Pension Risk Remains A Key Credit Consideration

The U.S. not-for-profit health care sector saw a slight dip to the median funded status of its pension plans in fiscal 2019 but levels remain very high overall. We view the dip in fiscal 2019 as primarily due to a decrease in the discount rate used to measure pension liabilities, which increased those liabilities. The discount rate is based on a conservative municipal bond rate, and we view the 2019 decrease as being within reasonable volatility expectations, so we don't consider it to be a fundamental change in the funded status of the plans. While funded ratios can be and often are affected by volatile investment markets, a consecutive dip in fiscal 2020 could more likely be driven by a number of factors beyond a decrease in the discount rate. These include more volatile investment markets compared to 2019 stemming from continued uncertainty related to the coronavirus; recessionary pressures with ongoing headwinds as the economy gradually reopens; and the upcoming presidential election. We therefore believe variable investment returns will likely characterize much of calendar year 2020 potentially curbing measurable increases to funded status levels by year end. Further, while a return to an increasing trend in the median funding status is unlikely over the next year, we believe the overall funded status of pension plans in fiscal 2020 in the health care sector will remain stable in the near term.

In the near term, S&P Global Ratings believes a stable funded status will likely result in a stabilizing of statutory minimum contributions to defined benefit (DB) pension plans, which tempers the benefit to overall financial profiles because operating performance in the health care sector remains under stress. (For more information, see "Not-For-Profit Acute Care Sector Outlook Revised To Negative Reflecting Possible Prolonged COVID-19 Impact," published March 25, 2020 on RatingsDirect, and "A Bumpy Recovery Is Ahead For Hospitals And Other Health Providers As Non-Emergent Procedures Restart," published May 26, 2020.) Further, the bond rate has proven to be volatile from year to year, the projected benefit obligation for many plans remains large in our opinion, and many plans have updated assumptions such as mortality to more accurately recognize longer lifespans over which benefits are to be paid out. Therefore, we believe pension risk remains a key credit consideration in the health care sector.

Current stressed markets notwithstanding, many not-for-profit issuers continue to focus on de-risking strategies that lower pension funding risks such as increasing annual contributions to improve the funded status, minimize their reliance on volatile markets through more conservative investments, closing current plans to new participants as a means of long-term liability reduction, freezing benefit accruals at current levels, and in some cases, terminating plans altogether.

The majority of not-for-profit hospitals report liabilities under the Financial Accounting Standards Board (FASB), which prescribes liability measurements using different sets of assumptions and methods than those under the Governmental Accounting Standards Board (GASB). Comparison between plans under different reporting entities is not straightforward and this is represented in our opinion of individual creditworthiness.

Future Developments – Lower Funded Ratios In Fiscal 2020 Driven by Heightened Volatility

We expect continued volatility in the funding levels of pension plans, given annual fluctuation in the bond rate and continued volatility in the investment markets as there is much uncertainty driven by the pandemic and the recession. The Chicago Board Options Exchange VIX index (a benchmark used to measure the stock market's expectation of future volatility) has increased significantly in 2020, demonstrating the continued volatility in the investment markets. Most providers have made adequate contributions to their pension plans, and we expect overall pension costs to be somewhat lower and manageable for most issuers. However, for some providers, the combination of lower funded ratios and lower bond rates will result in increased costs, causing budgetary pressure as well as potential liquidity pressure. In some instances, providers may look for budgetary relief in pension contributions, through direct contribution deferrals or weakening of amortization methodologies. We will continue to assess these situations on a case-by-case basis and also make negative liquidity or performance adjustments for providers that we believe will have rising future costs that could weigh on the credit profile.

In order to provide health care providers near-term financial relief, the Coronavirus Aid, Relief, and Economic Security Act includes a provision allowing certain providers with single employer plans to have pension funding holidays during calendar year 2020 with a due date for any deferred contributions of Jan. 1, 2021 (including accrued interest). We note that many providers are taking advantage of this provision to achieve near operating and liquidity relief. However, we expect this to result in weaker future funded ratios and will assess this impact on a case-by-case basis.

Overall Funded Ratios Dip Slightly And Are Likely To Do So Again

In fiscal 2019, the not-for-profit median funded ratio fell by roughly one percentage point to 83% after reaching a 10-year high in fiscal 2018 but the 2019 funded ratio remains the third highest since 2009 (see chart 1). This modest decline was primarily due to an decrease in the bond rate, perhaps influenced by the weak investment market going into fiscal 2019 following the S&P 500's posted loss of 4.4% in 2018 from a particularly weak fourth quarter. As noted above, unstable bond rates will likely characterize fiscal 2020 when considering the recent market volatility due to economic concerns related to COVID-19, coupled with the upcoming presidential election in the fall.

S&P Global Ratings has tracked the funding levels of the DB plans of not-for-profit hospitals and health systems since 2007, when, on average, they were at their highest level (90%). Funded statuses declined sharply in 2008 and 2009--by 20 percentage points--following the significant downturn in global investment markets during the Great Recession. After the recession, funded ratios were essentially flat at about 70% through 2012, despite hospitals' healthy contributions to plans.

Chart 1

While the year-over-year median funded ratio has remained fairly constant, variance among plans within our sample is high but the range tightened considerably in fiscal 2019 compared to fiscal 2018 (see table 1). In fiscal 2018, the variance was between a low of 44.6% and a high of 144.5%. However, in fiscal 2019, the variance was between a low of 60.8% and a high of 128.6%. The top 10 highest funded pension plans reflect a funded status that is considerably lower compared to those listed in fiscal 2018, respectively. Conversely, the top 10 lowest funded pension plans reflect a generally stronger funded status compared to those listed in fiscal 2018. Specifically, none of the ten lowest plans listed last year had a funded status of 60% or higher, while this year, all of the plans listed have a funded status of over 60%. While we use the same methodology to derive data from year to year, the timing of this publication shifts thereby sometimes including or excluding different credits captured in the tables as some audits are not available due to a difference in the fiscal year end.

Table 1

In our view, most hospitals and health systems manage their pension burdens well, isolating pension risk from contributing to negative credit implications (see table 2). At the same time, a high funding burden can inhibit improvement in credit quality, even if there exists no direct negative credit impact due to pension risk. In many circumstances and to varying degrees, underfunded pension plans contribute to credit stress, particularly for providers already struggling with thin income statements and balance sheets. A low funded ratio might not drive a credit rating, but can weigh down credit quality if these obligations appear to be swelling. We present table 2 for fiscal 2018 to capture the entire universe of credits in our portfolio with DB plans.

Table 2

In general, we consider fully funded plans (plans funded at 100% or more) or the absence of a DB pension plan as positive credit factors in our assessment of an organization's financial profile. Conversely, we view DB plans that are considerably underfunded, or expected to be underfunded in the near- to mid-term, as risks to the financial profile. Of course, we also understand the existence of a DB plan is a positive long-term recruitment and retention incentive in tight labor markets. Again, while we use the same methodology to derive data from year to year, the tables that appear in last year's report show different credits due to the timing of this publication and the availability of fiscal year audits for those credits with a Dec. 31 fiscal year end.

Net Periodic Pension Costs Remain Low And Manageable For Most FASB Issuers

In fiscal 2019, the median net periodic benefit cost to total operating expenses leveled off at a low 0.2%, compared with the much higher 1.3% in fiscal 2011 (see chart 2). Similarly, employer contributions as a percentage of EBIDA extended their largely declining trend and stood at a 10-year low of 6.5% in fiscal 2019 versus a high of 15% in fiscal 2012 (see chart 3), remaining well below 10% for a second consecutive year. We view this trend positively as the ever lower employer contributions as a percentage of EBIDA translate into greater financial flexibility for hospitals and health systems, as management teams spend less of their cash flow on pension contributions, freeing up cash for other uses such as debt repayment, capital expenses, and building reserves. In our opinion, the continued decline in the contribution-to-EBIDA measure of this ratio reflects growing EBIDA levels, improved funded ratios, and pension costs that remain manageable. While annual changes in pension contributions and cash flow influence this measure from year to year, pension contributions on a nominal basis have generally increased in the past 10 years, due to weaker market returns at times and updated demographic assumptions such as mortality (specifically, people living longer). However, an increase in the contribution-to-EBIDA measure does not necessarily indicate higher pension contributions, primarily because the ratio can be skewed during times of depressed cash flow; for example, in fiscal 2009, when cash flow was soft for many hospitals.

Chart 2

Chart 3

For many organizations, rising pension costs and underfunded pensions are issues that management teams, in general, are successfully managing with no immediate credit implications. However, we consider DB plans and their funded status as a component of our assessment of an organization's financial profile. Employer pension contributions can directly affect unrestricted reserve ratios and crowd out other cash needs, although at times, management teams have meaningful latitude on the timing of contributions. Furthermore, the effect of swings in funded status is evident in several key ratios that we analyze, including debt to capitalization and pension-adjusted debt to capitalization.

Assumptions Drive Valuation

The discount rate is one of the most influential factors used in the measurement of pension liabilities. Higher discount rates result in lower liability measurements and correspondingly lower annual pension benefit costs, while lower discount rates will raise these amounts.

Over the past 10-plus years, inflation and real market return expectations have decreased and assumed asset returns have trended down slowly, as assumption changes lag experience. The median return on plan assets declined to its lowest level in 10 years in 2019, at 6.8% (see chart 4). Pension plan valuations can be somewhat volatile and depend on realization of actuarial assumptions over which sponsors have limited control. In addition to the assumed asset return and the discount rate, mortality assumptions can play a significant role in liability measurement. The most conservative mortality assumptions consider annual future longevity improvements, called "generational mortality" assumptions. Updating to a generational approach from a static table has led to one-time jumps in liabilities, but likely will mean fewer increases in the future because the tables include the expectation of longer lifespans.

A look at rating actions over the past few years indicates pension funding has sometimes contributed to credit strain. However, the issue has rarely, if ever, been the primary reason for a rating action in the not-for-profit health care sector, in part because pensions are long-term obligations whose values move with investment markets and actuarial changes, and because DB pension plans in the sector are generally well funded in our view. Because we expect changes over long periods, our ratings accommodate some volatility in the valuation of pension liabilities. In addition, most not-for-profit health care providers have strong unrestricted reserves to cushion year-to-year pension funding volatility.

Whether the average funded status will remain at the current level or improve depends on a number of factors, including contribution sufficiency, market returns, discount and bond rate trends, other actuarial assumptions, and benefit design changes. Many hospitals and health systems are moving to mitigate risks through more conservative asset allocation strategies, while others are focusing on reducing liabilities by making benefit design changes. Some are reluctant to change or curtail DB plans that have long been a part of their benefits packages and that they see as a powerful recruitment and retention tool.

Chart 4

Projected Benefit Obligations Remain Relatively Consistent

Since 2018, we've noticed a slight decrease to projected benefit obligations as a share of assets. Total overall assets were steadily increasing through 2018. However, in 2019, total assets decreased to near 2017 levels. The specific reason for the slight decrease in this ratio in unclear, as inconsistent asset levels limit the certainty with which key drivers can be identified. That said, we view it as plausible that the slight decrease in this ratio reflects both an increase in assets from accelerated funding as well as an increasing population of recent hires that are in cheaper new benefit tiers.

Chart 5

Defined Contribution Plans Are Still A Minority But Are Growing

We identified hospitals and health systems that have defined contribution (DC) plans only (see table 3 for the 25 largest). This represents over 34% of our rated universe, indicating that DC plans are a significant minority in the not-for-profit health care world (see chart 6). This has ticked up from the 33% we reported last year and we expect this shift to slowly continue as providers move away from traditional DB plans, although most hospitals that freeze their DB plans still have pension obligations and contribute to them for many years. For those organizations, we still include the DB plan in our medians, and they aren't part of the "defined contribution only" group. In our analysis, we recognize the benefits of reduced volatility in plan contributions, and eventually reduced contributions in absolute terms, as a frozen DB plan winds down.

Table 3

Chart 6

Pension De-Risking Continues To Be A Dominant Theme

Many issuers have continued to focus on de-risking pension plans by closing their plans to new participants, terminating or freezing plans, or offering a lump-sum payout to some employees, though the current recession may alter these strategies toward cost-cutting, as opposed to plan funding.

When a sponsor closes a DB plan to new participants, new employees aren't eligible for that plan. Furthermore, an employer might freeze a DB plan, where not only is the plan closed to new participants, but existing participants stop accruing benefits while still being entitled to all vested benefits at retirement. In most cases, after the plan is closed or frozen, the employees earn all future benefits in either a cheaper DB plan or a DC plan, which is a far more predictable expense for management because the investment risk is shifted from the employer to the employee. However, because employees hired before the plan's closure are still vested in the DB plan, the plan continues to pay benefits for many years, generally for the life of all vested beneficiaries. The "tail" can be decades long. The closed plan no longer accepts new employees, so some funding methodologies and assumptions need to be made more consistent to ensure funds are available as the population ages into retirement. This typically translates into higher near-term contributions. Our credit analysis recognizes the benefits of reduced volatility in plan contributions, and eventually reduced contributions in absolute terms. Freezes often result in an actuarial gain, which increases the plan's funded status at the time of the freeze. However, the existing obligations still can pose financial risks for sponsoring organizations.

Outright pension terminations have been rare in recent years because when a plan is underfunded, the plan sponsor must first contribute enough to fully fund it before distributing assets to beneficiaries. Furthermore, low interest rates make it very expensive to purchase annuities for beneficiaries. The recent uptick in funded status and slightly higher interest rates could bring plan terminations into a more realistic price range for some sponsors.

Finally, we view cases where an entity issues debt to fund its pensions generally as an increased acceptance of market risk. Debt payments are made on a fixed schedule (as opposed to a schedule where management has some flexibility) and we incorporate this into our analysis through debt metrics.

Accounting And Reporting Considerations

FASB vs. GASB reporting

While it may be reasonable to compare liabilities between different pension plans, it's important to note that there are major differences between reporting standards. Liabilities reported under FASB, as is the case for most health care plans, are discounted using a very conservative bond rate and point-in-time liability measurement. These liabilities tend to be quite conservative due to the low discount rate, but volatile because they change annually, as they tend to trend in line with broad interest rate movements. Liabilities reported under GASB are generally designed to have long-term costs spread out smoothly over time by incorporating the assumed asset return within the reported discount rate along with a more conservative liability measure.

Unlike GASB reporting, under FASB reporting the service cost component is reported in the same line as other compensation costs in operating expenses. Remaining components of pension costs, such as interest cost, experience gains and losses, or changes in assumptions, must be reported separately from service costs and outside of operating income. We do not adjust the reported pension costs in our analysis.

There are also differences for reporting retiree (other post-employment benefit, or OPEB) liabilities. These liabilities may be more easily reduced than pensions, and this has led to the elimination of DB plans or conversion to DC plans for many GASB and most FASB OPEBs.

Church plans don't need to adhere to ERISA

The Employee Retirement Income Security Act (ERISA), a federal law that sets minimum standards for employee benefit plans, governs most rated not-for-profit health care universe pension plans. However, many not-for-profit DB plans are classified as "church plans," so while reporting under FASB guidelines, they do not need to adhere to ERISA guidelines, and therefore have more flexibility in determining underlying assumptions for funding levels and contribution amounts. While added flexibility can also mean added risk of underfunding, we review obligors with both church plans and ERISA plans in accordance with our criteria, examining funding levels, overall costs and cost trajectory, and underlying assumptions.

The June 2017 U.S. Supreme Court Advocate Health Care Network v. Stapleton ruling regarding the exemption of church-affiliated pension plans from ERISA did not fully settled the matter, resulting in further litigation. Lower court decisions have been mixed and not entirely favorable for hospitals with church plans. As the status of church plans remains uncertain, we expect the matter will continue to be challenged in the courts. Although this adds risk to the financial health of issuers with church plans, we continue to evaluate the plan characteristics of each issuer individually in accordance with our criteria.

Methodology And Criteria Considerations

Ratios in our analysis

In our "U.S. And Canadian Not-for-Profit Acute Care Health Care Organizations" criteria (published March 19, 2018), we measure the extent to which current, proposed, contingent, and off-balance-sheet liabilities could affect an organization's debt servicing capability. In our criteria, one of the four measures used to evaluate debt is the funded status of a DB plan. Our assessment includes a forward-looking view of funding requirements and management's plans to address the risks. We believe a low pension funding ratio could signal elevated risks after incorporating the appropriateness of actuarial assumptions. Similarly, we consider increased credit risk for pension contributions that aren't actuarially determined, based on weak actuarial methods, or if required contributions aren't regularly funded. We characterize stand-alone hospitals with a funded ratio of 55%-65% as vulnerable, and those below 55% as highly vulnerable. For health care systems, we characterize a funded ratio of 50%-60% as vulnerable, and those below 50% as highly vulnerable. We also assess the actual magnitude of the liability because some frozen plans may have a low funding ratio, but the overall size of the liability can be quite modest since many of these plans are in run-off mode.

Ratios provide insights, but each plan is unique

This is our ninth study of the not-for-profit health care sector's pension plans. We use combined data from a broad sample of our rated universe, including 188 of 284 (or 66%) rated providers that maintain defined benefit plans with fiscal years 2009-2019. The sample size is smaller for the 2019 data because most of the audits for the fiscal year ended Dec. 31 aren't yet available. Our analysis is based on the current universe of our ratings. Therefore, sample sizes each year may vary slightly and we may have less than 10 years' of data for any given rated provider. We view these analytical limitations of the sample as negligible given that the portfolio is broad and remains sufficiently stable enough from year to year and the sample size is large enough to support conclusions drawn from the identified trends. Tracking the broad sample allows us to monitor pension funding trends using our considerable sample size to smooth any unusual effects from individual organizations. However, each plan has unique characteristics, including actuarial assumptions, workforce characteristics, payout amounts and timing, asset allocation, and changes such as a freeze in benefit accruals. In addition, we review issuers' asset allocation strategies for DB plans to ascertain if there might be excessive market and liquidity risk accepted within the investment portfolio. These unique characteristics can have a substantial impact on a plan's funded status, and in any given year, some plans' asset and liability valuations could change in a direction contrary to the broad sample. In our analysis, we recognize each plan's characteristics while applying our understanding of the broader trends gleaned from the median ratios.

Plan sizes vary significantly

When analyzing DB plans, we believe that in addition to the funding level it's important to look at the plan's relative size. Some plans that are underfunded as a percentage of the projected benefit obligation (PBO) or total pension liability (TPL) are relatively small. This can happen when plans have been closed or frozen for a long time, or if only some employee groups or hospitals within a system are covered.

As a proxy for relative size, we use balance sheet and revenue measures. The balance sheet measure is the PBO (or TPL) as a percentage of the health system's total assets. We use this measure to gauge a plan's size relative to an organization's total asset base. For this measure, a lower value indicates a lesser risk, all other things being equal.