How does inflation affect insurance-linked securities?
How does inflation affect insurance-linked securities?
Headline inflation across key economies has hit levels not seen in over 40 years, and what was initially thought to be transitionary now seems to be more persistent. Seemingly caught by surprise, central banks are taking a more robust stance to tame it.
We ask what inflation means for insurers, and more specifically for investors in ILS.
How inflation risk affects insurance claims
The accepted wisdom in the insurance industry is that “the longer it takes to settle a claim, the more expensive it becomes”. For example, if a home is damaged in an earthquake, a policy may need to cover temporary accommodation for its occupants while repairs are being undertaken. The longer those repairs take, the more expensive the loss is for the insurance company. Covid-related supply chain disruptions are another good example.
It should be easy to see that higher inflation makes this situation worse – costs are increasing much more rapidly with time, impacting the ultimate cost of any repairs.
Inflation for insurers is not just about macroeconomic risk
For insurers, there are other forces which are driving up insured values, and ultimately how much they may have to pay out on a claim. For example, enhancements may be required to a damaged building (such as access for disabled persons) to comply with the newest building code. Or certain materials (such as wooden roof shingles) may no longer be allowed due to tightened requirements owing to past catastrophe losses (e.g. wildfires).
There have also been some observable trends in claims adjusting practices and in policyholder litigation which have impacted property insurers (sometimes referred to as “social inflation”). One of the more pronounced has been in the practice of the assignment of benefits by the policyholder under a property insurance policy to a contractor. In such instances, the first time an insurer may become aware of a claim is after the contractor has completed the work and presented the insurer with an invoice.
This practice was particularly pervasive in Florida and resulted in a significant increase in loss cost, particularly after Hurricane Irma. Attempts by the insurance industry to limit this practice by amending policy wordings met with limited success in front of the plaintiff-friendly courts. Insurers and reinsurers will need to adapt their pricing to reflect the observed long-term trends in loss cost that cannot be directly assigned to known identifiable parameters.
The examples above can be thought of as raising the insured value “per property”, or more generally “per exposure unit”. However, the growth rate of an insurer’s overall exposure is also a function of how many of these units it covers . This is tied to economic growth in a given region (which could be a state, city or even postcode). The most obvious example is the amount of new construction taking place in Florida because people are still moving there, not only to retire but increasingly to live and work in the era of more flexible working patterns. Even if there were 0% inflation in a macroeconomic sense, insurers’ exposures would still be on the up.
What can insurers do to manage inflation-driven exposure growth?
One of the main defences that insurance companies have against rising cost exposure is regularly adjusting the Total Insured Value (‘TIV’) of a given insured property. Most do this on an annual basis, which serves as the basis for calculating the premium. Changes in the TIV can also reflect projected higher costs, following a change in the building code for example.
As well as managing the insurance company’s exposure, this has the added benefit that it can limit the risk that a policyholder ends up under-insured - where declared insured values are less than the actual replacement cost, and where the insurance company may not be obliged to cover repairs in full.
Another tool in the insurance company’s kit to manage cost inflation is to restrict certain elements of the insurance cover to paying an actual cash value (ACV) for the damaged property instead of the full replacement cost. This has become increasingly common in the insurance of older roofs on residential properties.
Insurers have also changed their practice in the use of deductibles to mitigate the impact of inflation on their loss cost. Historically, US insurers usually had a flat dollar deductible of $100, $250 or $500 irrespective of the TIV of the property insured, although these are less effective in a world of increasing inflation and higher TIVs.
Consequently, many insurers are now applying a deductible that is a percentage of the TIV, and hence increases as the TIV does. This might be 1%, 2% or sometimes higher depending upon the peril insured. While not popular with consumers or regulators, the trend has been for these percentage deductibles to apply to catastrophe perils (where many policies are impacted by the same event) while sometimes maintaining the flat deductible for non-catastrophe perils.
Do catastrophe bonds offer any protection against inflation?
Many of our clients have raised the question of whether catastrophe bond values risk being eroded by inflation. Tradable catastrophe bonds contain certain features that seek to preserve the risk-reward relationship during the life of the bond.
The key protections are:
The annual reset
In a multi-year catastrophe bond, the amount of risk is re-assessed annually. The layer of cover is changed so that the risk in it is the same in a subsequent annual period as it was in the initial annual period.
The reset is accomplished by running the catastrophe simulation model against new/revised exposure data, reflecting the updated insurance values (capturing if they have increased for any reason, including inflation), adjusting the attachment point (the point at which the bond can start to suffer a reduction in principal) and layer structure (the tranching of the protection buyer’s protection tower) to maintain a consistent level of risk.
In some catastrophe bonds a limited amount of fluctuation in the risk from year-to-year is allowed and the future coupon is adjusted according to a pre-agreed scale.
The growth limitation factor
This feature is applied to bonds with an indemnity trigger. Indemnity triggers are insured loss levels which, when passed, result in a cash call (against the bond in favour of the protection buyer). Growth limitation is applied retrospectively, after the occurrence of a loss during the current risk period.
At the time of a loss event, the TIV in force at the time of the loss is compared against the TIV provided at the inception of the instrument. If the difference is greater than an agreed percentage (usually limited to 10%) then any potential recovery is reduced proportionately for any growth in excess of the limitation factor.
This feature was originally designed to provide protection buyers with some reduction in their basis risk (the mismatch between what can be recovered and what might otherwise have been recoverable were it not for the limitation being imposed) by allowing for some limited growth in the exposure base during the risk period. But it also protects investors in times of high inflation, when the TIV could be rising fast.
Floating return on the collateral
The floating rate nature of catastrophe bonds means that, as central banks hike rates to combat inflation, their coupons will rise.
What else can ILS managers do to reflect the impact of inflation properly?
A prudent ILS manager should also make adjustments in its view of risk to accommodate for the risk posed by inflation.
First and foremost, one needs be conscious of when the exposure data submitted for analysis was prepared. For an indemnity trigger catastrophe bond, or private collateralised reinsurance transaction, it is not uncommon for the exposure data to be captured three to six months prior to the inception date of the instrument. Then one must consider the time elapsed until the potential event may occur.
Consider a catastrophe bond with an indemnity trigger providing protection against North Atlantic Hurricane risk. It incepts on 1 January 2022.
Losses will then be expected in the late third or fourth quarter (August to October), when peak hurricane activity is observed. If the exposure data submitted for modelling the transaction is as of 30 June 2021 – six months prior to 1 January 2022 - this means that the in-force exposures (without any new business growth) will have experienced a further 12 months at least of inflation-driven exposure growth.
This is not customarily included in the modelling projections provided by the sponsor when the instrument is offered to ILS investors.
In times of very low or no inflation, this timing gap was not much of a cause for concern. In the current environment, however, a prudent ILS manager should adjust its expected loss and attachment probabilities to reflect the additional exposure growth due to inflation (as well as other variables).
Another thing that one must be conscious of is that the modelling of index-triggered ILS instruments (i.e., Industry Loss Warranties and Property Claim Services (PCS) or Perils Index catastrophe bonds) is based upon the modelling agent’s most recent release of the model and the underlying industry exposure database (IED).
Crucially, while the models are subject to regular updates or releases, the underlying IED may or may not be. In some peril regions, IEDs are not updated annually. Also, the timing of when a reset is calculated may not correspond with the release schedule of an updated model or IED which can cause further slippage. As in the case described for an indemnity triggered instrument, a prudent ILS manager would also adjust the IED by applying a proper inflation factor to reflect that the assessed underlying exposure might be out-of-date.
Catastrophe models used by the insurance and capital markets are increasingly sophisticated, regularly updated and calibrated to reflect the latest scientific research on the physical perils and the damage that they can cause. Even so, a residual degree of model risk remains.
This is why protection sellers receive a healthy multiple over the model expected loss value as compensation for the risk assumed. However, as we all know, a model’s output is only as good as the quality of the data it has as its input. Ensuring that the data used in simulating projected losses properly reflect the most current view of exposure (including potential effects of inflation) is critical to achieving a realistic view of the risk and rewards posed by a given transaction.
Inflation risk: Not slain, but tamed?
One of the inherent challenges of selling insurance protection is that one must set a price for the costs of a future event. As we have outlined, the insurance industry and ILS managers have different tools and features at their disposal that seek to compensate or mitigate the impact of inflation.
Industry practices such as insurance-to-value, restricting coverage to ACV or using dynamic deductibles are designed to ensure that the inflation does not result in an increase in uncompensated exposure. Similarly, features such as the annual reset, the growth limitation factor and the floating rate nature of catastrophe bonds also offer an element of protection against the impact of inflation.
Last, not least, a prudent ILS manager can adjust the exposure data used for a specific transaction to compensate for the inherent time lags between when the exposure data was collected and the point in time when a loss might be expected to impact the instrument being analysed. While these various tools and practices may not eliminate the risk posed by inflation to investors in ILS, they should tame them. We’ll have to leave the slaying of the inflation tiger to the central bankers.
- Economic and Strategy Viewpoint - Q4 2022
- China’s latest Covid crisis, war in Ukraine: observations for investors
- Outlook 2023, US multi-sector fixed income: yields up risk down
- Outlook 2023, Sustainability: five trends to watch
- Outlook 2023, Japanese equities: re-opening to provide boost
- The small cap sustainability opportunity