Ruchir Sharma had an NYT column warning about the risks of a trade war from the tariffs Trump is imposing on steel and aluminum imports. At one point the piece tells readers about rising protectionism across the world and says that as a result, "trade has yet to recover to its pre-crisis level."
The measure of trade the article gives is merchandise trade as a percent of world GDP. This measure is misleading since a major factor reducing this ratio is a fall in oil and other commodity prices. Before the crisis oil prices rose sharply, peaking at $150 a barrel in 2008. Other commodity prices also were very high in the years just before the recession.
The reduction in prices for commodities is a major factor in reducing the ratio of trade to GDP. In the case of oil, with more than 40 million barrels trade daily, a drop of $50 a barrel in the price would reduce the volume of world trade by more than $750 billion a year, or roughly one percent of world GDP. There is a similar story with other commodities.
It is also worth noting that weaker and shorter patent and copyright protections would also lead to a lower ratio of trade to GDP. If drugs are traded across borders at generic prices rather than patent-protected prices, this ratio will fall even though the volume of trade can be rising.
Royalties and licensing fees are not picked up in this measure since it is explicitly merchandise trade, which would exclude payments for services. This is another factor that would tend to depress the ratio. As the world economy shifts from goods production to services, it is pretty much inevitable that the ratio of merchandise trade to GDP drops over time.
This doesn't mean that Sharma is necessarily wrong about a rise in trade barriers over the last decade (certainly patent and copyright protections are getting stronger), only that the measure he uses to back up this assertion is not very useful.