[Skip to content]

Sign up for our daily newsletter
The Actuary The magazine of the Institute & Faculty of Actuaries

Risk management: Opportunity knocks

Many insurers understand the benefit of evaluating reinsurance counterparties’ creditworthiness. 
However, this risk is often evaluated simply on a qualitative or statutory basis, or only when the cover is placed. Credit risk is often overlooked, underquantified or discounted as negligible once reinsurance balances are booked on financial statements, and even where quantified is rarely part of a strong feedback loop aimed at improving future placements.

Further advances in quantifying the credit risk inherent in reinsurance are possible, and may change current thinking about reinsurance transactions and liabilities.

The focus on modelling over the past few years, driven in large part by the Solvency II 
regulations, offers the opportunity to enhance the quantification of reinsurance counterparty risk. These modelling tools provide insurers with an opportunity to take a more quantitative approach to evaluating the bad debt inherent in past, current and future reinsurance transactions, and achieve a number of new benefits.

Credit risk and the reinsurance industry
Credit risk is the chance that a financial contract’s terms may not be met because of bankruptcy, insolvency or other financial impairment. Typically, credit risk has three components: the chance that the counterparty will experience a credit event; the balance due from the counterparty at the time of default and beyond; and the proportion of the value of the obligation that becomes unrecoverable upon default. The second point in particular is often overlooked but, of course, recoveries already received and in the firm’s bank account are usually safe.

Many insurers evaluate reinsurance counterparty risk using a variety of 
partial and limited approaches rather 
than comprehensive assessments. 
These approaches include:

• Approved reinsurer lists. Many primary companies that cede insurance exposure have a list of reinsurers deemed acceptable. These lists provide some protection against ceding to less creditworthy counterparties. However, companies differ in their approaches to constructing and maintaining these lists. Furthermore, lists do not help to quantify the extent of potential bad debt inherent in ceded liabilities.

• Maximum exposure limitations. 
As an extension to the approved reinsurer list technique, many insurers set a cap on the aggregate credit exposure to any single counterparty, often differentiated by the credit rating of the counterparty.

• Credit enhancements. When the reinsurer is deemed less creditworthy by the ceding company, there are several ways to amend the credit status of the reinsurer, including withholding all or part of the ceded premium, the posting of a letter of credit or a trust fund as collateral for future obligations, and the taking over of the reinsurer’s financial responsibility by a parent company or third party.

• Cancellation/termination/commutation. A special termination clause in a typical reinsurance contract allows the ceding company to cancel or commute the reinsurance contract.

• Differentiated reinsurance prices. 
In rare situations involving quota share reinsurance, ceding insurers select a group of reinsurers to cover the same underlying exposures at different reinsurance prices.

Critics question qualitative assessments
Critics often regard these qualitative approaches as incomplete and inadequate in their assessments of reinsurance counterparty risk. For example, the approved reinsurer list approach fails to further distinguish the credit status of approved reinsurers.

Without a comprehensive risk analysis, ceding companies would have difficulty determining how much security is really needed in the credit enhancement approach. The cancellation/termination/commutation approach can, at best, stop the bleeding when the reinsurer’s credit is impaired in a credit event, and comes at the cost of the removal of the reinsurance cover originally seen as beneficial.

ERM and quantitative tools
Enterprise risk management (ERM) is prompting insurance companies to embrace quantitative approaches to evaluating reinsurance counterparty risk as part of an overall risk profile. Quantitative approaches in use today include:

• Factor-based approaches. Solvency II 
and other regulatory regimes apply a series of factors to reinsurer balances, often grouped by financial strength ratings of counterparties, to calculate an estimate of uncollectable reinsurance. However, these factors only offer simplifications of real potential uncollectible reinsurance.

• Modelling. Some insurers use differing sophistication levels of deterministic or stochastic modelling to quantify potential uncollectable reinsurance balances. 
These models evaluate the expected amount and, for the most sophisticated, the distribution of potential reinsurance bad debt. Various modelling approaches, including deterministic, expanded 
factor-based and stochastic methods, are currently used in the insurance industry.

A credit event/recovery rate approach
One stochastic approach to modelling reinsurance counterparty risk is to use information available from the bond market, where credit risk has been studied in great detail. Each of the following inputs is publicly available and updated at least once a year:

• Financial strength ratings.

• Transition matrices or aged probabilities of default, which provide probabilistic expectations of the movement in financial strength ratings over time, including 
default rates.

• Recovery rate, or the amount of financial recovery historically available in the case 
of default.

Reinsurance balances are grouped into ceded reserve balances at a certain point in time and expected balances to be ceded in a relevant future period (for instance an upcoming accident year). Future balances are often split between catastrophic and non-catastrophic exposures, reflecting differences in variability and payment pattern of future ceded balances. In this Monte Carlo simulation approach, the ceded reserve balances also vary by trial.

For each bucket of reinsurance recoverables, many trials are run, and in each trial the simulated ceded balances 
are repaid over a multi-year period. 
During this time, the default of reinsurers is also modelled, either via modelling the reinsurers’ financial strength ratings over time using the predefined transition matrix to capture potential changes in each time period, with reference to which rating the reinsurer had been left with in the previous period, or modelling straight to a default event by reference to observed historical probabilities of default which vary by time and initial rating.

Once the beginning ceded balance by counterparty has been determined, the first modelling step is to determine whether a stress scenario is triggered. Several events can be used to trigger these stressed environments. In the case of quantifying reinsurance counterparty risk, the idea is to capture events that would shock the financial strength of much of the reinsurance industry:

Catastrophic events (natural and 
• Hurricanes

• Other windstorms

• Earthquakes

• Liability catastrophes

Financial catastrophe events

• Bond market downturn

• Stock market crash.

A large catastrophic event, a series of catastrophes over a period, or some combination of a large catastrophe and a downturn in the financial markets may trigger a stressed environment. If so, we would expect more negatively biased transition and recovery rates.

Care must also be taken to consider for each firm any potential impact of acceleration of downgrades (that firms downgraded are more likely than an average firm to be downgraded to the next step) or reaction of the firm (that it will fight and manage to achieve a bounce back in its rating).

When a reinsurer defaults, a credit event occurs. The unrecoverable amount depends on the remaining unpaid ceded balance and the simulated recovery rate.

An insurance company can determine its distribution of potential bad debt due to financially unrecoverable reinsurance by running all ceded balances with counterparties of various financial strength ratings through this type of model. The distribution can then be incorporated into an economic capital model.

The future
While many insurers understand the importance of evaluating counterparty risk, advanced evaluation techniques have not been widely embedded. For example, it is rare for firms to evaluate different potential providers of reinsurance with reference to both the cost savings they would achieve through accepting lower rated security, and the magnitude of the risks that lower rating would result in (which may be very different for different lines of business), especially 
at a whole account level. The use of an approved list provides some protection. 
But this approach does not begin to differentiate the relative creditworthiness 
of the entities within each group.

Other qualitative or basic quantitative approaches are valuable in helping to control exposure to and approximate 
the magnitude of counterparty risk. 
But they represent only partial solutions and do not allow for full quantification of the exposure.

Sophisticated modelling improves measurement of the true costs of reinsurance transactions. These approaches consider all sources of loss, including counterparties’ ability and willingness to pay. As Solvency II develops, and firms move closer to a single model of risk, appropriate reinsurance pricing differentiation and better quantification of expected bad debt can be achieved.

Martin Cairns

Martin Cairns is a senior consultant with Towers Watson