.

First, Hovey points out that SDCERS’ average rate of investment return during the past three decades (9 percent) has exceeded the investment gains assumed by the actuarial model. This fact is not particularly helpful in identifying the implications of investment losses, however, because it is the recent, short-term rate of return — rather than the long-term average — that plays the pivotal role in determining taxpayer pension contributions.

To understand this, consider two states of the world, both of which would produce a long-term average annual rate of return of 9 percent: In the first, the pension system would make a consistent 9 percent each year. In the second, the pension system would make an 18 percent return on investment in each of the first 15 years, and 0 percent in each of the second 15 years, also producing a 9 percent long-term average over the three decades.

In the first state of the world, there would be no problems, and the retirement system would remain well funded. This would not be true in the second, however.

During the first 15 years, high investment gains would exceed those assumed in the actuarial model, making the pension system look “overfunded.” As a result, the city’s pension contribution would fall. In an extreme case — like the University of California retirement system during the past two decades — taxpayers would make no pension contribution, taking a “pension holiday” and living off these high earnings. The problems would come during the second 15 years, when investment gains would systematically underperform relative to the actuarial assumptions, opening up a huge unfunded liability.

This temporal component — the fact that the taxpayer pension contribution is based on the current snapshot of the pension system — explains why long-term averages are not particularly informative. The variance around that average is what explains the volatility in the city’s pension payment, and causes the payment to rise when the city has the least amount of revenue to pay for it. Because the retirement system has increased its investment in riskier asset classes over the past three decades, the variance and volatility have both grown.

Second, the long-term average ignores the reality that the city has treated good investment years differently than bad ones. During much of the past three decades, the city has defined investment returns that exceeded the assumed actuarial model as “surplus earnings,” and spent it on benefit enhancements and other budgetary purposes. This left the pension even more underfunded during bad years. Put simply, the long-term average cited by Hovey systematically overstates the actual returns for the pension system, because many of the returns during good years were withdrawn and spent on other stuff, exacerbating the effect of investment losses.

Finally, in arguing that investment losses have not caused the current unfunded liability, Hovey notes: “The $1 billion of investment losses sustained by SDCERS during the 2008 financial crisis has since been fully recovered.”

This may be factually correct, but beside the point. The actuarial model used by the pension system assumes that the value of investments will grow 7.75 percent each and every year. Although the value of SDCERS’ investment portfolio today may be the same as three years ago, prior to the start of the financial crisis, our actuarial model has assumed that they would be worth 25 percent more this year. The gap between what the value of our assets should be under the actuarial model and what it is today translates into an unfunded liability in the hundreds of millions of dollars.

Vlad Kogan lives in University City.