By Mathieu Boudreault, FCIA, PhD
Following the Calgary floods in 2013, the federal government called for the Canadian insurance industry to play a more active role in the flood insurance market. Many discussions have occurred between the industry and the government since then to improve the financial management of flooding in Canada. This has led to the creation in 2021 of the Task Force on Flood Insurance and Relocation (TF) that published a report in August 2022 in which I have contributed as an actuary and professor of actuarial science.
In this article, I summarize some of the key findings of a research project* that I am leading with Public Safety Canada and the Insurance Bureau of Canada, meant to further investigate the financial management of flooding in Canada in the face of climate change. This project follows from the TF’s work and represents a deep dive into how actuarial science can help design better public policies for one of the most significant climate-related risks in the country.
Financial management of flooding in Canada
Financial management of flooding in Canada is a shared responsibility between municipalities, provinces and territories, the federal government and the insurance industry. Flood insurance is available as an optional rider in homeowner’s insurance and since 2015-2016, it typically covers sewer backup and overland flooding. The product has an approximate take-up rate of 60% over Canada.
However, for high-risk homeowners, insurance is difficult to purchase: it might be too expensive or coverage is inadequate. Therefore, provinces (and territories) provide emergency financial assistance to flood victims whereas the federal government financially supports the provinces for significant events through the Disaster Financial Assistance Arrangements (DFAA).
Municipalities are responsible for land use planning and maintenance of infrastructure with the oversight of provinces that also supervise the mapping of floodplains. In addition to managing the DFAA, the federal government supports provinces and municipalities through various programs, including the National Disaster Mitigation Program and the National Adaptation Strategy.
Flood risk in Canada
A first research objective was to establish an independent view of flood risk in Canada. We have therefore built a Canadian exposure dataset using various products from DMTI Spatial, Statistics Canada and Opta. This work consisted of assigning property characteristics and exact location (geocoding) to 10 million dwellings in Canada. With this exposure dataset, KatRisk then ran their fluvial and pluvial flood model across the 10 provinces leading to the following portrait of flood risk in Canada.
“The total average annual losses (AAL) in Canada amounts to $1.4 billion, whereas there is a 1% chance every year that total flood losses are beyond $13 billion and a 0.1% chance that such losses are above $34 billion.”– Mathieu Boudreault
Although not as significant as earthquakes, flooding still yields a sizeable systemic risk. This is however a more optimistic view than what the TF showed in their report, with an AAL of nearly $3 billion per year. This is because the TF used an average of different vendor flood models that showed a large discrepancy and focused on an extra 5 million homeowners who reside in multi-dwelling units. This illustrates how difficult getting a clear portrait of flood risk is in Canada (and elsewhere).
Flood risk is also heavily concentrated in a small share of homeowners: 39% of the AAL is concentrated in the top 1% of homeowners, 78% in the top 10% of homeowners and this is mostly aligned with the results of the TF. It is therefore no surprise to find that flood insurance is not available for high-risk homeowners. As such, and even if we had an 80% take-up in the country, flood insurance would only cover about 20% of the AAL. This is a clear motivation for the federal government to discuss potential solutions for the risk-sharing of floods in Canada, including a national flood insurance program for high-risk homeowners.
Annual costs and initial capitalization
A second objective of the research project is to evaluate the recurrent costs and the initial capital required to set up a flood insurance program. This largely depends upon coverage provided, who the program is targeting and the expected participation, use of international reinsurance and support from the federal government as a backstop, etc.
Knowing that nearly 80% of losses is concentrated in 10% of homeowners, we investigated plans with full and partial participation. We analyzed various policy terms, with or without a coverage limit of $300,000, and deductibles ranging from $5,000 to $25,000. The costing analysis assumed the plan would be run by a not-for-profit organization with the federal government acting as a financial guarantor to a fully funded plan. Including additional living expenses, loss adjustment expenses, cost of international reinsurance and overhead, it would cost between $1.7 billion to $2.4 billion to run an insurance program under full participation. Because of the heavily skewed distribution of AAL across homeowners in the country, the cost of the plan depends largely upon the risk profile of homeowners in the plan and not only the take-up rate. For example, excluding the top 1% of homeowners yield annual savings of more than 40%.
We investigated the capitalization of various plans as well as the frequency and severity of federal interventions with different solvency levels, policy terms, risk profiles, investment policies, reinsurance, etc. Having the federal government back the program in case of catastrophic losses is important because the aggregate loss distribution in Canada has very heavy tails as discussed earlier. This however comes with political risk as well for both the program and homeowners. Therefore, we also analyzed the case where such interventions would be minimized.
For a 1% probability of insolvency over 30 years, the required initial capital is in the range of $30-50 billion depending on policy terms. Again, excluding the worst 1% of homeowners from the plan, the required capital drop by more than 40%. When federal interventions are allowed, the frequency and severity of the cash flows required obviously depend on the amount of initial capital. With minimal capital, the interventions are frequent, early and often total tens of billions of dollars over 30 years. It shows there is no free lunch: high initial capital provides safety against large and frequent interventions, therefore minimizing political risk for the plan. Underwriting tighter reinsurance from the international market until more capital builds up also helps considerably reduce the required capital, but with a significant additional annual cost.
There are obvious difficult questions that arose from the actuarial analysis. For example, who is going to pay the recurrent and non-recurrent costs of a flood insurance program in Canada? What to do with the top 1% of homeowners who are driving a very significant share of losses? The TF has taken the first steps towards answering these questions by comparing four risk-sharing models: two high-risk insurance pools, a public insurer model and a public reinsurer model. As a result of the TF’s work, the federal government announced in its budget in March 2023 that it would create a national flood insurance program, an insurance subsidy program and a new reinsurance Crown corporation to support the financial management of flooding. Parametrization of these programs is still being discussed with key stakeholders and a bill is needed to initiate a Crown corporation.
The future creation of a Crown (re)insurance corporation is good news for Canada as it provides a flexible and powerful tool to foster pooling of risks from natural hazards (not just flooding, perhaps, earthquakes as well) while making sure it remains autonomous, apolitical and viable in the long run. There are however challenges ahead for Canada to become resilient to flooding.
The funding model of the DFAA needs to be reviewed. As of today, Canadian taxpayers are funding the DFAA from the general budget (pay-as-you-go model) and that does not provide the financial incentives for the provinces to tighten flood mapping and land use planning regulations.
“If and when the DFAA is phased out, the funding scheme of the new flood insurance and affordability programs should involve provinces and possibly municipalities to incentivize risk reduction in the long term, or in other words, to slow down construction in or near the floodplains and build protective works.”– Mathieu Boudreault
Canada also lacks key data to appropriately prepare against flooding in the long term. On the flood modelling front, each municipality and province prepares flood maps for regulatory purposes with different methodologies and levels of risk (return periods). As a result, the insurance industry typically contracts and even the TF contracted foreign vendors to model flood risk all across Canada with a consistent methodology. Our preparedness to flooding is therefore in part in the hands of foreign entities. This is in addition to urban areas in the country that do not have high-resolution terrain modelling yet, which is essential for a more precise view of flood risk.
On the vulnerability front, claims data history for overland flooding is limited in the industry, whereas provinces have provided financial assistance for years. As such, the information that would be important to better understand the financial vulnerability of Canadian homes to flooding is locked in with provinces and therefore, difficult to access to analyze, e.g., the profitability of relocation or other risk reduction measures, or for insurers, to appropriately price flood insurance.
And finally, the elephant in the room: climate change. An Environment and Climate Change Canada report states, “While inland flooding results from multiple factors, more intense rainfalls will increase urban flood risks. It is uncertain how warmer temperatures and smaller snowpacks will combine to affect the frequency and magnitude of snowmelt-related flooding.” And therefore, what will be the financial impacts of climate change on flooding in Canada, and how is it going to affect the financial management of flooding for the future? This is the topic of the second part of our research project* with Public Safety Canada and the Insurance Bureau of Canada, whose results will be available in 2024-2025.
*The research project, Canadian flood-risk sharing models in the face of climate change, is funded by Public Safety Canada, the Insurance Bureau of Canada through the Natural Sciences and Engineering Research Council of Canada Alliance program, with Michael Bourdeau-Brien (Université Laval) and Jason Thistlethwaite (University of Waterloo).
Mathieu Boudreault, FCIA, PhD, is a professor in the Department of Mathematics (actuarial science) at the Université du Québec à Montréal.
This article reflects the opinion of the author and does not represent an official statement of the CIA.