When will ocean carriers shed their obsession with vessel size? Not anytime soon, according to one economist.

For decades now, carriers have been introducing progressively bigger containerships into the oceangoing trades. The largest vessel in 1970 was 1,800 twenty-foot equivalent units (TEU); by 2014 it had surpassed 18,000 TEU. Currently the record is held by the OOCL Hong Kong, with capacity of 21,413 TEU — the largest containership ever built.

Driving this obsession with gigantism is the desire to take advantage of scale economies: the bigger the ship, the lower the per-slot cost of operation. But there’s a downside to these gargantuan structures as well. The number of ports that can accommodate them, with sufficient water depth and berth length, is severely limited. And that translates into fewer service options for shippers.

One might think the size race would eventually yield diminishing returns, especially given persistent overcapacity in major trades. But carriers aren’t backing off yet, according to Walter Kemmsies, economist and chief strategist with Jones Lang LaSalle Inc. (JLL).

The addition of a third lock to the Panama Canal, allowing for the passage of wider vessels, is one reason. Ships of up to 15,000 TEU can now transit the canal, Kemmsies notes, so carriers moving imports from Asia can sail all the way to the U.S. Gulf and East Coast, instead of turning around at the West Coast. Kemmsies believes the canal maximum could be stretched to 18,000 TEU with additional widening.

The next class of containership will range between 18,000 and 25,000 TEU, he predicts. Already the orderbooks contain more vessels topping 20,000 TEU, and there’s no structural barrier to their getting even bigger. The only real obstacle is infrastructure limitations at ports (and, of course, market demand).

Global containership fleets also favor smaller ships in feeder service and lower-volume routes. They range between 1,000 and 5,000 TEUs. Ships of up to 4,500 TEU can use the Panama Canal without paying higher fees for the third set of locks, Kemmsies notes.

What needs to change, he says, is where American importers source their goods. Currently shipments from China account for around 35% of containerized tonnage entering the U.S. India, the next largest source, lags far behind with 4%, followed by Germany with 3.8%. Taken as a whole, the European Union is responsible for 18% of import volume by tonnage — just over half of China’s share. Given uncertainties caused by the U.S.-China trade war, along with rising costs of production in China, a diversification of sourcing is necessary. When it comes to offshore manufacturing, India and Vietnam are among the Asia nations that are poised to gain market share at the expense of China, Kemmsies says.

With congestion a serious problem at ports in Southern California and New York/New Jersey, shippers are looking to maximize their use of container space. That means greater participation by non-vessel operating common carriers and consolidators. But in times of tight capacity and potential delays, smaller shippers remain at a disadvantage to their larger counterparts.

“Anybody who doesn’t ship a substantial number of boxes is getting priced out of huge congested markets,” Kemmsies says. “If you don’t have to be there, you should get out.”

One answer is to route cargo through less-crowded ports. Busy as they are, Oakland, California; Seattle and Tacoma, Washington; and Prince Rupert, British Columbia offer alternatives to the dominant facilities at Los Angeles and Long Beach. For shippers operating on thin margins, such as agricultural exporters, a few pennies per ton of freight cost can make the difference between profit and loss.

Inland congestion continues to be a problem as well, especially where containers are shifted between modes of transportation. Ports such as Savannah, Georgia and Charleston, South Carolina are spending heavily on building out their intermodal transfer facilities, which must be able to accommodate both international and domestic containers and trailers, moving by rail and truck.

The nationwide driver shortage continues to act as a barrier to the smooth flow of freight over the roads. Kemmsies says the driver shortfall could increase from its current level of 50,000 to as many as 160,000 over the next 10 years. To skirt the problem, shippers will increasingly turn to rail for long-haul moves. At the same time, the economics of that decision will shift. The distance at which a shipper would consider rail over truck used to be around 700 miles; now it’s closer to 500.

With on-dock facilities facing heavy congestion, inland ports become another important piece of the transportation puzzle. Kemmsies cites projects either underway or on the drawing board in Portland, Oregon; Albany, New York; Charleston and Savannah, and Southern California’s Inland Empire.

“We just have to put more cargo on trains,” he says. “It has to become all about intermodal.”