Many countries have formalised solvency insurance programmes that provide payments to policyholders of insolvent insurers. Like most insurance schemes, these programmes operate most effectively when failures are firm-specific. Using Canada as a case study, experts examined the issue of using ex post guaranty fund assessments following a peak peril loss resulting in the simultaneous failure of multiple insurers. This research paper, written by Mses Mary Kelly, Anne Kleffner and Mr Grant Kelly, was published in Geneva Association’s quarterly journal.
Many countries have formalised solvency insurance programmes, or guaranty funds, that provide claims payments and sometimes refunds of unearned premiums to policyholders of insolvent insurers.
Guaranty funds operate most effectively when failures arise from idiosyncratic issues such as poor underwriting, inadequate pricing or reserving practices. Like insurance, the efficient operation of a guaranty fund requires a large number of similar exposures independently exposed to loss. As such, the failure of one insurer should not be highly correlated with the failure of another insurer. However, a large catastrophic loss could result in several insurers being impaired and possibly wound up, violating the required independence property.
Due to the dramatic growth of catastrophic losses over the last decade, a concern for all stakeholders is the effect of catastrophes on insurer solvency. Catastrophes create challenges for the private insurance market and the guaranty fund because losses are not independent. Furthermore, the larger the potential losses, the more difficult the challenge.
Catastrophic losses have been growing dramatically over the past decade. Swiss Re in 2017 noted that in 2016, catastrophic losses worldwide accounted for $175bn in total damages and $54bn in insured losses. As catastrophic losses increase, a concern for all stakeholders is the potential impact on insurer solvency.
Historically, catastrophes have not been a primary cause of insurer insolvency in Canada. The costliest natural disaster in Canada was the May 2016 Fort McMurray wildfire, with estimated insured losses of C$3.58bn ($2.7bn).
This size of loss does not pose a threat to insurer solvency. However, a severe earthquake under Vancouver or Montreal would cause losses greater than ever experienced in Canada. In a report commissioned by the Insurance Bureau of Canada, the 1-in-500 year earthquake that was modelled – a 9.0 earthquake in the Cascadia subduction zone, approximately 75 km off the west coast of Vancouver Island and 300 km from downtown Vancouver – was estimated at C$75bn in economic losses and C$25bn in insured losses.
The study of Canada
This paper addresses the specific problem of ex post guaranty fund assessments following a catastrophic loss that results in the impairment and potential failure of multiple insurers. We consider the problem where the ex post assessment mechanism itself results in insolvency for insurers that survive the initial catastrophe. This issue is of particular importance in countries where the peak peril exposure is large relative to the size of the insurance industry, such as in Canada, Australia, and Chile, and where the private insurance market provides coverage for catastrophic losses, such as earthquakes.
Using Canada as a case study, we examine the impact on the property/casualty insurance market1 of large earthquakes in Vancouver, British Columbia, that result in estimated insured losses ranging from C$15 to 40bn. We show that for total losses below C$20bn, no insurers fail and at C$20bn, two insurers are distressed but are rescued by a capital infusion from other group members.
Thus overall, the private insurance market is able to respond without significant negative effects. For losses of C$25bn, several regional insurers would be distressed. One standalone insurer would fail, but we anticipate that insurers that are part of a larger group would receive a capital injection from the group, as this injection would not impact group solvency.
Twenty-six insurers would be distressed if insured losses were C$30bn. Most insurers that are part of a larger group would receive a capital injection and be saved. However, nine insurers would fail due to the earthquake and two more would fail because of the resultant impact on the group. Losses of this size would not create a contagion effect. However, they would create liquidity problems for the guaranty fund: assessments would not be received fast enough to pay claims arising from the catastrophe.
Once losses reach C$35bn, we estimate that 18 insurers (including members of still solvent groups) would be wound up as a result of the earthquake. Another seven insurers would fail due to failures of group members. In addition, the ex post assessment mechanism of the guaranty fund would cause the failure of several large insurers because of the size of the resulting assessment, the manner in which it is distributed across solvent insurers, and the accounting treatment of the assessment.
Once a large national insurer fails, the remainder of the industry cannot cover the assessment without failing, and the end result is the collapse of the entire Canadian property/casualty insurance market. The assessment mechanism itself creates contagion risk.
Our analysis relies on the Canadian market, but the results are also relevant to other jurisdictions where guaranty funds are financed through an ex post assessment mechanism and the jurisdiction faces a peak peril that is large relative to the size of the industry. In addition, the implications are relevant for guaranty funds financed through an ex ante system where the prefunding is too low relative to the potential payout arising from the peak risk, since most ex ante funds have mechanisms to assess surviving insurers to cover excess losses.
Given the upward trend in catastrophic losses in recent decades, this issue as to how to pay for large losses should be addressed before such a loss happens.
Public-private partnership combats contagion
Insurance guaranty funds play an important role in protecting policyholders from insurer insolvency. While ex-post funded guaranty fund systems can be designed to fund claims for failed insurers with relatively infrequent and idiosyncratic causes of failure, peak perils pose the challenge of placing very large claims demands on the guaranty funds. With the likelihood of mega-catastrophes increasing, due to such factors as climate change, increasing urbanisation of the population and wealth accumulation, it is important for countries to analyse the impact of extremely large losses on the functioning of guaranty funds.
We examine this problem using Canada as a case study. Canada is the only G7 nation that does not have a public–private partnership to pay for large catastrophic losses. To date, PACICC has not been stress-tested with an event large enough to cause multiple insurer insolvencies. However, given the not unrealistic scenario of an earthquake that causes C$35bn in insured damages, consideration of the impact of such event on the insurance market and the guaranty fund is prudent.
Our analysis indicates that even with a degree of regulatory forbearance, a loss of this magnitude would cause the failure of 18 insurance companies. The resulting assessment arising from these failures is then allocated across solvent insurers by market share within the relevant jurisdictions. The loss is so large that all amounts are not collectable in the short term and insurers retain most of the assessment as a liability.
The result is contagion: the assessment causes seven more insurers to fail the federal risk-based solvency test. One of these insurers is the largest insurance company in Canada, and once it fails, the remaining insurers in the country do not have the capital resources to pay the assessment.
Not paying claims arising from a catastrophic loss has welfare implications, including reduced and longer path to recovery in the disaster area. As such, we argue that jurisdictions need to plan how to fund losses arising from a peak peril before the event occurs. For jurisdictions with a mature insurance system that is of a size comparable to the peak peril, such as the US, a strictly private insurance mechanism may be sufficient.
For other jurisdictions with less capacity relative to peak perils, increasing prudential requirements and alterations of the guaranty fund itself are not sufficient to protect the solvency and sustainability of the insurance market. A public–private partnership addresses the problem of financing mega-catastrophes without resulting contagion, while at the same time preserving the expertise and efficiency of the private market. A
Ms Mary Kelly hails from the School of Business & Economics, Wilfrid Laurier University; Ms Anne Kleffner hails from Haskayne School of Business, University of Calgary; and Mr Grant Kelly is vice president and chief economist, Property and Casualty Insurance Compensation Corporation.