Amtrak Transfers $275 Million Risk in First Cat Bond for Railway

  • Coverage protects against infrastructure damage in Northeast
  • GC Securities and Goldman Sachs are bookrunners on deal

Amtrak obtained $275 million of natural-disaster protection from fixed-income investors in the first catastrophe bond for the U.S. long-distance passenger railway.

Marsh & McLennan Cos.’ GC Securities joined Goldman Sachs Group Inc. as bookrunners and initial purchasers of the debt, Amtrak said Wednesday in a statement. The bonds will help protect Amtrak if a catastrophe damages infrastructure in the corridor from Boston to Washington.

Insurers and reinsurers typically sell catastrophe bonds to investors to help cover extreme hazards, and buyers of the securities get above-market interest rates for taking the risk that they could lose principal in the event of a qualifying disaster. Amtrak’s coverage would protect against storms similar to Sandy in 2012, when sea water flooded New York tunnels, causing more than $1 billion in damage, according to the statement.

“This is the first time Amtrak has used the capital markets to broaden our base of insurance coverage,” Chief Financial Officer Gerald Sokol Jr. said in the statement. “The catastrophe bond market provides us with a means to diversify our sources of insurance in a cost-effective manner.”

The coverage, issued through a Bermuda-based special-purpose entity known as PennUnion Re Ltd., will protect against extreme storm surge, wind and earthquakes. The bonds yield 4.5 percentage points more than the benchmark and mature in late 2018, according to data compiled by Bloomberg. Advisers on the deal included RMS Inc., a catastrophe-risk modeling firm, and Mayer Brown LLP.

Support from more than 25 investors “allowed the transaction size to be increased by 37.5 percent from the initial offering size while at the same time pricing at the lowest end of price guidance,” Chi Hum, global head of insurance-linked securities distribution at GC Securities, said in a separate statement.