Current global trade is valued at about US$14.8 trillion, of which the US accounts for 14 per cent. Picture / Getty Images
"Trade wars are easy to win," according to Donald Trump. But history teaches another lesson: trade wars have no winners. When one government starts slapping tariffs on imports to protect its own, others feel obliged to retaliate.
The result is a downward spiral in trade, one that can happen with astonishing speed. Figures collated by Llewellyn Consulting show just how fast.
In January 1929, global trade was valued at about US$2.9 billion ($4b). Four years later, it had shrunk by two thirds to just US$992 million. Current global trade is valued at about US$14.8 trillion and the US accounts for 14 per cent of that. It has a lot to lose.
This doesn't mean much to Trump. For him, trade is a game of winners and losers. Exports are wins, imports are losses.
This isn't because, like the UK, he's worried about the sustainability of his country's current account deficit. It's because he likes the idea of American cars, American steel, American coal, and he certainly does not like seeing "made in China" on everything, even if that does include his campaign's signature "Make America Great Again" red baseball caps.
Underlying this instinctive nationalism is a serious problem: a large loss of American working class jobs. From 2006 to 2016, the US lost 1.8 million manufacturing jobs. It gained 10 million services jobs, but the two aren't always directly fungible.
Services jobs require different skill-sets and might be in different locations or require more flexibility than a traditional manufacturing job.
Most of the jobs are thought to have been lost to automation rather than outsourcing, but for Trump and his voters, the culprit is clear: international trade. His proposed solution, masterminded by his trade adviser Peter Navarro, is simple.
Last week, he announced tariffs of 25 per cent on imports of steel and 10 per cent on aluminium. This is probably the start, not the end, of a broad protectionist programme, which aims to cut the US off from the global economy so that it is forced to "buy American".
Steel tariffs might sound like good news for working class voters. More than 80,000 people are employed by the US steel industry.
That sounds like an awful lot - until you read another statistic: American carmakers employ around 700,000 people.
That same automotive industry is one of the steel industry's biggest customers. It buys more than a quarter of all steel used in the US. The effect of making steel more expensive, therefore, is to raise costs for US industry.
When it comes to steel, Trump is at least correct in one thing: there is far too much of the metal being produced.
The glut has been led by China, whose steel production has increased from 100 million tonnes 20 years ago - about the same amount as the US - to 800 million tonnes today. To be fair, this enormous build-up in capacity was driven by its demand. Chinese steel consumption rose from 100 million tonnes to 750 millions.
Now, however, as China's mega building boom fades, demand is flatlining. According to research by Bank of America/Merrill Lynch (BAML), per capita steel demand in China has stayed around the same level for the last five years. What should happen, of course, is that market forces kick in. Thousands of unnecessary steel mills should be shutting down.
Of course, no country likes to see large industries shut down and lay off workers, and China's authoritarian political leaders are particularly nervous about letting it happen.
And so they have increased the flood of subsidies flowing towards steel producers, dumping their loss-making, cheap steel on global markets.
For years now, Europe and the US have been responding with heavy tariffs, to the point where the vast majority of exported Chinese steel is now sold in Asia. Chinese steel already makes up only a very tiny fraction of US imports. Meanwhile, Beijing has very slowly started to address its overcapacity problem. BAML's figures show that Chinese production is now creeping downwards.
Trump's tariffs won't directly hit China, of course, but will have a bigger impact on the US's biggest foreign steel supplier: Germany.
It's not entirely clear what Trump's trade gripe against the EU is, although he has said several times that Berlin benefits from an artificially low currency due to the dysfunctional euro.
Given that this negatively affects Europe's southern economies as much or more than it boosts Germany, this in itself does not quite explain his reasoning for threatening to slap tariffs on all EU-made cars.
The protectionist fallacy is that tariffs can reverse global trends.
Former car workers in Detroit, who helped get Trump into the White House, believe that cutting down on trade with the outside world will revive their factories and warehouses.
But a trade war won't bring those jobs back, even if American workers were also prepared to pare down their living standards to compete with their Chinese counterparts. The problem is that many jobs have simply been replaced by innovative machines. Once productivity has risen in this way, there's no putting the genie back in the bottle.
A better policy would be to focus the US' considerable resources on educating and training its workers for the millions of new jobs available in the modern economy. That, of course, isn't a quick or easy fix and therefore it's not likely to get Trump re-elected.
Meanwhile, US trade policy could prepare for the future better by striking deals with those countries that do adhere to similar standards on technology ownership and safety, expanding the sphere of global trade that works according to Western norms.
Instead, Trump's retrenchment will make his own country poorer, cede ground to China's totalitarian model and risk setting off a catastrophic contraction in global trade. Trade wars are easy to start, but they're certainly not easy to win.