The article does a poor job of describing the Coase Theorem. This New Yorker article [1] does a much better job. I like the example they give, which comes from Coase's paper:
"To illustrate this, Coase used the example of a cattle
farmer’s herd straying onto the cultivated fields of a
neighboring grain farmer.
If the law said the cattle farmer was legally responsible,
then mostly likely he would end up paying for the cost of a new fence.
If putting up a fence was prohibitively expensive, he might end up paying the
grain farmer for the occasional damage to his crops. But what if the cattle
farmer wasn’t legally responsible, perhaps because the property lines were
unclear? In these circumstances, Coase said, the grain farmer would likely end
up paying for a fence. Or, if the grain farmer believed that constructing a
fence would be more costly than the occasional damage done by straying
cattle, he would reluctantly accept the odd incursion.
The assignment of property rights would determine who pays. But from a
broader economic point of view, Coase argued, the outcome would
essentially be the same, regardless."
This doesn't hold up in real life with negative externalities and transaction costs. The article cites an example of pollution, but pollution can affect thousands of people, and there's no way to get all of them together to reach an efficient outcome without transaction costs.
[1] http://www.newyorker.com/online/blogs/johncassidy/2013/09/ro...