- Pension funds have been adding in the order of £100bn of notional interest rate exposure a year for the past couple of years.
- Given the high level of hedging that is already in place, we conclude that the current pace of rate hedging can only continue for at most another three years
- We are approaching Peak LDI, as detailed below
Liability Driven Investment (LDI) has been one of the dominant trends in UK Defined Benefit (DB) pension scheme investment during the past 15 years. But this trend may be coming to an end: the current pace of increasing hedging levels can only continue for at most three more years.
Our survey of key market participants suggests that notional interest rate and inflation hedging exceeded 75% of private sector DB assets by March 2017. It is clear that, since its emergence around 15-20 years ago, LDI has grown incredibly rapidly. Concurrently, over the last 20 years, UK real yields have dropped by over 5%. While some of this reflects global trends, UK real yields have fallen by almost 2% more than global real yields. In part this will have been due to LDI activity by DB pension funds. This drop in yields has meant LDI asset portfolios have delivered very strong returns.
Section 2 of this paper provides a brief history of DB pension scheme investment. The second half of the 20th century could be characterised as the Age of the Equity Risk Premium. The early 21st century has been the Age of Unrewarded Risk – when increasingly interest rate and inflation risk was seen as unwelcome and schemes sought to hedge it. Our view is that we are now close to a transition in the UK where DB schemes will only increase rate hedging at the margin, in an opportunistic manner. We call this the Age of Peak LDI.
Having looked at drivers of global real yields in Section 3, and UK specific factors in Section 4, Section 5 sets out the analysis that shows we will reach the Age of Peak LDI by 2021 at the latest. In short, we believe that: c£1.2tn of notional interest rate risk is hedged today; in the past couple of years schemes have added around £100bn of notional hedging a year; and schemes will not materially hedge above assets levels of c£1.5tn. Our conclusion is not materially changed by most factors that affect pension scheme funding and investment: deficit payments, accrual, equity returns or even payments of substantial levels of transfer values. The main factor that could delay the peak is if there is a rapid slowing of hedging, although one could argue that this implies the peak has already occurred.
Given the scale of the flows of pension scheme money into hedging assets, an abrupt change to those flows will likely have a material impact on yields (both nominal and real) and consequently on DB scheme funding levels. Buying will be affected by sentiment, and sentiment will be affected by the pace of buying as well as broader asset price and yield moves. We have not attempted to predict how the transition will play out, but do believe it has meaningful consequences for both schemes and gilt markets. Our analysis builds on previous work, and extends it by considering: leverage; non-gilt sources of hedging assets; rate hedging rather than inflation hedging; and when hedging might peak and slow dramatically, rather than when it will actually stop.
We examined a number of economic research papers that seek to assess why the global real yield is at the levels seen today, and where it might go in future. Our summary is in Section 3. While the approaches differ, the main points of consensus are that: We should not look to real GDP growth as an indicator for where real yields might end up. The central expectation, globally at least, is for modest yield rises (<1%). There is a great deal of uncertainty around the central expectation, and real yields could shift very materially up or down. This would be an argument in favour of LDI to reduce this risk, unless one had a strong directional view.
Having explored the global outlook, in Section 4 we then focus on supply and demand issues in the UK. We can think of UK yields as having a global real yield component and an idiosyncratic UK component:
- Global real yields have dropped due to factors economists can largely explain, in hindsight; and
- The UK idiosyncratic component appears to have been stretched considerably due to the weight of UK pension scheme demand.
We examine some time series of real yields in the UK vs the US (where we utilise the US as a global proxy) which demonstrates evidence of this distortion in the UK. This present over-valuation (high price) of index-linked gilts is in no small part due to perceived and actual UK DB demand. Though we do not forecast a large move in global real yields over the medium-term, there is room for considerable cheapening in the UK as a result of the following related forces which could come into play:
- Any softening in the LDI “unrewarded risk” philosophy prevalent in the UK pension fund industry.
- A move from the Age of Unrewarded Risk to the Age of Peak LDI whereby UK pension funds’ demand for gilts is largely satiated and becomes marginal to yield levels.
- Market sentiment recognises that we are close to the Age of Peak LDI, and that this transition could happen soon.
While we are chastened by the fact that commentators have been saying ‘yields will rise faster than predicted by markets’ for the past decade, when in practice yields have continued to fall further, at some point this will play out. However, the key question is how the combination of global and UK components will move in aggregate over the medium term. In writing this paper we have relied, in part, on a proprietary survey of key market participants and also extrapolated from known data. The reason for this is that key data simply does not exist. In particular, despite the wealth of data captured on DB schemes, there is no definitive measure of the degree of aggregate hedging undertaken to date, or the pace at which those hedging levels are altering. Given the importance of yields and yield risk to pension schemes and capital markets, in our view this gap in data could and should be closed. Our thoughts on further work are set out in Section 6.
UK DB pension schemes promise pensions to over 10 million people, underwritten by thousands of UK companies. Collectively they invest c£1.5tn of assets, hundreds of billions of which supports the UK Government bond market. Our work suggests that a fundamental shift will occur in the next few years. That is important to other market participants both UK and overseas based; the “buyers of last resort” are soon to become more price sensitive. The ramifications of this shift are important to understand, given the consequences for all involved.