• Home
• About Us
• Services
• Bio
• Ramble
• Contact

California Pension Systems: Ranking their Funding Assumptions

Roeder Financial       June 28, 2009

This survey (click here to see the survey in a new window) ranks the funding assumptions used by California’s public pension systems from “most conservative” to “most optimistic.” There is no “right” or “wrong” in setting assumptions. There can be a number of valid reasons that an assumption package for Entity A differs from Entity B. Entity A might have a larger equity allocation than Entity B. Entity C might wish to have more conservative assumptions to be able to fund an ad hoc COLA in most years. The nation’s largest state plan, CalPERS, may be able to have certain size-related investment efficiencies unavailable to smaller sponsors.

Due to the current financial crisis facing the state and most local entities, there will be a great temptation or need for plan sponsors to do everything in their power to minimize pension contributions. If there is a request by the plan sponsor to change certain actuarial assumptions, this survey may have value in terms of clarifying what “the herd” is doing. This temptation to reduce near-term contributions is not limited to plan sponsors.

Even though active employees benefit in their retirement years by having well funded systems, making assumptions more conservative also has a “cost” for actives. Excepting Alameda-Contra Costa Transit District, all plans in the survey are contributory. Lower assumed investment assumptions often directly or indirectly translate into higher employee contributions. Further, certain plan sponsors have said in past recessions that if they do not get contribution relief, it will result in the layoff of x employees.

Defining “most conservative” versus “most optimistic” is useful.

Most Conservative	Most Optimistic
Lower Assumed Investment Return	Higher Assumed Investment Return
Higher Assumed Pay Increases	Lower Assumed Pay Increases
Shorter Amortization periods	Longer Amortization Periods
Explicit Expense Load	No Explicit Expense Load
Entry Age Normal Funding	Projected Unit Credit Funding
Level Dollar Amortization	Level Percent of Pay Amortization

Using comparative funded ratios, to determine how well funded a plan is, can be misleading for several reasons:

• Actuarial assumptions will often not be comparable.

• The steep market decline will make comparability harder for entities with different valuation dates -- pre and post 2008’s 4^th calendar quarter.
• A high funded ratio could be largely attributable to Pension Obligation Bonds (POB). In looking at the financial viability of a plan, it is essential to look at more than just than the computed actuarial rates if there is also POB debt service.

However, to say that funded ratios do not matter is both false and foolish. Consider the City of Fresno systems. Both Fresno Systems will have assumptions ranked toward the bottom, reflecting “most optimistic” assumptions. However, their funded ratios have easily been among the highest in the state during the current decade. Both Systems have an extremely intelligent approach to handling a 100+% funded ratio in that the vast majority of “excess” assets in any one year was retained -- with a relatively small residual being paid to retirees as a contingent benefit. If there is one healthy byproduct of the recent stock market decline, it would be to discredit the use of the concept and term “excess” asset. Traditionally, all assets that were in excess of liabilities were characterized as “excess.” However, yesterday’s excess assets would have been a helpful buffer as today’s “rainy day money” instead of being used to pay out new and contingent benefits during better times.

There is strong preponderance of phasing-in the differences between actual investment returns and expected returns over a 5-year period. However, CalPERS instituted 15-year averaging earlier this decade. Tulare County instituted 10-year smoothing earlier this month. Will other entities be tempted to go to a longer period to temper the steep market declines? Is this appropriate? More Retirement Boards and their actuaries will be addressing this question in the next year.

CalPERS combines its 15-year asset smoothing with a rolling 30-year amortization policy on actuarial gains and losses. This approach does is give very little weight to the actual actuarial experience in any one year. Even though this does not result in a pegged contribution rate, it produces results in the proximity of a pegged approach. There are critics of corridor funding. The CalPERS approach has many similarities to corridor funding. Both the CalPERS approach and corridor funding have the significant advantage of a higher degree of budgetary certainty from year-to-year in most years.

Are there any limits to the extent of smoothing that can be used in the actuarial process?

Yes, but those limits are fuzzy. Actuarial Standard of Practice #44 was issued in 2007 and opines in Section 3.3, “The actuary should select an actuarial value method designed to produce actuarial values that bear a reasonable relationship to market value.” This survey addresses whether entities have a maximum permitted divergence from the market value in asset value determination. The most common limitation, when it exists, is that the actuarial value is not permitted to diverge more than 20% from market value. Stakeholders in certain systems (ie, CalPERS, Alameda County, Tulare County, City of San Diego, Alameda-Contra Costa Transit) have been exploring whether it is sensible to expand a 20% divergence. Kern County recently adopted a 50% divergence from market. ASOP 44 stipulates there be no bias in the smoothing method. If the recent market decline is treated as a “special circumstance”, ASOP 44 may be violated.

Last week, the CalPERS Board approved for its contract agencies a recommendation from its in-house actuaries to temporarily expand its 20% permitted divergence (“80/120 smoothing”). For 2009 valuations, to the extent that the excess of smoothed actuarial value over market value exceeds 20% (but not more than 40%), a closed 30-year amortization base is established for actuarial gains/losses. For 2010 valuations, a similar approach is taken for any divergences within a 30% corridor. For 2011 valuations, any divergences over 20% are amortized on a closed, 30-year basis. CalPERS reverts to existing policy commencing with 2012 valuations. Obviously, the purpose of this special policy is to offer some temporary funding relief in trying times – especially to those entities who adopted increased benefit levels pursuant to SB 400 earlier this decade.

It is important to note that the assumed actuarial rate of return is almost always net of expenses incurred. The City & County of San Francisco System and the Alameda – Contra Costa Transit District are virtually the only Systems which have an explicit load for expenses as part of its computed contribution.

Several points should be noted on the amortization of unfunded liabilities. “Open” or “rolling” methods will use the same number of years in a future valuation as is been used in the current valuation. “Layered” means that there is a new amortization base established each year which is funded on a “closed” or “declining” basis. If one believes that it is best practice for an individual’s benefit to be fully funded at their anticipated retirement date, it is a sound practice to have the amortization period be closely correlated with the average future working lifetime of the active member group (typically between 12 and 15 years). 30-year amortization passes a significant part of the cost, attributed to current participants, to a future generation of taxpayers. All amortization approaches noted in this survey should be assumed to be level percent of payroll unless otherwise indicated. Level-percent-of-payroll amortization will produce a lower contribution than level dollar amortization over the same period.

In the spreadsheet, the “most conservative” system is ranked as #1 and “most optimistic” system is ranked as #40

Rick Roeder, FSA

The source for survey data has largely been from system web sites. Plan administrators were sent a draft report to give them the opportunity to make any corrections and updates, as needed. The final version will be on roederfinancial.com. Thanks to the many who helped make the survey more up to date.

Guest authors are welcome to submit articles and their thoughts to the ROEDER RAMBLE for potential viewing on this site.