1 June 2018

2018 Atlantic hurricane season - the calm after the storm?

Last year's Atlantic hurricane season was one of the most catastrophic in recent memory with insured losses estimated at over $130bn. Now 12 months later and on the first official day of the season, forecasters believe relative calmness awaits. But forecasters are often wrong. Paul Walsh reports.

The 2017 season saw 17 storms including 10 hurricanes and six major hurricanes in the north Atlantic and led to a total of 882 fatalities. It was the only one of the last six seasons to feature multiple category 5 hurricanes, including Harvey, Irma and Maria.

As the 2018 season commences, forecasters are predicting something of a polar opposite.

Colorado State University

Colorado State University (CSU) predicts a "near-average" season citing abnormal cooling in the tropical Atlantic as the key driver for this. Equally CSU predicts 14 named storms, including the most recent, Alberto, of which six will become hurricanes and two will reach major hurricane strength.

According to CSU, hurricane activity in the upcoming season will be in line with the average. In comparison, the extremely active 2017 season was around 245% of the average. In regards to landfall, CSU predicts a probability of 51% of major hurricanes making landfall across the entire US coastline.

Tropical Storm Risk

Tropical Storm Risk (TSR) is also predicting a "well-below average" season, anticipating less than 50% or more below the long-term expectation. In a similar vein to CSU, TSR believes the lower forecast is due to considerable cooling of sea surface temperatures in the tropical north Atlantic and Caribbean Sea. Furthermore, TSR believes there is a 69% likelihood that hurricane activity in 2018 will be in the lowest one-third of years since 1950.

Alternatively, TSR predicts there is only a 7% chance hurricane activity will be above the firm's 68-year average and only a 19% chance landfalling in the USA will be above average on its index.

National Oceanic and Atmospheric Administration

Arguably the worst case scenario comes from the National Oceanic and Atmospheric Administration (NOAA) which predicts a 75% chance the season will be "near or above normal."

Specifically, NOAA predicts a 35% chance of an "above-normal" season and a 40% chance of a "near-normal" season. NOAA states an average season would produce 12 named storms, of which six could become hurricanes including three major hurricanes. In its prediction for the upcoming season, NOAA cited a 70% likelihood of 10-16 named storms of which five to nine could become hurricanes and four could become major hurricanes.

No complacency

In spite of the seemingly shared predictions of a calmer season, industry participants said forecasts should not be relied on blindly.

"The predictions are interesting but I don't think they are that important for insurers/reinsurers for a couple of reasons," said Karen Clark, president and chief executive at Boston- based catastrophe modelling firm Karen Clark and Company (KCC).

"One is we know from their track record there is a lot of uncertainty around the predictions and they're as likely to be wrong as right, and we have a whole spectrum now anyway from below average to average to above average.

"Secondly, an insurance company is not going to buy less reinsurance because the prediction is for a below average season. I don't think any insurance company would base a major decision on these forecasts—not only because there is a lot of uncertainty but because of the nature of expected hurricane losses."

Moreover, despite the forecasts of an average storm season, the potential loss impact of the storms could remain severe given factors like storm surge hitting homes along the Atlantic and Gulf coastlines across 19 states.

"While industry predictions for this year's storm season indicate average activity levels, associated storm surge risk remains an important consideration for residential and commercial properties in the 19 states analysed," said Dr. Tom Jeffery, senior hazard scientist at CoreLogic.

"Depending on the location of a storm's landfall and that area's population density and reconstruction costs, lower category storms can cause just as much damage as storms in higher categories."

Lessons learned?

The magnitude of losses of last season had a number of knock-on impacts. Arguably most notable, was the vast difference in loss estimates produced by catastrophe modelling firms.

Following the impact of Maria on Puerto Rico, AIR Worldwide put estimated losses in the Caribbean island at between $40bn and $85bn in September. Rival firm RMS initially estimated insured losses from hurricane Maria would be between $15bn and $30bn.

The following December the estimate on maximum insured losses from the storm was halved. And, adding to the uncertainty, AIR Worldwide said the top end of insured losses in the region would be $48bn - but they could also be as low as $27bn.

"The biggest surprise last year to the industry was the size of the differences between loss estimates," said KCC's Clark.

"The differences between estimates, particularly for Maria, where some of the modellers gave large ranges and the ranges didn't even overlap.

"For the hurricane models that have been in the industry for almost 30 years, I believe most companies thought they should be timelier in the estimates and more accurate."

As a result of such differences in modelling, Clark adds the industry needs to pay attention to "not just the science in the model" but the quality of "the estimated losses arising from that science."

Despite the gloomy forecasts, a report from Fitch Ratings revealed US insurers are "well-prepared to" withstand any future catastrophic events in 2018. According to Fitch, aggregate industry policyholders' surplus grew by 7.5% in 2017 to a record $765 billion, as elevated underwriting losses were offset by higher total investment returns.

Clearly forecasts are just predictions and the true extent of the season will not be known until it is well-underway.