Tag Archives: Tianjin

No More Guessing Games for Marine Insurers

Huge ports mean huge amounts of cargo. Huge amounts of cargo mean huge accumulations of risk.

As a guiding principle about where marine insurers are exposed to the highest potential losses, it seems reasonable enough. But in fact, as RMS research has proven this week, this proposition may be a bit misleading. Surprisingly, a port’s size and its catastrophe loss potential are not strongly correlated.

Take the Port of Plaquemines, LA which is just south-east of New Orleans. It is neither well known nor big in comparison with others around the world. Yet it has the third highest risk in the world of insurance loss due to catastrophe: our analysis revealed its 500-year marine cargo loss from hurricane would be $1.5 billion.

Plaquemines is not an isolated case. There were other smaller ports in our ranking: Pascagoula, MS in the United States ranks 6 on our list with a potential $1 billion marine cargo loss due to storm surge and hurricane; Bremerhaven in Germany (ranked 4th at $1 billion) and Le Havre in France (ranked 10th at $0.7 billion).

Asia-Pacific ports featured less frequently, but worryingly one Asia port topped the list: Nagoya, Japan was number 1 ($2.3 billion potential losses) with Guangzhou, China a close second ($2 billion). Our analysis modeled risk posed by earthquake, wind, and storm surge perils in a 500-year return period across 150 ports – the top ten results are further down this blog.

Ports At Risk For Highest Lost
(500 year estimated catastrophe loss for earthquake, wind, and storm surge perils)

Estimated Marine Cargo Loss in Billions USD
1 Nagoya, Japan 2.3
2 Guangzhou, China 2.0
3 Plaquemines, LA, U.S. 1.5
4 Bremerhaven, Germany 1.0
5 New Orleans, LA, U.S. 1.0
6 Pascagoula, MS, U.S. 1.0
7 Beaumont, TX, U.S. 0.9
8 Baton Rouge, LA, U.S. 0.8
9 Houston, TX, U.S. 0.8
10 Le Havre, France 0.7

* Losses rounded to one decimal place.

Our analysis demonstrates what we at RMS have long suspected: outdated marine risk modeling tools and incomplete data obscure many high-risk locations, big and small. These ports are risky because of the natural perils they face and the cargos which transit through them, as well as the precise way those cargos are stored. But many in the marine sector don’t have these comprehensive insights. Instead they have to make do with a guessing game in determining catastrophe risk and port accumulations. And with the advanced analytics available in 2016 this is no longer good enough.

Big Port or Small – Risk Can Now Be Determined

Back to that seemingly simple proposition about the relationship between port size and insurance risk which I began this blog with. As the table above demonstrates, smaller ports can also present a huge risk.

But the bigger ships and bigger ports brought about by containerization have led, overall, to a bigger risk exposure for marine insurers. Not least because larger vessels have rendered many river ports inaccessible forcing shippers to rely on seaside ports, which are more vulnerable to hurricanes, typhoons, and storm surge.

The value of global catastrophe-exposed cargo is already huge and is likely to keep growing. But the right tools, which use the most precise data, can reveal where the risk of insurance loss is greatest. Leveraging these tools, (re)insurers can avoid dangerous cargo accumulations and underwrite with greater confidence.

Which means that, at last, the guessing game can stop.

In a box: Our ranking of high risk ports used the new RMS Marine Cargo Model™, with geospatial analysis of thousands of square kilometers of satellite imagery across ports in 43 countries. RMS’ exposure development team used a proprietary technique for allocating risk exposure across large, complex terminals to assess the ports’ exposure and highlight the risk of port aggregations. The model took into account:

  • Cargo type (e.g. autos, bulk grains, electronics, specie)
  • Precise storage location (e.g. coastal, estuarine, waterside or within dock complex)
  • Storage type (e.g. open air, warehouse, container — stacked or ground level)
  • Dwell time (which can vary due to port automation, labor relations and import/export ratios)

Tianjin Is A Wake-Up Call For The Marine Industry

“Unacceptable”  “Poor”  “Failed”

Such was the assessment of Ed Noonan, Chairman and CEO of Validus Holdings, on the state of marine cargo modeling, according to a recent report in Insurance Day.


China Stringer Network/Reuters

The pointed criticism came in the wake of the August 12, 2015 explosions at the Port of Tianjin, which caused an estimated $1.6 – $3.3 billion in cargo damages. It was the second time in three years that the cargo industry had been “surprised”—Superstorm Sandy being the other occasion, delivering a hefty $3 billion in marine loss. Noonan was unequivocal on the cargo market’s need to markedly increase its investment in understanding lines of risk in ports.

Noonan has a point. Catastrophe modeling has traditionally focused on stationary buildings, and marine cargo has been treated as somewhat of an afterthought. Accumulation management for cargo usually involves coding the exposure as warehouse contents, positioning it at a single coordinate (often the port centroid), and running it though a model designed to estimate damages to commercial and residential structures.

This approach is inaccurate for several reasons: first, ports are large and often fragmented. Tianjin, for example, consists of nine separate areas spanning more than 30 kilometers along the coast of Bohai Bay. Proper cargo modeling must correctly account for the geographic distribution of exposure. For storm surge models, whose output is highly sensitive to exposure positioning, this is particularly important.

Second, modeling cargo as “contents” fails to distinguish between vulnerable and resistive cargo. The same wind speed that destroys a cargo container full of electronics might barely make a dent in a concrete silo full of barley.

Finally, cargo tends to be more salvageable than general contents. Since cargo often consists of homogenous products that are carefully packaged for individual sale, more effort is undertaken to salvage it after being subjected to damaging forces.

The RMS Marine Cargo Model, scheduled for release in 2016, will address this modeling problem. The model will provide a cargo vulnerability scheme for 80 countries, cargo industry exposure databases (IEDs) for ten key global ports, and shape files outlining important points of exposure accumulation including free ports and auto storage lots.

The Tianjin port explosions killed 173 and injured almost 800. They left thousands homeless, burned 8,000 cars, and left a giant crater where dozens of prosperous businesses had previously been. The cargo industry should use the event as a catalyst to achieve a more robust understanding of its exposure, how it accumulates, and how vulnerable it might be to future losses.