Long-Run Global Implications of the UK's LDI Crisis

17 October 2022
14 min read

The recent crisis in the UK pension industry stemming from the “mini” budget and its impact on UK bonds and policymaking has implications far beyond the UK. It also raises questions about asset allocation, macro policy and the broader investment environment.

UK defined benefit (DB) pension schemes have adopted liability-driven investment (LDI) overlays to better match their liabilities. The UK “mini” budget announced on September 23 pushed UK bond yields up rapidly, with the 30-year gilt surging from 3.7% to 5.1% at the peak. This spike triggered margin calls on LDI instruments, forcing pension plans to rapidly sell liquid assets, including gilts—putting more upward pressure on yields. The Bank of England (BoE) was forced to postpone the start of quantitative tightening and resume buying long-term gilts.

This episode might be billed as a failure of LDI; we think that take slightly misses the point.

In the UK market, many DB schemes are closed to new members, with well-known liabilities and an explicit time limit on investments. With the encouragement of regulators, DB funds de-risked by taking derivative positions to match liabilities—and those LDI instruments did what they were meant to do. Any definition of risk has nuances: while this de-risking tool removed one kind of risk, it created another in the form of margin calls and the resulting liquidity demand when yields move in extreme ways.

This incident has both UK-specific and global implications. A key UK-specific aspect is the large relative size of DB pension liabilities and the gilt market, which can translate into a substantial impact from pension funds liquidating these securities to satisfy margin calls.

From one perspective, the relative sizes seem similar between the UK and US: total UK DB liabilities are GBP 1.6 trillion and the government bond market is GDP 2.2 trillion, so DB liabilities are 71% of the size of the equivalent asset that would best hedge them (Display). In the US, total DB liabilities are just over US$17 trillion and the government bond market US$22.5 trillion, implying a proportion of 76%.

However, if we leave aside US state plans and focus on private plans, liabilities are US$3.8 trillion, making liabilities only 13.7% the size of the government bond market (Display). What’s more, the US has a much larger investment-grade corporate credit market. Including that market increases the size of the relevant US bond market to US$27 trillion.

Sizes of DB Pension Liabilities and Government Bond Markets
A relative comparison of UK and US aggregate pension liabilities and government bond markets

Current analysis does not guarantee future results.
Left display: Gilts and T-bill data as of June 30, 2022; pension liabilities estimate as of September 6, 2022. Right display: DB pension liabilities data as of June 30, 2022, US Treasury data as of October 12, 2022 and IG (investment-grade) corporate data as of April 30, 2022.
Source: Federal Reserve Economic Data, Fitch Ratings, Securities Industry and Financial Markets Association, United Kingdom Debt Management Office, XPS Group and AllianceBernstein (AB)

In the UK, many pension schemes are closed to new members, implying a default approach in long-term government bonds where there’s enough funding. Beyond this, we see strategic global implications that extend beyond the narrow confines of UK pensions and the allocations of DB plans.

Systemic Risk Has Shifted

At the time of the global financial crisis (GFC), banks were the point of systemic risk in the economy. Since then, declining bank leverage has reduced that risk. By contrast, investors—including the pension system, have taken on a larger part of the mantle of systemic risk in both the US and UK. We can see evidence of this in fixed-income allocations that have descended the quality spectrum, and more so in the structural increase in allocations to illiquid assets, such as private markets and loan provision. The increase in exposure to illiquid assets, coming as the trend path in rates has shifted upward and the business cycle has made a comeback, constitutes a risk buildup that could be systemically important.

Retirement income remains a need, but the expected real return on financial assets and the quality of (nonfinancial) credit are lower, so risk levels must rise—even if it doesn’t show up in metrics like the trailing reported volatility of fund assets. Yet.

To be clear, we’re not claiming that this allocation is wrong. Far from it. We think that the combined pressures of lower nominal returns, higher inflation and less diversification highlight a need to increase risk in pension portfolios in order to deliver the required outcomes to beneficiaries. We’ve long argued that the true risk measure for a funded pension system, in aggregate, is the probability of a hardship outcome for beneficiaries—not the expected measured volatility of the portfolio.

The macro implication: a gradual transition in the 14 years since the Lehman Brothers bankruptcy has the pension system now carrying more of the systemic risk. The recent panic among UK pension schemes is a canary in the coal mine in this respect.

Pension Health and Economic Policy Are Inextricably Linked

This shift in risk points to a broader conclusion: economies in which there’s an attempt to fund pensions face an intimate, inescapable link between pension funds and economic policy. We’ve seen this painfully in the UK, which has been forced to effectively restart quantitative easing just when the BoE was about to head the opposite way. More broadly, democratic governments can’t let pension plans fail but are also desperate to keep those liabilities off their own balance sheets. The issue is amplified by the observation that debt/GDP ratios for developed economies have risen back to levels last seen at the end of WWII.

We face years of central bank hawkishness to clamp down on very high near-term inflation, but policy ultimately can’t be divorced from pension health. Separately, we’ve argued that governments may eventually be tempted to monetize debt as a way to reduce burdens. Both points suggest the possibility of real interest rates being lower in the long run than they otherwise would be (not that it feels that way at the moment). They also suggest that pension regulators may be encouraged to find ways to sanction higher risk tolerance, if it can be done with a focus on appropriate long time scales, a point we’ll touch on below.

Switching from DB to defined contribution (DC) has been one path. The initial migration from government to corporate balance sheets was proven unsustainable. That corporations could offer such schemes was really just an artifact of the macro backdrop, as well as favorable demographics and taxation. In the latter part of the 20th century, the system shifted to DC, with individuals bearing the risk. That said, the de-risking of the long tail of DB liabilities has been a long-winded affair.

Despite Recent Headlines, LDI Has Served UK Pension Schemes Well

We can assess how the funding status for UK DB plans has evolved over time (Display) with the assumption that a fund allocates 60% to gilts and 40% to equities. The equities are a proxy for what would, in reality, be a mix of risk assets, such as credit. We illustrate this under three different LDI hedging assumptions. The 100% LDI fund would have outperformed through the Brexit period, and the reverse is true for 2022.

The analysis also illustrates something counter to the sensationalist headlines: while the current “LDI crisis” has created cash flow problems for some UK pension schemes, it has far from created a solvency crisis. In fact, in many cases the discounted value of liabilities shrank at a much higher rate than assets. Indeed, LDI has generally helped reduce the volatility of funding ratios—a success marred only by poor execution issues in recent times.

Typical UK DB Plan Funding Ratio Development
Versus Buyout: 70% Funded as of March 31, 2015
A comparison of plan funded ratios under different percentages of LDI usage.

For illustrative purposes only.
As of September 30, 2022. Analysis assumes that a UK DB plan is 70% funded on a gilt-matching basis as of March 31, 2015. Liabilities are assumed to move 30% in line with the FTSE All Stocks Gilt Index and 70% in line with the FTSE All Stocks Index Linked Gilt Index. The investment strategy is assumed to be 40% invested in global equities (represented by the MSCI World Index 50% hedged to GBP) and 60% in gilts that match the plan’s liabilities. The LDI hedge is based on buying a swap to match the remaining unhedged liabilities at the ratios shown. The strategy assumes monthly rebalancing of both the physical asset allocation and the LDI hedge ratio—with the swaps settled and restruck at the end of each month. No allowance is made for fees.
Source: FTSE, MSCI and AB

Reduced Liability Values Make Annuitization More Attractive

The recent crisis suggests a potential allocation shift for UK DB pension plans. LDI using swaps may be viewed as less attractive, suggesting that a reallocation would likely favor more physical gilts, credit and securitized assets. Inevitably, leverage in the UK DB system will come under scrutiny, along with the liquidity implications of the swaps and the impact of using pooled vehicles that can lead to a rapid unwinding of demand. This issue does, however, go to the heart of the appropriateness of risk metrics. US DB funds have taken on a much smaller amount of leverage through this mechanism. As a result, the nominal asset risk of UK funds is higher. However, through the lens of funding-ratio risk, it could be argued that UK plans haven’t taken a high-risk position.

After their recent moves, UK DB plans have been left with less liquid investment assets and a relatively higher share of private assets and cash posted as margin. This is probably not an ideal long-run asset allocation; how it’s redeployed may depend on whether the BoE extends its temporary support for gilts.

If we look beyond the drama of recent margin calls, the net result has been a surge in UK bond yields, pushing down liability values. Lower liabilities will make it more attractive for many DB funds to annuitize via insurance companies. Yield increases earlier in the year have already made annuitization a prominent conversation point with clients, and we expect this to intensify. As a result, the crisis could ultimately hasten the end of DB as a significant force in UK pensions. This was always going to happen, but it now seems destined to happen much faster than anyone thought.

The DB to DC Shift Could Accelerate…with a Pivot Toward Real Assets

The investment industry tends to focus somewhat narrowly on current market structure, but there’s a bigger picture to lay out. Even as DB comes to an end as a significant force in the UK, individuals will still need to save for retirement. The shift of retirement-savings risk to individuals via DC pension arrangements is accelerating, with individuals buying target-date or similar solutions. Viewed this way, recent events imply an even larger asset-allocation issue.

It makes sense for a DB scheme closed to new members to invest the most of its assets in low-risk nominal and inflation-linked bonds that match the shape of known liabilities. When individuals save for their own retirement and bear that risk, the “liability” tracks inflation more closely, rather than a discounted cash-flow proxy, because retirement cost is an attribute of the real economy. That’s why we think the long-run implication is a changing mix of UK pension assets: away from long-dated bonds and toward shorter-dated instruments and real assets, such as equities and real estate, as well as away from UK-focused assets to more global allocations.

Questions About Future Demand for Long-Term Gilts

An asset-allocation change of that nature, on its own, has implications for the UK. The positive angle is that a long-horizon DC system that needs real assets should be willing to provide risk capital that can be used for productive investment. As we noted earlier, this may require regulators to allow a higher risk tolerance (illiquid asset exposures, for example), but that shift can also be viewed in the context of genuinely long-run investment time scales. This is a sharp distinction from DB investments in nominal high-grade bonds, which are less useful economically.

The flip side of this reallocation is the questions it raises about the future demand for long-maturity UK government debt. Consistent demand at the long end of the government yield curve has been a funding boon to the UK government for many decades, but we suggest that the shift in the pension industry that we’ve outlined here could call time on that benign state of affairs.

While heavily regulated insurers and DB schemes may tolerate low long-term bond yields, there’s little evidence to suggest that individual investors without such constraints would tolerate them. Indeed, this approach has enabled new DC retirees to be the greatest potential beneficiaries of recent market moves. As with many investors, they’ve seen capital values decline, (Display) but the income purchasing power of their diminished assets has increased substantially. That’s a marked contrast to DB plans locked into LDI and unable to benefit from this move.

Experience of 65-Year-Old UK DC Saver
Income vs. Capital Growth Since March 31, 2015
The value of income and capital of a DC saver from 2015 through September 2022

Past performance does not guarantee future results.
This chart illustrates the capital growth of £100 invested on March 31, 2015 in the AllianceBernstein 2020–2022 Retirement Strategy Target Date Fund and is net of fees. The conversion to an annuity is by reference to an annuity purchased with the capital available for such an investor, scaled to £100 per month as of December 31, 2015. The reference annuity is for a healthy, non-smoking 65-year-old living in Chelmsford, UK, is level in payment and has an attaching 50% reversionary annuity on the death of the policyholder. Annuity rates are the observed best-in-market rates by AB at the end of each month for a purchase amount of £20,000.
As of September 30, 2022
Source: AB

There’s (Still) No Such Thing as a Risk-Free Asset

The “doom loop” between LDI margin calls and the need to sell more gilts is a useful illustration of a larger point—that there’s no such thing as a risk-free asset. We’ve made the point in previous research that the existence of risk-free assets isn’t a given; it’s a contingent state of affairs, as can be seen over the longer arc of history (Display).

Is There a Risk-Free Asset?
Indexed Real Returns of Select Assets (1851 = 100)
Indexed cumulative returns of select real assets from 1851 through June 2022

Past performance does not guarantee future results.
December 31, 1851, through June 30. 2022
Source: Global Financial Data, Thomson Reuters Datastream and AB

The practical result of this point is that sovereign risk may need to be priced. The structural issue of high public debt/GDP, combined with the prospect that rate volatility remains high, implies that this sovereign risk may need to be reflected in a larger margin haircut for Organisation for Economic Co-operation and Development government bonds as well as in higher borrowing costs for all asset classes.

The Private-Asset/Liquidity Trade-off Has Become More Challenging

The UK LDI incident raises questions about the appropriate liquidity for pension funds—and therefore the appropriate level of private asset allocation. Ultimately, this is more a question for DC funds than DB funds closed to new members, but it was surfacing before the UK mini budget and recent events will amplify it.

An investment environment with less-established trends in interest rates, inflation and market returns implies, all else equal, a greater need for liquidity. However, things aren’t equal, which starkly clashes with the need for return and diversification in a higher-inflation world. This trade-off might be one of the toughest asset-allocation issues today.

We’ve argued that the force to increase private asset exposure stems from:

  • Prospects for lower nominal returns in public markets
  • The hunt for diversification, if bonds can no longer provide as much of it
  • Strategic inflation protection (public markets play a role here, too, but under the aegis of private assets there are important real-return providers)
  • The dearth of young, high-growth companies issuing listed stock

All those factors remain today, and are arguably more important now than pre-pandemic, resulting in a record allocation to alternative assets by institutional investors, most of which are highly illiquid allocations to private equity, private real estate and infrastructure (Display).

Record Alternatives Exposure Raises Liquidity Questions
Alternatives Allocation (Percent)
Annual percentage allocation to alternative investments since 2008.

Historical analysis and current estimates do not guarantee future results.
*2022 estimate prorates the 2019 stock and bond allocation with year-to-date returns of S&P 500 and US 10-year government bonds, respectively, and assumes no change in the alternatives allocation.
As of October 12, 2022
Source: CME, Datastream, McKinsey and AB

What’s the way out of this?

Recent questions from clients (leaving aside those from UK DB plans in recent weeks) have been from those seeking more liquidity and those who have more capital to deploy into private assets. We think the questions can be answered only in the context of time horizon and appropriate benchmarks. A liquidity threshold will still be needed to cover beneficiary payments and worst-case margin requirements. Beyond that, however, a fund will still need a high level of private-asset exposure if inflation is its true benchmark. The question is one of governance and setting measurement goals that reflect a long time horizon.

Conclusions: Implications of the LDI Crisis

Just as Victorians are often said to have invented childhood, the 20th century saw the invention of retirement. Will it be a permanent feature of society or merely a contingent state made possible by demographics, healthy real returns on financial assets and the belief in risk-free assets?

Retirement will clearly persist, but how its achieved—and funded—will evolve. The recent UK LDI crisis may be only a ripple compared with the sum of all global retirement provision. However, it could be a useful reminder that a much larger pool of global retirement savings has relied for quite a while on a particular set of investment conditions. The implications are far-reaching:

What It Means for UK DB Pensions

  • If pensions aren’t fully funded, then expect less LDI in the future and higher allocations to credit and securitized assets.
  • If the upward shift in yields has significantly reduced liabilities, pensions are more likely to enter into insurance arrangements to annuitize plans.
  • An alternative would be to buy and hold bonds directly—either way, there’s a significant reduction in the need to hold risk assets.

What It Means for the Investment Industry

  • The crisis accelerates the shift from DB to DC plans and expands the provision of vehicles for individuals to save for their own retirements.
  • Insurance solutions, such as annuities, that enable individuals to hedge longevity risk may make a re-emergence—after having been deeply unattractive to individual savers in the past decade. And the scope for insurance innovation is vastly enhanced in a higher interest-rate environment, as the cost of guarantees diminishes.
  • The DB/DC shift will also shift more retirement assets, implicitly or explicitly, from nominal to real return benchmarks, with a commensurate need for more real assets. That need is magnified by the macro shift in the likely path of inflation.
  • Expect a big debate about the allocation to private assets and the need for liquidity in pension plans. There are reasons for significant private allocations, but this incident puts liquidity front and center in allocation debates—and the tension goes to the heart of how funds’ governance structure is set.

What It Means for Global Policymakers and Regulators

  • Systemic risks have shifted over the past decade, with a larger proportion now lying in pension plans. In this light, pensions inherit some of banks’ unenviable position in the GFC era. Pension plans are arguably better equipped to take on risks such as holding illiquid assets, given their long time horizons. But regulatory and governance structures must reflect that pensions require this long horizon, while also reflecting that robust solvency requirements applied to these investors can remove it again.
  • On a related note, the pension system and policy setting are inevitably linked, a relationship that has long-run implications for real yields.
  • A positive angle from the turmoil could be the ability of long-horizon individual DC pension funds with inflation benchmarks being able to provide risk capital to fund useful activity in the real economy (something that closed DB funds weren’t incentivized to do).
  • As DB plans fade as a force in coming years, the UK government admittedly may need to find alternative ready buyers for its long-maturity debt. New, less highly regulated, buyers will be less supportive of negative or near-zero real rates—boosting the cost of government debt and shortening its duration. This will require governments to refinance debt more often, reducing policy flexibility.
  • There’s no such thing as a risk-free asset, an observation that could show up in the pricing of government bonds.

The views expressed herein do not constitute research, investment advice or trade recommendations and do not necessarily represent the views of all AB portfolio-management teams. Views are subject to change over time.

About the Authors