On March 1, 2018, US President Donald Trump announced his decision to impose tariffs on imported steel and aluminum. The decision comes on the heels of a report from the Commerce Department released last month that concluded that such imports threaten national security. In response to the announcement, US stocks prices fell, in particular those of automobile and aeronautic manufacturers – all heavy users of metals. Trump defended the measure on Twitter, saying that “trade wars are good”.
While the longer-term impact of such protectionist measures are difficult to predict, much can be learned from similar cases. In a report drafted for the European Parliament, we review one such (well-known) protectionist episode, the global “safeguard measure” enacted by US President George W. Bush in 2002. Tariffs ranging from 8% to 30% were imposed on a wide range of steel products for a three-year period, starting on March 20 of that year.
Our research found that any claimed protection was dwarfed by unintended consequences: The measures destroyed employment in the US steel industry, fostered outsourcing overseas and hurt downstream industries. This case illustrates the consequences of using trade-defense instruments to pursue protectionist purposes. And while the official justification was allegedly unfair trade practices, the real motivation was clearly to secure the political support in steel-producing “swing states”.
What can be learned
This case is particularly informative about the potential consequences of Trump’s decision. To better understand this, let’s start with the impact on trade.
A detailed analysis of Bush’s steel tariff shows that the trade impact was very heterogeneous across products and countries, reflecting the numerous exclusions from the measure. US imports of steel products from countries on which the measure was applied plummeted, falling by 28% on average in 2002 and by a further 37% in 2003 (Bown, 2013, Table 5). Even products not covered by the measure were significantly affected in 2002. Meanwhile, imports of steel products from countries excluded from the safeguard increased strongly. For example, in 2002 they rose 40% for preferential agreement trading partners, and over 2002-2003 they climbed 28% from exempted developing countries. Furthermore, imports of non-safeguarded product categories increased. Overall, there was a 3% increase in US steel imports in the 12 months following the safeguard. Trade diversion – not import reduction – was the main effect on trade flows.
The impact on jobs is more difficult to assert, because the safeguard was decided against the background of a severe cyclical downturn in the steel industry, combined with a structural declining trend as a share of manufacturing. Monthly statistics show that the decline in the 12 months following the safeguard (-4.9%), was slower than in the 12 months before (-13.8%). Since the same pattern is also found for manufacturing as a whole (although less pronounced), there is no clear sign of a significant employment impact. The steel industry went through significant restructuring and consolidation in the period following the safeguard implementation, but the most detailed analysis available does not establish a direct link between this trend and the measures (United States International Trade Commission, 2003).
In assessing the broader impact of protection, this example is especially insightful because steel is an important input for many other industries. At the time the safeguard was decided, steel-using industries employed roughly 57 times more workers than the steel industry itself: 12.8 million compared to 170,000, respectively (Liebman & Tomlin, 2007). The price-increasing effect of protective measures is a major concern in such a situation, especially because most steel-using sectors are highly competitive, so that steel-using firms face difficulties is passing on price increases.
In practice, outcomes were heterogeneous across categories, but they exhibited strong price increases for important categories, with spot prices of steel sheets increasing by 40% or more in the four months following the safeguard, while producer price indices increased by 20% to 30% (although the initial increase was somewhat tempered later on; see USITC, 2003, Vol. III). According to one estimate, 200,000 jobs were lost in steel-using industries as a result of the safeguard, which is more than total employment in the steel industry itself at the same time (Francois & Baughman, 2003).
The 2003 assessment carried out by the USITC, which includes a detailed firm survey, sheds light on these issues. In addition to price increases, it found that almost half of responding steel-using firms (and many more in some cases) reported difficulty in obtaining steel in the quality and quantity desired. 11% of all responding firms reported that they had shifted to sourcing finished parts from overseas as a result of the safeguard measures, and this proportion reached 16% in steel fabricators and motor vehicle parts sectors, 29% in furniture and hardware, and 50% in household appliances.
Asked if the safeguard measures led them to relocate US steel-consuming facilities abroad, 7% responded that it did; this share was significantly higher in some industries, 11% among motor vehicle parts makers, 12% among steel fabricators, 19% in furniture and hardware, and 33% in household appliances. On other words, the safeguard not only caused trade diversion, but also the threat of diversion of production abroad. This suggests that the indirect costs were disproportionately high compared to direct benefits. The central, model-based estimate of the USITC for the resulting impact was a real-income cost of $42 million, but this does not factor in the cost of indirect job losses, which could be far larger. Just for the period of February to November 2002, Francois and Baughman (2003) calculate that the safeguard had originated a wage loss worth $4 billion)
Shares up, and down
Another illustrative aspect of this safeguard is the impact on share prices. In accordance with the remarks above, shares of firms in steel-consuming industries experienced significant negative abnormal returns in response to the initiation of the safeguard investigation and the affirmative injury decision by the USITC. In contrast, stock prices of steel producers significant increased. Within days of initiation of the investigation, steel producers’ shares increased by 6% to 8% beyond what might have been expected otherwise. They increased further by 5 to 6% within days of the decision to impose the safeguard, while losing more or less the same proportion of their value when the negative ruling of the WTO panel was announced. (The estimates are drawn from Liebman and Tomlin, 2006.) All these impacts are consistent with the rent-seeking motivation of those filing for protection.
To complete the description of the unintended costs of the safeguard, we emphasise the importance of the WTO dispute. Losing such an important and widely commented case in a multilateral arena involved significant reputational costs for the US, especially as the ruling emphasised that the measure was in breach of the country’s commitments in several respects. The case was also illustrative of the potential importance of retaliatory measures. As soon as May 2002, the EU notified the WTO that it reserved its right to rebalance the adverse effect of the US steel safeguards. It subsequently issued a list of products concerned by these would-be measures, which encompassed a wide range of goods, from orange juice and apples to sunglasses, knitwear, motor boats or photocopying machines, representing a total $2.242 billion of US exports to the EU. The EU list was intended to respond to the political motivation by political targeting: it targeted products whose production is important in politically sensitive states. The result for the United States was that even the political benefit of the safeguard was quickly undermined.
The bottom line of George W. Bush’s 2002 steel tariffs is clear: share prices increased for the firms filing for protection, at the expenses of everyone else. While the impact of Donald Trump’s protectionist measures remain to be seen, the 2002 case provides important insight.