DISCLAIMERS: The opinions expressed herein are those of the participants and do not reflect the opinions of Milliman or its employees. Facts cited by the participants are presumed to be accurate and have not been externally verified. This document may not be used in whole or in part without the express written permission of Milliman.
A panel of senior insurance investment professionals joined Milliman in early July 2021 to share some thoughts on current themes in asset-liability management (ALM) and capital management, in light of the events of the last year.
This meeting was conducted under the Chatham House rule, and below we summarise the key points from our engrossing and wide-ranging discussion. We are grateful to those who took part.
Moderators
- Russell Ward and John Jenkins of Milliman.
Participants
- David Campbell
Financial Risk Director in Lloyds’ Insurance Risk division. - Joseph Cannavan
Head of Capital Management at Royal London. - Chandra Chadalawada
Head of ALM at Legal & General. - Michael Eakins
Chief Investment Officer at Phoenix Group. - Paul Fulcher
Group Capital and Investment Executive at Just Group. - Andrew Hague
Head of Financial Risk (Group) at NFU Mutual. - Simon Iversen
Director of Annuity Portfolio Management at M&G.
Is the current uptick in inflation transitory or something longer-term? For example, with evidence of recent wage inflation in the US, how does the panel feel about the outlook for inflation and interest rates?
Participants agreed on the immediacy and importance of this issue. One common theme was that, as ALM practitioners, there is a need to become as comfortable as possible with the implications of all reasonably plausible interest rate and inflationary scenarios. However, we are now facing a range of plausible outcomes wider than it has been for many years. This raises some challenges, for example that:
- In a perfect world, rates and inflation would be immaterial to ALM, but most manage against more than one metric, and these metrics may point to different courses of action.
- From a first-order perspective, any detrimental impacts of inflation or rate changes are fairly straightforward to assess, though convexity effects could see significant changes to risk exposures. However, it is much harder to gauge other consequential impacts if rates and/or inflation shift significantly. What would be the effect of a yield spike on liquidity? Would the risk of defaults and downgrades rise, how quickly and by how much? What happens to credit risk in an undesirable inflation scenario? Forming views is difficult as we have little recent historical experience to guide us as it has been almost 30 years since we last saw significant rises in interest rates.
We also discussed the role and possible action taken by central banks in this regard. As their core remit is inflation control, it is likely rates could be raised if this inflationary uptick becomes more pronounced or persistent. Participants foresaw a risk that central banks could act, whether by reducing quantitative easing (QE) or by hiking policy rates, faster than markets have priced in with negative implications for financial markets and asset prices.
There was also some concern over the potential for differences to emerge in the timing and extent of action taken by central banks in different countries. Participants described an undesirable situation for significant investors in non-sterling fixed income instruments; if the US Federal Reserve, for example, were to move before the Bank of England, we could see the dollar appreciate significantly against the pound, driving increased collateral requirements for those hedging with cross-currency swaps. This worry was compounded by the current environment in which almost any Fed announcement can generate significant market movements in response.
Some participants saw the prospect of increased market volatility as not necessarily a bad thing as it may offer opportunities to deploy new investment at more attractive rates than have been available in the tight credit spread environment for the first half of 2021. Being prepared to widen the range and geography of asset investments to minimise exposure to any particular asset class or region was noted as increasingly important.
Some participants saw the prospect of increased market volatility as not necessarily a bad thing as it may offer opportunities to deploy new investment at more attractive rates than have been available in the tight credit spread environment for the first half of 2021. Being prepared to widen the range and geography of asset investments to minimise exposure to any particular asset class or region was noted as increasingly important.
Participants observed that the difficulty is not always in the finer technical points of hedging, but in the high-level question: “What do we hedge?” We discussed the example of the Solvency II Risk Margin, widely mooted to be lowered and made less rate-sensitive in an upcoming review. Should one hedge against the Risk Margin as it is now, or the Risk Margin as one expects it to be in say five years’ time?
Many of us would like to see a move to a higher-yield environment in the future. Isn’t the difficulty then in the transition from where we are now to that new environment? Do we feel central banks really can raise rates in the face of current levels of public and private sector debt?
Participants observed that, on credit spreads, the ideal situation may be one in which there is more differentiation available between credits but without an across-the-board broadening of spreads indicative of a significant increase in the risk of downgrades and defaults. Though there would inevitably be a reckoning for some so-called “zombie” firms.
Industry-wide, UK insurers may view a rise in interest rates as broadly positive, as it could make a number of insurance products more attractive for consumers. However, there is likely to be some price to pay in terms of an adverse impact on the value of current assets. Participants observed that the best situation we can hope for is to find a sweet spot with better long-term returns but where asset values don’t do too badly on the way there.
We also discussed a situation in which inflation abates, removing a factor that might mitigate rising ratios of debt to gross domestic product (GDP), bringing back into sharp focus questions around the sustainability of debt loads across Europe and in the UK. Opinions differed here, with some concerned that debt trajectories might drive significant sovereign downgrades over time and in themselves might provoke unpalatable responses from central banks. Others saw relative performance as a key factor noting that if comparable economies all had similarly high debt-to-GDP levels, market impacts from an overall elevation of debt levels may be limited.
Central banks (ECB and the Fed, not just the Bank of England) were noted to have handled the COVID-19 crisis well, especially their coordination of monetary policy loosening in the first quarter (Q1) of 2020, and the stability this helped bring to markets in an otherwise very turbulent period. However, looking now in the other direction, towards a potential tightening of monetary policy, some questions were raised over the competing priorities of central banks and their ability to maintain coordination as they try to normalise policy.
One point of debate was around how any tightening might be implemented—increase to policy rates or a reduction and then reversal of previous QE programmes. Some participants noted that, for the Bank of England (BoE), while the former was the priority under previous governor Mark Carney, under Andrew Bailey the latter appears to be the more pressing concern. This stance was felt to be consistent with the view that QE has been a useful tool for the BoE in tackling recent crises but needs to be unwound now, restoring headroom for it to be deployed again in future.
Broadening the discussion, have we relied too much on equity-bond diversification in recent years? Do we need to look elsewhere and possibly extend hedges if this strategy is becoming less effective?
Participants noted that the overall impact of the correlation between equity and bond returns is not always straightforward and in particular depends on the product mix. For example, during the 1990s with-profits business suffered as equities fell but bonds appreciated as interest rates dropped.1 On the other hand, this combination should be less problematic for the value of unit-linked business. Some participants agreed that the correlation might now have flipped, with rates up causing falls in real asset prices—the inverse of the relationship seen in the '90s and '00s. In terms of crises, the bursting of the dot-com bubble, the global financial crisis and the COVID-19 pandemic all involved falling rates and falling real asset prices.
Even putting aside the direction of the equity-bond return relationship, the current situation regarding the up/down asymmetry of prospective rate changes was noted as a likely constraint on the ability of bonds to mitigate equity falls. In the UK, rates up 300 basis points (bps) would take us to familiar territory of only a few years past, whereas rates down almost certainly means the terra incognita of negative interest rates.
Moving on to liquidity. During the COVID-19 crash, we had a so-called “dash for cash.” Do we feel liquidity is being managed effectively? Are these considerations starting to weigh more heavily on ALM practice, particularly around derivatives?
Some participants contended that liquidity risk has always been a forefront consideration, but one that has increased in importance since the COVID-19 crisis. Participants noted that many insurers are increasing their exposure to high-yield, illiquid assets, but that this introduces additional liquidity risk. However, others noted that even traded corporate bonds, for example, may not be eligible collateral and may be hard to liquidate in a short-term liquidity squeeze.
Turning specifically to derivatives, participants noted that, as insurers extend their range of assets and in particular increase exposure to non-sterling fixed income, you need liquidity to post on cross-currency swaps. However, cash and gilt credit support annexes (CSAs) take up a lot of liquidity and so widening the range of assets eligible to be posted as collateral can be attractive—though there is a cost to that.
Overall, the sense was that liquidity risk does not need to be an inhibitor of an insurer’s ability to use derivatives. However, it is an area that can be very complex and challenging to model and manage, requiring a serious commitment of resources. The rewards though are improved oversight and insights into the performance of hedges alongside scope to increasingly include liquidity as part of a well-rounded optimisation of investment.
Do we inevitably end up with a so-called “barbell” liquidity strategy in which the middle ground on the liquid-illiquid spectrum is hollowed out?
There was agreement that this strategy may become more popular, but the participants did not see this as a bad thing. During the market stress of March-April 2020, it was noted that semiliquid assets tended to become illiquid thus providing neither liquidity support nor attractive returns.
Finally, the Quantitative Impact Study (QIS) exercise of the Prudential Regulation Authority (PRA) is imminent. The details are still to emerge but we would welcome the panel’s thoughts on possible implications.
The panellists noted the considerable practical challenge of delivering high-quality responses to the QIS alongside regular planning and reporting cycle activity and the bank’s Climate Biennial Exploratory Scenario (CBES) work.
Some also felt the rewards may be somewhat limited, given PRA indications that any changes should be expected to be broadly neutral in terms of overall industry capital requirements. Other participants noted that even if overall levels are similar, there are likely to be winners and losers in terms of individual firms and strategies.
Finally, it was noted that a less volatile matching adjustment (i.e., less loss-absorbing of spread risk) would make a buy-to-hold investment strategy tougher to follow. Looking back to the earlier discussion around the risks to spreads, potential for a less loss-absorbing matching adjustment (MA) to combine with a credit spread expansion would represent a scenario of some concern.