Protection gap can be closed

Closing the protection gap is not impossible, AIR Worldwide's Ashish Jain told APIC17.

The ideal environment would be that "everything is insured".

AIR Worldwide's Singapore vice-president and MD said while the increased cost of natural catastrophes was widely documented, if losses were normalised after considering population increases, greater wealth and property values, and inflation, there was "no economic loss trend in the past two decades".

Over the last decade, nat cat losses were below the expected global average of $US80 billion a year. Uninsured economic losses were $US300 billion a year.

Much of north America had no flood or earthquake cover and, in Asia, the insurance penetration gap was even higher, with only 5% to 8% of losses insured. The US Government was paying more for nat cats than in the past. For example, after Hurricane Diane in 1955, the US Government's contribution was less than 10% of the cost. After Hurricane Sandy in 2015 it was 70%.

Mr Jain said the protection gap was caused by:

• Poor risk awareness – people not being sufficiently educated about potential losses
• Government bailouts – when governments are lenders of last resort, people buy less insurance
• Capital allocation inefficiencies – many agencies try to help but there’s insufficient co-ordination
• Affordability – insurers need to make profits and cover is sometimes too costly
• Immature insurance regulatory frameworks generate a loss of confidence in the market; particularly in developing markets there is fear about insurers’ claims-paying ability
• Cultural issues – insurance is not viewed as a need.

Mr Jain said the protection gap affected the middle class and "aspirers" more than those at the top or bottom of the income scale because they had assets. They got nothing from governments because they were not considered "deprived".

General insurance was growing in ASEAN markets, but still covered only a fraction of the population. For example, in the Philippines, annual insurable losses were $US70 billion, but only $6 billion of that was insured. In Indonesia, for earthquakes only, $40 billion was insurable, but only $3 billion was insured. In Vietnam, $12 billion was insurable, but only $1.3 billion was insured.

In many ASEAN nations, "generally only large commercial facilities are insured, so there is a large burden on governments when an event occurs".

Mr Jain said public-private partnerships were needed to fund uninsurable risks. AIR was working with the World Bank, the Rockefeller Institute's 100 Resilient Cities campaign and a host of other organisations to develop new products. For example:

• A catastrophe risk program in The Philippines
• Pacific Island risk pooling
• A pandemic emergency financing facility involving AIR, Swiss Re, Munich Re and the World Bank.

In a separate presentation, Mr Jain said catastrophe modelling was more than just a reinsurance buying tool.

Cat modelling began to become popular after 1992's Hurricane Andrew in the US, which cost $US26.5 billion.

Cat modelling programs mathematically represented the characteristics of perils. Traditional methods were "not good predictors" of possible losses because "the constantly changing landscape of exposure data limits the usefulness of past loss experience".

Cat modelling could provide the probability of a given loss level; where events were likely to occur; how intense they would be; frequency of events; and the estimated range of damage and insured losses.

Cat modelling considered historical data, physical information about locations, construction methods, population density, and policy conditions, limits and deductibles to calculate insured losses.

It provided a range of outputs with one of the most important being the exceedance probability curve – what's the probability of a portfolio having X amount of losses.

Cat modelling was used at all stages on the insurance value chain — insurers, reinsurers, reinsurance brokers, cedants and capital markets.

Mr Jain said companies with superior analytics and modelling would force adverse selection or risks to competitors that did not.

He said AM Best had identified cat risks as the primary threat to insurer solvency, so insurers and reinsurers needed to assess risk aggregation and accumulation, the potential impact on surplus layers, and the impact on ratings.

Cat modelling was originally designed to help insurers decide how much reinsurance to buy and was still used as merely a reinsurance buying tool in some developing markets. However, it could now be used to buy optimal reinsurance protection, re-evaluate treaties, discover pricing abnormalities, and save on reinsurance costs. It should be "embedded into the underwriting process" for better decision making and to navigate volatility in financial markets.

Cat modelling could assist in portfolio optimisation – deciding where to write business by examining risk characteristics like construction methods, occupancy and year-built bands.

In claims, it could be useful for advance planning, resource deployment and post-event communications.

After an event, insurers need to know:

• the total loss
• estimated claim numbers by zip/postcode
• the impact on reinsurance programs and cat bonds
• the impact on the balance sheet and income
• whether illiquid assets would need to be sold to pay claims.

New areas in modelling included:

• Cat cyber events
• Cat lability events, which Mr Jain said had the potential to be bigger than property catastrophes
• Supply chain risk quantification
• Life and health
• Crops and agriculture
• Climate change.