PRA consults on funded reinsurance

The PRA's consultation on funded reinsurance is a welcome development. This is an area in which the PRA has signalled interest for quite a long time, and clarity on its expectations is helpful.

A number of the proposals will be familiar to insurers, albeit they will likely lead to some increased formality and documentation. In some areas, many insurers will welcome the PRA's suggestions as they will help to ensure that the bulk annuity market is not being distorted by a small number of parties taking risks that others regard as inappropriate.

Aspects of the consultation seem to us, however, to raise practical concerns. Some of the proposals relating to how reinsurers are assessed, including the obligation to establish whether reinsurers’ default risks derive from correlated sources or business models, appear difficult to implement. This may be particularly problematic for larger, more established reinsurers. Other proposals may be reasonable, but it will be important to see how they are implemented (eg ensuring collateral assets are in a form that would allow MA compliance on recapture).

A lot depends on how the PRA interprets the idea of proportionality. Some of the measures would, if applied uniformly without regard to context, seem over the top. Others could potentially even lead to the opposite of the PRA's objective or cause.

Background to CP

The PRA has seen an increase in funded reinsurance in recent years, which it believes to have been fuelled by more, and higher value, transactions in the UK bulk purchase annuity ("BPA") market. If the reinsurers were to fail, the consequences for UK insurers in this market (and potentially UK pensioners) would be significant. As a result, the PRA wants to set out its expectations clearly.

In consultation paper CP24/23, released on 16 November 2023 (the "CP"), and its accompanying draft supervisory statement (the "SS"), the PRA focusses on three key risks which are raised by funded reinsurance, namely credit, collateral and systemic risks. As the PRA seems to be saying that it does not, as of yet, have concerns about systemic risks (ie it will monitor those risks for now, rather than address them), we do not comment on those.

The PRA's launch of this consultation is generally to be welcomed. This is an area the PRA has indicated an interest in for some time, and it is helpful for the PRA to make its expectations clear. Most of its comments are unsurprising albeit that some, as discussed below, may give rise to practical difficulties. Furthermore, this is an area of real importance to the wider economy, relating to what will be, for many UK pensioners, their main source of income.

Counterparty risk

As part of its risk management system, the UK insurer ceding the risk will be expected to consider the impact of a potential recapture of all ceded business, including on its SCR. With this in mind, insurers will need to set:

• internal investment limits for exposures to an individual counterparty, which should be "focussed on the idiosyncratic risk of a counterparty" , such that it may default independently of other counterparties in the market;
• an additional limit for concentration risk, based on the simultaneous recapture from multiple "highly correlated" counterparties; and
• an aggregate limit based on ensuring a diversified asset strategy as well as operational capabilities on recapture.

To implement those requirements, a cedant firm will need to understand the extent to which its reinsurers are "highly correlated". The PRA states that firms should assess "similarities in the risk profile of counterparties" operating in this market, but does not say how that should be done.

This seems like a sizeable burden. Assessing the credit risk of an individual counterparty can be a significant exercise, but the PRA's proposal would require a much deeper understanding of the various reasons reinsurers might default. Even after developing that deeper understanding, the insurer would need to be able to compare the risks between different reinsurers and then identify ways in which those risks are correlated.

That kind of exercise would likely need substantially more data than reinsurers provide at the moment, which raises a point for contracts. It is also not a given that reinsurers would agree to disclose the additional data. Even if reinsurers could be persuaded in that regard, the additional processes and expertise that UK insurers would need to put in place seem likely to result in significant expense. As a result, it seems likely that there would be practical difficulties with implementing the PRA's suggested approach.

We have a similar concern about the PRA's expectation that insurers should "analyse how the solvency ratio of their counterparties changes under various market stresses", and how this could inform their assessment of a stressed probability of default. This is said to reflect an existing requirement of the PRA's rules on internal models. However, if read too literally, we believe it sets an impossibly onerous challenge, given the detail the cedant would need about the reinsurer if it were to conduct that exercise properly.

In a presumably related point in the context of the SCR, the PRA indicates that firms should look to gather non-public information as part of their counterparty approval processes. It goes on to say, however, that such information should not be used to assign a lower probability of default to a counterparty than would otherwise be used. Is this an example of the PRA both having its cake and eating it? Why can the information provided only result in a more negative view of the reinsurer?

The practical issues seem important in and of themselves. Perhaps the bigger point, however, is the incentives they would create. The exercise of assessing large, well-diversified (across products, assets and geographies) reinsurers would be more complex, and so require more time and expertise (ie cost), than in the case of a small, mono-line reinsurer. If the same assessment needs to be done regardless of how likely the reinsurer is to default, it creates an incentive to transact with reinsurers with simpler balance sheets. That incentive would nudge the market in the very direction that the PRA is presumably trying to avoid. As such, a statement in the final SS to the effect that the measures can be applied in a proportionate way would be welcome. Even if the PRA's view is that this is implied, the factors it would take into account (and not take into account) when assessing whether proportionality has been properly applied would be a useful addition.

Collateral risk

The PRA is concerned that, in a competitive BPA market, firms might be able to offer pension scheme trustees a lower price by accepting lower quality collateral from their funded reinsurer, with the result that firms would then be in a more vulnerable position on recapture.

To address this concern, the PRA expects firms to establish clear collateral policies as part of their risk management processes. These policies should be closely linked to firms' limit setting process (see above) and would need to cover, at a minimum:

• approaches to credit assessments;
• valuation methodologies by asset class;
• matching adjustment ("MA") eligibility monitoring;
• SCR modelling; and
• investment management approaches on recapture.

In the context of calculating the extent to which collateral supporting funded reinsurance arrangements mitigates risk for SCR purposes, the PRA's expectations include a requirement to stress assets held as collateral on a look-through basis to reflect the risks that a firm would face on recapture. Other requirements include considering possible mismatches between the stressed value of the underlying insurance liabilities and the stressed collateral required under the funded reinsurance arrangements. Firms should also consider the risk that counterparties would not be able to replenish the collateral portfolio in stressed conditions.

These proposals may all be reasonable, but, as with the requirements related to credit risk, much depends on their application. For example, requiring firms to consider how they would meet the MA criteria if a recapture occurred seems fair (assuming MA reliance would be part of their recapture plan (see below), which does seem highly likely). However, the majority of reinsurers in this market are high-quality counterparties with a very low risk of default. Disregarding this, with the result that investment management strategies are forced to be highly conservative, will impact the costs of BPA transactions (ie trustees will pay for this conservatism with higher premiums) and could well work against the objectives behind the liberalisation of the categories of MA eligible assets, as reflected in the PRA's CP19/23. These increased costs need to be justifiable and rooted in real risks. For example, the importance of collateral, and so the need for prescribed restrictions, is less when a reinsurer's SCR is 200% than at 100%.

Recapture plan

Firms will be required to formulate and document a recapture plan for their funded reinsurance arrangements. These would need to demonstrate that the firm's business model can survive a single recapture event and multiple recapture events from correlated counterparties.

Recapture plans should also establish a clear and structured decision-making process for assessing whether ceded business should be recaptured when optional contractual termination event clauses are triggered.

The burden of this exercise could be significant. It would be helpful if the PRA could clarify how firms should approach this. In particular, a confirmation that an exception, or at least a significant transition period, will be granted for existing relatively low risk reinsurance (eg if a reinsurer is in a sophisticated jurisdiction and is highly capitalised) would be a positive addition.

Contractual mitigations

The PRA notes that appropriate contractual protections should be introduced into funded reinsurance arrangements. Firms should adopt internal guidelines setting out the minimum protections that should be sought and the rationale for seeking them. These include the approach to termination rights, substitution of collateral assets, rules on investment management (including valuation), concentration limits, rights to obtain information (including information that may be commercially sensitive) and choice of applicable law.

In our experience, UK insurers already include many of these types of contractual protections. We would not, therefore, expect this aspect of the SS to change the high-level approach in most cases (though the detail of the clauses may need to be different). If, however, arrangements are being entered into without including these protections, clarity from the PRA is a positive development to ensure that some are not seeking to gain commercial advantage by taking risks that most insurers in the market regard as inappropriate.

Conclusion

In general, moving towards a more detailed framework in the manner proposed by the PRA is prudent, particularly in light of a growing BPA market and the fundamental objective of ensuring the insurance industry operates effectively for underlying policyholders.

Firms will need to consider how best to meet the PRA's expectations. This may mean looking at existing funded reinsurance arrangements and considering how to respond if those arrangements fall short of expectations. In our experience, most insurers already spend a lot of time thinking about the types of contractual protections the PRA has referenced. They also put considerable effort into understanding how a recapture would be implemented in practice, albeit that the new proposals may result in increased formality in that respect.

However, some aspects of the CP proposals (in particular in relation to assessing and monitoring counterparty credit risk and reinsurers' solvency ratio changes under stress) impose obligations that seem likely to give rise to practical issues. Insurers will need to assess if these are proportionate and provide feedback on the CP if they feel that they are not. In our view, the PRA should consider these issues in detail, and should in particular consider if being unclear about how proportionality will be applied might incentivise the types of risks the PRA is looking to discourage.