Volatile economic conditions combined with larger funding gaps have prompted concern about the risk defined benefit (DB) plans can pose to their sponsors. Willis Towers Watson has been measuring dynamic pension risk for more than a decade,1 and this year’s analysis finds only minor changes from last year’s risk landscape. The analysis measures pension risk for companies in the Willis Towers Watson (WTW) Pension 100.2

The average funded status of DB plans sponsored by the WTW Pension 100 climbed to 90.4% by the end of 2013 — boosted by rising interest rates and a prosperous stock market — but fell back down to 82.4% by the end of 2014, mostly due to lower interest rates and the widespread adoption of new mortality tables.3 At year-end 2015, funding levels were roughly the same as they were in 2014, although fund performance was volatile over the year. Heavy losses in the third quarter of 2015 were relieved by higher investment returns during the last quarter, and rising interest rates helped keep funding levels constant as well.

Among companies in the WTW Pension 100, the median Pension Risk Index (PRI) score ticked up slightly compared with the previous year — from 1.7% in 2015 to 1.9% in 2016.

Willis Towers Watson Pension Risk Index

Fluctuations in retirement plan assets and liabilities can represent a significant risk to plan sponsors. Willis Towers Watson developed a measure of one-year value-at-risk (VaR) in a pension plan relative to the sponsor’s market value under a simulation of adverse financial conditions. The VaR is the increase in the funding deficit (as measured for financial accounting purposes) under adverse market conditions, given the plan’s asset allocation and funded status at the beginning of the year. The PRI quantifies the pension risk in additional pension underfunding to a company’s core finances over the following year and allows for a risk comparison among plan sponsors.

The adverse market condition is modeled as a 5% probability outcome. For example, a PRI score of 4% suggests that an increase in the pension plan’s funding deficit over the coming year will exceed 4% of the sponsor’s market value 5% of the time.

To perform the analysis, pension plan assets and accounting liabilities are stochastically simulated for 25,000 scenarios, with the net funding liability realized over the year then divided by the company’s market capitalization. Scenarios are generated using Willis Towers Watson’s capital market assumption model (the expected return estimator) and reflect a composite of asset returns and interest rate changes projected by the model.

The PRI analytic also reflects the linked/combined ups and downs of pension assets, the company’s stock price and changes in the pension plan’s asset allocation.4 The company’s market value is assumed to be affected by the stock return in each scenario, and the 5% probability outcome is used to decrease the market value accordingly. The analysis calculates standard corporate beta for each company because all companies do not respond to market fluctuations in the same way.

Beta measures the risk/opportunity potential of a stock or investment portfolio expressed as a ratio of its volatility to overall market volatility, with 1.0 representing the market. This equity market benchmark is often estimated using a representative index, such as the S&P 500. For example, a portfolio whose beta exceeds 1.0 is expected to experience greater volatility than the overall equity market, while a portfolio whose beta is less than 1.0 is expected to have less volatility than the market.

Estimates of the duration of plan liabilities are based on the ratio of benefit payments to the pension benefit obligation (PBO). The durations in our model range from 9.5 to 15 years and reflect the net impact of discount rates and any other linked assumption changes.

PRI values for the WTW Pension 100

This analysis focuses on companies in the current WTW Pension 100 and looks at the VaR for 2015 and 2016. Among this group, the median PRI score inched up from 1.7% over 2015 to 1.9% over 2016. Scores were less than 0.5% for 13% of these pension sponsors in 2016, down slightly from 16% of sponsors in 2015 (Figure 1). A score of less than 0.5% means that the adverse financial outcome realized in the model would impose little disruption to the company’s core finances. PRI scores were 0% for five of the 14 sponsors whose risk scores were less than 0.5%. These companies’ pension surpluses insulate them from the adverse 5% probability outcome.

Figure 1. Distribution of PRI scores for 2015 – 2016 in WTW Pension 100
Click image to enlarge

Source: Willis Towers Watson

PRI scores of 10% or higher (indicating that adverse market conditions could trigger a pension loss of more than 10% of market capitalization) pose the risk of a large disturbance to the sponsor’s core business. PRI values were 10% or more for 6% of companies in 2015 versus 8% in 2016. These sponsors typically maintained a relatively large pension plan (relative to firm value),5 with the asset allocation skewed toward equities. Poor equity returns in 2015 caused declines in these companies’ market value, which substantially increased the relative size of their pension plans. Although higher interest rates helped reduce pension liabilities, the reduction was not always enough to counteract declining market capitalization.

Among WTW Pension 100 companies, PRI values increased for 72% of plan sponsors, decreased for 22% and held steady for 6% between 2015 and 2016 (Figure 2).

Figure 2. Changes in PRI scores from 2015 to 2016

Changes in PRI scores from 2015 to 2016

Source: Willis Towers Watson

Where PRI scores declined, the improvement was often at least partly attributable to a shift to fixed-income assets, as well as a marked decline in relative pension size. Among these companies, the median pension size dropped from 19.5% at the beginning of 2015 to 16.3% at the beginning of 2016. Over the same period, median equity allocations fell from 38% to 35%. Lower equity allocations reduced the negative effect of adverse equity markets. Additionally, among companies whose PRI scores declined, current risk — as measured by the current pension deficit/surplus over market value — fell from 3.6% at the beginning of 2015 to 2.4% one year later. Median PRI scores for this group declined from 1.4% in 2015 to 1.1% in 2016.

For the 72% of these companies whose PRI scores increased, relative plan size remained at 26%, and median allocations to equities and fixed-income investments remained at 46% and 37%, respectively. At the beginning of 2015, current risk was 4.2% for this group, which ticked up to 5.2% over the next year. Median PRI scores for these companies climbed from 2.2% in 2015 to 2.7% in 2016.

What affects PRI scores?

Relative plan size is a primary driver of PRI scores (Figure 3). By 2016, the median pension size was less than 20% for almost half of the WTW Pension 100 (45%), and the median PRI score for this group was 0.8%.

Figure 3. Median PRI scores by beginning-of-the-year pension size (PBO/market capitalization), 2016

Median PRI scores by beginning-of-the-year pension size (PBO/market capitalization), 2016

Source: Willis Towers Watson

Even after controlling for different asset allocations, the association between pension size and PRI scores remained strong. For every percentage point increase in pension size, PRI scores increased by 0.19 percentage points. Where pension liabilities present a challenge, sponsors might want to find ways to reduce retiree or inactive participant obligations, such as purchasing annuities from a third-party insurer.

Asset allocation also plays a role in PRI scores, albeit to a lesser degree than plan size. To reduce investment risk, more plan sponsors continue to allocate a larger share of pension assets to fixed-income investments, often as part of a liability-driven investment (LDI) strategy.6

Figure 4 compares median PRI scores with the percentage of assets held in fixed-income instruments. The lower PRI scores among companies whose allocations to fixed income were 50% or higher support the presumption that these sponsors are successfully employing an LDI strategy (as explained below). Plan sponsors that held the majority of assets in bonds generally had lower PRI scores than other sponsors. For 2016, the median PRI score was 0.9% for DB plans with fixed-income allocations of 50% or more, versus 2.8% for plans whose allocations to debt were less than 50%.

Figure 4. Median PRI scores by percentage allocated to fixed income, 2016

Median PRI scores by percentage allocated to fixed income, 2016

Source: Willis Towers Watson

LDI strategies typically use fixed-income assets as a hedge against the downside risk of lower interest rate scenarios, which drive up plan obligations. In years when long-term, high-quality corporate bond interest rates decline and plan obligations rise, corporate bonds should realize positive returns and vice versa.

Our PRI model assumes that sponsors that allocate the majority of their assets to fixed income have adopted an LDI strategy. To get a better sense of how higher debt holdings can reduce risk, Figure 5 shows median PRI scores under three asset allocation scenarios and bond durations, holding plan size, funded status, market capitalization and other parameters equal.

Figure 5. PRI outcomes under different asset allocations and bond durations, 2016

PRI outcomes under different asset allocations and bond durations, 2016

Source: Willis Towers Watson

As shown, overall pension risk declines as the asset portfolio becomes more risk avoidant. With their lower volatility, bond returns can mitigate losses in adverse financial markets. Moreover, pension risk scores drop even further among companies that allocate assets to long-duration bonds, which further hedge against lower interest rates and their impact on liabilities. While large equity positions can — if all goes well — reduce long-term pension cost faster, the trade-off is higher market risk and funding volatility.

Is the company’s risk profile reflected in its PRI score?

To evaluate whether pension risk is factored into a company’s overall risk profile, we divided the sample into two categories by bond rating: investment grade or junk bond.7 Within the investment-grade group (88 companies), the median PRI score was 1.7%. For the junk-bond group (12 companies), the median PRI score was 15.7% — roughly nine times higher.

Figure 6 depicts the PRI distribution among investment-grade and non-investment-grade companies. PRI scores were less than 5% in 82% of investment-grade companies versus only 17% of those with a junk-bond rating. Further, 58% of those with a junk-bond rating had PRI scores of 10% or more, versus only 1% of the investment-grade group.

Figure 6. Distribution of PRI scores by credit rating, 2016

Distribution of PRI scores by credit rating, 2016

Source: Willis Towers Watson

Finally, only one of the eight sponsors whose PRI scores were 10% or more was rated as investment grade (BBB). Therefore, the pension risk profile seems to correlate to the sponsor’s overall risk profile as measured by the S&P credit rating.

The total VaR measured in dollars8 of non-investment-grade sponsors represents only 12% of the sample’s additional risk (Figure 7), thus limiting the potential impact on the entire market.

Figure 7. Total dollars at risk by credit rating ($ thousands)

Total dollars at risk by credit rating ($ thousands)

Source: Willis Towers Watson


According to our analysis, the risk from pension plans to companies’ core finances increased slightly from 2015 to 2016. The mild uptick in PRI scores over the last year reflects a combination of factors. While higher interest rates reduced liabilities, declines in equity returns resulted in funding shortfalls in some cases and lower company values by year-end 2015.

So far in 2016, market conditions have not been favorable for pension sponsors. Over the past few months, equity returns have been poor, while interest rates for high-quality corporate bonds fell by almost 60 basis points from year-end 2015 to early May, pushing liabilities higher. If these conditions persist, sponsors’ financial positions will deteriorate further by year-end 2016. Plans using an LDI strategy will likely fare much better, as long bond returns have been strong so far in 2016, which should mitigate the negative impact of falling interest rates.

Companies whose pensions pose significant risks to their core business might want to consider de-risking their plans. Risk-reduction strategies — such as asset-liability duration matching and shifting from equity to debt (and other less market-sensitive investments) — can help sponsors lower their risk profile. Moreover, measures oriented to reduce obligations through lump sum buyout offers or annuity-based strategies can also lessen the pension’s exposure.


1. See “Pension Risk Increases Slightly for Fortune 1000 Plan Sponsors in 2015,” Willis Towers Watson Insider, November 2015.

2. The 2015 WTW Pension 100 consists of sponsors of the 100 largest U.S. pension programs among U.S. publicly traded organizations, ranked by pension benefit obligation at year-end 2014.

4. PRI results are based on three alternative allocations of fixed-income assets. If less than 40% of plan assets are invested in fixed income, the model assumes the fixed-income assets are invested in Barclays Capital Aggregate Bond Index (duration generally five to five and one-half years). If 40% to 50% of assets are invested in fixed income, the model assumes the fixed-income portfolio is invested in a blend of 50% Barclays Aggregate Bond Index and 50% Barclays Capital Long Government/Credit Index (duration roughly 10 years). For plans with more than 50% invested in fixed income, all fixed-income assets are assumed to be invested in Barclays Capital Long Government/Credit Index (duration generally 14 to 15 years).

5. Pension size is the ratio of PBO to market capitalization.

6. See “2014 Asset Allocations in Fortune 1000 Pension Plans,” Willis Towers Watson Insider, October 2015.

7. The categories reflect the S&P rating, which classifies companies rated BBB– or higher as investment grade, and those rated BB+ or lower as junk bonds.

8. Dollars at risk equal PRI score times market value.