Blame foreigners first. This strategy is always the companion of aggrieved nationalism. It can be seen in Donald Trump’s ban on immigrants from seven countries. It will be seen in his protectionism. A kernel of truth – terrorism and the direct impact of imports on jobs – bolsters a lie: my actions are enough to keep you safe and restore the prosperity you once knew. At the heart of the US debate on trade policy is the story of jobs in manufacturing. The single most important fact is the steady decline in the share of jobs in manufacturing from about 30 per cent of total employment in the early 1950s to just over 8 per cent at the end of 2016. The main explanation for the long-term decline in the share of manufacturing employment in the US (and other high-income economies) has been the rise in employment elsewhere. In 1950, employment in manufacturing was 13m, while that in the rest of the economy was 30m. By the end of 2016, it was 12m and 133m, respectively. Thus, all the increase in employment between 1950 and the end of 2016 occurred outside manufacturing. Yet output of US manufacturing was not stagnant. Between 1950 and 2016, output rose 640 per cent, while employment fell 7 per cent. Even between 1990 and 2016 output rose 63 per cent, while employment fell 31 per cent. The explanation for the contrast between output and employment is rising productivity. Yet no one is proposing to stop this. Indeed, the problem is rather productivity’s recent stagnation: in manufacturing, output per hour rose only 1 per cent between the first quarters of 2012 and 2016. As a result, employment rose a little. Yet this is a bad outcome: the economy needs rising productivity if it is to generate sustained improvement in living standards. Between 1997 and 2005, the US trade deficit in manufactures increased by 2.6 per cent of gross domestic product. But today it is at much the same level as in 2005, after shrinking during the financial crisis. How much bigger might output of manufacturing be if this increase in the trade deficit had not occurred? Make the plausible assumption that the impact on value added is some two-thirds of the gross value of the goods. Then the value added in manufacturing might be some 1.7 per cent of GDP higher. Suppose that the effect on employment would be proportionate. Employment in manufacturing would be some 2.5m bigger than it now is. This might have prevented half of the job losses in manufacturing since 1997 and raised manufacturing’s share in employment to more than 10 per cent. In sum, the increase in the trade deficit in the early 2000s had a significantly negative effect on employment in manufacturing, but next to none on the long-term decline in the share of overall employment in manufacturing. Even if the trade balance had remained unchanged in the early 2000s, there would still have been a big reduction in the share of employment in manufacturing from the late 1990s. For that, the main reason was weak demand: not surprisingly, the absolute declines in employment in manufacturing occurred during the two recessions, in the early 2000s and again in 2007-09. What part of the rise in the trade deficit was due to the North American Free Trade Agreement and China’s accession to the World Trade Organisation? Bradford DeLong of Berkeley concludes that it is quite modest. A more sophisticated analysis by Daron Acemoglu of MIT and others concluded that trade with China directly caused the loss of about 10 per cent of the total number of jobs lost in manufacturing between 1999 and 2011. But analysis of linkages among firms and the impact upon local demand gives far larger negative effects of between 2m and 2.4m jobs, though this is still less than 2 per cent of total employment. Two big points emerge. The first is that the effect of import competition is often geographically concentrated. This is a particularly big challenge in a country as large as the US. The best response must be a combination of helping affected communities to generate new sources of employment and assisting workers (and not just those directly affected) to gain skills and so new jobs. A part of the strategy must also be to help restore lost US mobility. The second point is the need to sustain demand and so ensure that new jobs replace the old ones in the economy as a whole. The orthodox view is that the US can always achieve full employment by active use of fiscal and monetary policy tools. Experience since 2000 and especially since the financial crisis suggests this may be difficult. As I have argued elsewhere, huge current account surpluses in some countries forced deficit countries into financial excesses as an (ultimately unsustainable) way to maintain demand in line with potential output. The crisis vindicated the concern of John Maynard Keynes about the potentially malign role of surplus countries in the global economy. Alas, the policies proposed by Mr Trump and the congressional Republicans – a combination of piecemeal protectionism with a large fiscal stimulus as well as elimination of much of the social safety net – is likely to impose large costs on unprotected sectors, while leaving supporters even more desperate. Nothing he does will reinstate manufacturing to its lost role as the dominant provider of “good jobs”. Cheap imports and the ability to supply them have also brought big benefits to domestic consumers and foreign workers. The right approach would be proactive, not defensive: it would open global markets; it would force countries with huge surpluses to rely more on domestic and less on external demand; it would help workers and communities hit by adverse change, not abandon them; it would stop blaming foreigners for the “crime” of selling goods cheaply. Such policies would make excellent sense. Alas, they are not what we will see.