Commentary: Stress testing a critical need for public pension funds

Just days before he left office in January, then-New Jersey Gov. Chris Christie signed a law requiring the state pension funds to regularly stress-test their systems and disclose the results to the public.

Several other states, including Connecticut, Hawaii and Virginia, already require regular stress testing by public pension plans. Minnesota and Pennsylvania might soon join the ranks as their legislatures are considering a similar requirement.

​I believe all public pension plans should regularly conduct stress tests. Those that don't risk having stress testing thrust upon them by lawmakers. More importantly, stress testing is critical to the long-term success of public pension plans.

Many public pension plans are experiencing a funding crisis as contributions become a larger share of government budgets. They are investing in riskier assets as pressure mounts to maintain their assumed rates of return. Some already have reduced benefits or closed their plans to new employees. However, in some cases this predicament could have been avoided if the plans had conducted regular stress tests to determine the likelihood of achieving their goals and the financial implications if they didn't.

As an actuary for the past 35 years and licensed pilot for more than a decade, I get the adage that what isn't measured can't be managed. To safely navigate a pension plan or an airplane it's essential to not only carefully measure and monitor the current course, but more importantly to anticipate what might lie ahead. Just as pilots navigate around thunderstorms instead of staying on autopilot, pension plan executives would be wise to routinely stress-test their funding models.

In February 2016, the Actuarial Standards Board's pension task force suggested several changes to actuarial standards to improve disclosures to the public, but stopped short of recommending that all pension plans be required to provide projections and conduct stress tests, even if the pension plans are in danger of becoming insolvent.

The case for pension plans to perform stress tests is straightforward. Pension funds are akin to banks and insurance companies. All three institutions collect money — whether deposits, premiums or contributions — and pay out money to the customers, the insured and participants. In the case of pension plans, the benefit payments they make can stretch out for decades. Unless pension plans test the viability of their assumptions, they cannot determine if they need to change course and develop contingency plans to manage future risks. Public plans that conduct stress tests can identify unsustainable financial practices and risks before they become critical.

In response to the financial crisis a decade ago, Congress passed the Dodd-Frank Wall Street Reform and Consumer Protection Act requiring large financial institutions to conduct annual stress tests to ensure they have enough of a cushion of capital to withstand even the biggest shocks to the financial markets. The law also requires banks to report the results of these tests to regulators.

State insurance regulators began requiring regular stress tests by insurance companies only after the near collapse in 2008 of American International Group Inc. Stunned insurance regulators and actuaries concluded that the existing solvency testing requirements of the state insurance departments were woefully inadequate because they relied solely on the actuaries' best estimates and did not test the sensitivity of the assumptions to determine if they were faulty. Today, the solvency testing requirements for insurance companies are far more rigorous.

Unfortunately, it often takes a big jolt such as these events to shift from autopilot. Too many pension fund advisers and trustees operated on autopilot until the plans were buffeted by the Great Recession. After the financial crisis, it was difficult for many plans to get back on course without taking such painful steps as cutting benefits or requiring plan sponsors to make unsustainably large contributions.

Stress testing, sometimes also known as scenario or sensitivity testing, is an analysis or simulation aimed at determining the ability of a financial institution to withstand volatile economic conditions. For pension plans, that means scrutinizing key assumptions about the future, especially investment returns under different economic scenarios.

Instead of performing financial projections using only "best estimates," insurance companies, banks and pension funds — or their regulators — stress-test underlying assumptions to gauge how vulnerable they are to market crashes. In other words, stress testing analyzes the risk of not meeting goals. For pension plans, the biggest risk is running out of money to pay promised benefits.

Here's a simplified example of how a stress test might work. The first chart shows a best estimate of a 15-year projection of a state's pension plan contributions measured as a percentage of member payroll assuming the plan will earn exactly the assumed 7% on its assets every year. Under this scenario, the state's contribution rate remains very stable, ranging from 11.4% to 12.2% of payroll over this period.

Although the pension plan assumes it will earn 7% annually, in reality, the actual returns will vary from year to year.

The second chart is an example of a simple stress test where the pension fund's returns duplicate the investment returns of the 15-year period from 1930 to 1944 with high market volatility. This test shows that instead of contributions remaining relatively stable, the state's contribution rate more than doubles the rate under the best estimate and exceeds 26% of payroll.

The trustees of this plan must answer tough questions: What would happen if the plan's contribution rate exceeds 26%? Would the plan sponsor be able to afford the contributions necessary to restore the plan's financial health? Would the plan need to cut benefits? What can they do now to avoid this scenario? And, what is the likelihood of this happening?

A more dynamic stress test would answer some of these questions through a Monte Carlo stochastic simulation.

Using thousands of computer simulations and the latest information, meteorologists create weather charts such as this to enable people to decide what to do when facing dangerous hurricanes. This particular chart plots not only the path of Hurricane Maria but also the areas most likely to face the strongest winds.​

Pension trustees have similar charts available.

In this pension stochastic chart, we show the plan sponsor's expected contribution as a percentage of payroll, as well as the probabilities of the contribution rate being higher or lower.

​Like the weather chart, these stochastic charts enable pension trustees and plan sponsors to better understand the risks their plans face, including the probability that the funding ratio might drop below a certain level in a given year or that contributions may become unsustainably high. This insight allows trustees and plan sponsors to make calculated decisions today to lessen these risks.

For public pension plans to become financially secure, it is essential that they regularly conduct stress testing and sensitivity analyses as part of annual valuations and disclose the results of these tests to the public. I believe trustees cannot fulfill their fiduciary duties without routine projections and stress testing.

Some of my colleagues in the actuarial profession have resisted stress testing, arguing it would be an unnecessary additional expense for pension plans. Advances in technology have invalidated that argument. The costs to society of the collapse of a large public pension plan would be far greater.

Gene Kalwarski is founder and CEO of actuarial consulting firm Cheiron Inc., in McLean, Va. This content represents the views of the author. It was submitted and edited under P&I guidelines, but is not a product of P&I's editorial team.