Britain’s LDI crisis: When things nearly fell apart

• The Truss government’s mini-budget of September 2022 sparked a sell off in Gilts and a crisis centred on LDI strategies
• A year on, and with litigation looking likely, what lessons have been learned?

Not only had the UK’s new political leaders ignored that policy, they hadn’t done the maths to cost how their giveaway was going to be funded. The markets in London responded by dumping government bonds. In under five trading days, the yield on the 30-year Gilt rocketed by an unprecedented 150 bps (figure 1). It took a promise by the Bank of England to buy up to £65bn to bring yields most of the way back down.

The Bank justified its intervention thus: given the way yields were going, certain pooled vehicles for pension funds would have been left with net negative asset values and their collapse would have caused even greater damage to the Gilts market.

Those certain pooled vehicles were leveraged Liability-Driven Investing (LDI) funds, investing on behalf of just over one-third of all of private-sector pension schemes in the UK.

LDI funds are not purely invested in Gilts. If they were, this problem would not have mounted into a crisis. LDI funds do, however, rise and fall in response to Gilt yields, and that shock rise at the end of September – which became a spike as soon as the BoE announced its intervention – created all kinds of havoc that lasted well into October (there was another spike that month of similar magnitude). 

The trustee of one small scheme said it got a demand from the LDI manager late one evening to find £2m by 10.00 the following morning. “It was a mess,” he said, adding that the investment consultant to the scheme didn’t know what he was doing. The liquidity was taken from a diversified growth fund.

One global equities manager based in London said that its 15-year relationship with one pension scheme client was ended in a phone call as the mandate had to be liquidated to meet collateral calls.

There was clearly a lot more going wrong in this crisis than just selling down government bonds. The first thing was that so much of the LDI funds in the UK were leveraged. In a speech last November, the BoE said poorly managed leverage was the root cause of its intervention, referencing the likes of Archegos and LTCM. Leverage can amplify shocks in any asset class, even a stable one like UK government bonds. 

“We reckoned that in a crisis there was no way we would cope with the volume of trades to be processed”

Rob Skelton

Average leverage in LDI funds was 3-4 times, according to a survey in the first half of 2021 by IC Select, an Edinburgh-based consultancy. 

So record volatility in the bond market caused the fire sale of credit, equities and even illiquid assets to replenish collateral for those derivatives. 

Foreign investors were exiting Gilts because they didn’t have faith in the new political administration’s plan, or ability to pay for them. Why did UK pension funds keep faith instead of joining the exodus? The answer is that LDI was a hedging strategy. When yields went out of normal trading ranges, funds had to desperately try to keep up to hold the hedge. The complication was leverage, which amplified the spikes. 

In theory there is no need to add leverage to LDI. One of the first examples in the UK, the pension scheme of pharmacy giant Boots, more than 20 years ago, never did. But as Gilt yields dropped to record lows after the Global Financial Crisis and stayed there, ‘doing a Boots’ became prohibitively expensive as liabilities increased commensurately and funding levels fell. So UK pension funds opted to add extra returns from the leveraged Gilts. 

Embedded risk

The next problem was how widespread this practice was. While the blow-ups of Archegos and LTCM had considerable ripple effects across financial markets, each was a single event. By contrast, approximately 60% of private-sector DB plans in the UK used LDI. At the time of the crisis total assets in these strategies were £1trn, according to the Bank of England.

As a result of the crisis, thousands of schemes, each with nuanced liability profiles and fiduciary responsibilities, were in emergency communications regarding hundreds of funds with tens of providers and a similar number of investment consultants.

Fallout was not uniformly distributed. Many schemes emerged unscathed. To understand further requires exploring the mechanics of leverage in pooled LDI. According to Rob Skelton, head of investment research at consultancy, First Actuarial, look beyond any indicative, average leverage ratios to the maximum any fund can go to before it takes corrective action. 

This maximum – the tolerance permitted in response to yield movements – isn’t written on every fund factsheet but it is discussed among LDI fund providers and consultants. On one popular 20-year-duration product back in 2020, Skelton saw a maximum of 10-times leverage. “We told the provider that was crazy. They would get stopped out before they could raise the necessary collateral. The manager cut maximum hedging on that product to 5-times a few months later.”

In the discussion, it turns out the provider had modelled its tolerances on a 50bps shock movement. 

“One of the reasons we questioned their assumptions was that Gilts had just experienced a 90bps move in a few days,” explains Skelton. This was the hedge fund ‘dash for cash’ in March 2020, which held the record for causing the biggest move in Gilt yields prior to last Autumn (figure 2).

So upper limits on leveraging depend very much on buffers and vice versa. Modelling to a shock move of 50bps in government bonds might seem unsound now but pre-crisis, most organisations, including The Pensions Regulator, envisaged Gilts moving no more than 100bps within a few days. They had not moved this much in March 2020; First Actuarial’s message to clients was that they might.

Regulation in hindsight

How the world has changed. Post-crisis, the new guidelines for pension schemes recommend a 250bps shock absorber. There seems to be no pushback from the industry on this.

Related to buffers and maximum leverage is the time period for transferring collateral calls. Post-crisis, The Pensions Regulator has laid down five days as the new standard. But in the years prior to the crisis, LDI fund managers were being pushed to extend calling periods. Some managers, in response to some consultants’ demands according to Skelton, would allow investors up to a month to meet these calls. 

That record spike in Gilt yields in late September lasted five just days. When it began, some pooled funds were still waiting on liquidity called on 14 September. 

Then there is the variety of processes for getting collateral going into the crisis. According to Skelton, only one or two pooled LDI managers offered a kind of credit before trades were actually settled – this proved to be very beneficial once the panic started and speed was of the essence. 

Meanwhile, many pension funds were relying on their investment consultant to carry out all the trades needed to meet collateral calls. “That’s fine most of the time but we reckoned that in a crisis there was no way we would cope with the volume of trades to be processed,” said Skelton. In spite of the revenue appeal, First Actuarial were not prepared to take on that risk and had delegated this part of the process to a fund manager or platform provider. 

Other consultants tell a similar story. Cartwright, an actuarial and pensions consultant based in Farnborough, UK, had arranged for such a facility for its pension fund clients with Mobius Life, an independent fund platform. “Mobius could pull through capital into the LDI fund from clients’ other holdings, such as short-dated corporate bonds,” says Sam Roberts, director of investment consulting at Cartwright.

Thinking on their feet

Knowing how ill-prepared the machinery of some LDI arrangements were for a crisis helps understand why things fell apart, even though rates had been rising steadily in the UK since December 2021 and consultants had been exhorting clients to make increased liquidity provision throughout 2022.

Pooled managers introduced faster trade times as the crisis happened around them but old-fashioned protocols to effect contractual changes hampered adaptation. “Big pooled managers had arcane rules governing the movement of assets, such as the need for wet signatures,” said one trustee. “It was as if digitalisation had never happened.”

The industry was thinking on its feet, which is not an image it likes to portray. Even if blame is laid at the door of the hapless politicians, Kwarteng and prime minister, Liz Truss; even if Gilt yield moves were unprecedented record-breakers, the experience of operational checks and balances being overwhelmed is embarrassing.

Fast forward one year, however, and the situation appears much sunnier. Alongside their new LDI oversight services to explain all the revised guidelines to clients, pension fund consultants are discussing congestion around schemes offloading their liabilities to insurers. 

The queue is long because UK pension schemes find themselves, in aggregate, in their best ever funding position. This has everything to do with steadily rising interest rates and in spite of last autumn’s havoc. 

“Many schemes experienced trading losses from hedges lost at the peaks in September or October which would not otherwise have occurred,” says Simeon Willis, CIO at XPS Pensions Group. “This temporary loss of hedging effectively depleted reserves they otherwise would have emerged with once yields had normalised in late October.”

Levels of leverage have fallen in LDI funds, in part because of greater cautiousness but also because Gilts are offering reasonable nominal returns. 

Whether any advisers, funds or pension scheme trustees will be found negligent for their actions prior to normalisation remains to be seen. There may yet be litigation to clarify culpability. The website of leading London law firm, Pinsent Masons says it is advising on a claim for LDI losses against an investment consultant and asset manager. IPE could obtain no further details before going to press.