Sharon Oosthoek | March 6, 2017
Imagine for a moment you are a business owner competing with a firm in Ohio to sell your product to a customer in Michigan. You may be able to match the competition’s quality, productivity, service, and price. But you have to do better than that because you have the border to deal with and the person in Ohio doesn’t.
Enter Bill Anderson. The director of the University of Windsor’s Cross-Border Institute wants to level your playing field by answering a deceptively simple question: “How can we make the border work better?” With a newly installed U.S. president skeptical of the benefits of international trade, the question is more timely than ever.
Currently, most goods pass between Canada and the U.S. tariff-free. But they are still subject to inspection and paperwork, and everyone who helps move those goods must pass immigration and security screening. All of this takes time and none of it is cheap.
As the Ontario Research Chair in Cross-Border Transportation Policy, funded by an endowment from the government of Ontario, Anderson has delved deeply into this topic. His research shows trucking goods across the border—the most common form of transport—boosts costs by 15 to 40 per cent, compared to moving the same goods the same distance domestically.
What truck traffic tells us
Reducing that expense and the hassles of border crossings could have significant payoffs. Almost 80 per cent of Ontario’s exports go to the U.S. Canada is in turn the largest destination for U.S. exports, with Ontario taking more than half of them.
Much of this trade crosses bridges at the international border between Ontario and the U.S., with the Windsor-Detroit Ambassador Bridge handling the lion’s share. It accounts for more than 2.3 million truck crossings a year.
Even when goods travel from Asia or Mexico via boat or rail, the last leg of the journey is almost always on a truck. Which is why Anderson and his team are paying particular attention to truck traffic between Ontario and the U.S.
Understanding how trucks move across the border and predicting their future movements can give us a good handle on how our economy is performing. It also gives us insight into when and how to intervene to make sure trade flows smoothly.
Anderson has come up with a novel way of using GPS data to keep track of all this traffic. While freight carriers have been using GPS to track their trucks since the late 1990s, they don’t collect data with an eye toward understanding what happens when trucks stop. Is it making a delivery, or is the driver having dinner or catching some shut-eye?
It’s important to understand the difference. “Once you filter out stops for food and fuel, you have your origin and destination points,” says Anderson. “Then you can figure out what industries are involved—what’s going from where to where.”
Not only does this help build a picture of Canada-U.S. trade, knowing where the demand is for specific products has commercial value.
Researchers traditionally differentiate between a pit stop and a delivery by interviewing truckers. But this approach takes time and money, and it relies on people’s imperfect ability to recall events. Anderson and his team have come up with a more accurate and cost-effective system: They calculate an entropy number for each stop.
Entropy is a measure of disorder, so if GPS data show a given stop attracts a large number and diversity of trucks, it gets a high entropy number. A lot of trucks from various carriers in one place suggests a pit stop for food, fuel, or accommodation. In contrast, one or two trucks stopped at a particular location yields a low entropy number, indicating a deliver stop.
Just to make sure, Anderson and his team cross-checked locations with high entropy numbers with Google maps. As expected, they found that of the 150 locations with the highest entropy numbers, 148 corresponded to truck stops, gas stations, or motels.
Heads up on border slowdowns
He and his team are also working on making better predictions for border crossing times, even several hours in advance of reaching a bridge.
“It’s possible to find the time for crossing right now,” says Anderson. “But we’d like to make a prediction that tells a driver in London, Ontario, which is 190 kilometres from the border, what to expect when they get there.”
Armed with such information, the driver can then decide whether to take the Ambassador Bridge or divert to the Blue Water Bridge connecting Sarnia, Ontario to Port Huron, Michigan. Anderson’s team is in the midst of building an algorithm that can work with complex and variable information, including the number of trucks between the driver and the bridge, accidents, weather, computer system breakdowns, and the number of lanes open at the border.
These last two variables touch on another area of Anderson’s research: the resiliency of our border crossings and the need to build in redundancy, such as the new bridge downriver from the Ambassador, approved by the Canadian and U.S. governments.
As Anderson puts it: “If there is a flu epidemic in Detroit and three customs officers call in sick, that could back up traffic significantly.”
That in turn could wreak havoc with just-in-time supply chains, including automotive plants on one side of the border waiting for parts from the other side.
“The level of co-ordination is really interesting. The parts delivery has to happen during a precise window—not just Thursday, but between 10 and 11 a.m.,” he says.
And if Anderson has anything to do with it, those parts will make it to the assembly line right on time.