Protectionist trade policy impact on Employment

An economic paper by Alessandro Barattieri & Matteo Cacciatore


In 2018, the U.S. administration imposed new tariffs on roughly 12% of imports, sparking debates on the effects of protectionism to unprecedented levels. A distinguishing feature of recent U.S. trade policy is the focus on the access to global supply chains (Krugman, 2018, and Baldwin, 2018). For instance, tariffs against Chinese imports have heavily targeted intermediate inputs, with nearly $50 billion of imports of steel and aluminium being affected (Bown, 2018a). Such a shift of trade protection towards intermediate inputs is not an isolated episode. Hidden behind unchanging tariff policies, governments have been using temporary trade barriers (TTBs)—antidumping, countervailing duties, and safeguards—to restrict trade in intermediate inputs since the last two decades (Bown, 2018b). In particular, while governments have maintained lower tariffs on factor inputs relative to final goods, TTBs on imported intermediates have been on average higher and growing relative to TTBs on final goods

In light of these events and considerations, it is not surprising that much of the discussions on the effects of protectionism contrast potential gains in protected industries and possible negative effects on downstream sectors—the producers that use protected goods as intermediate inputs.1 However, despite the relevance of supply-chains considerations, econometric evidence on the effects of protectionism through vertical production linkages remains scant. We address this issue by studying the employment effects of TTBs in protected and downstream industries

Our contribution to the literature is threefold. First, we provide novel evidence on the role of production networks in propagating protectionism targeted to specific industries. Second, while the trade literature typically focuses on the long-run effects of permanent tariff reductions, we provide evidence on the dynamic consequences of TTBs. Third, we exploit a novel high-frequency identification of temporary trade-policy shocks at a disaggregated industry-level

The analysis proceeds in three steps. First, we identify movements in protectionism that are plausibly unanticipated and not correlated with economic fundamentals. Our approach builds on a consolidated strategy in the monetary and fiscal policy literature, following the seminal work by Romer and Romer (2004). The idea is to purge the series of interest (TTBs protection) of movements taken in response to past, current, and expected dynamics in the outcome variable of interest (employment growth). The remaining variation allows us to identify the effects of protectionism on employment within and across industries, even in the scenario in which such variation is not strictly exogenous, i.e., when other factors not affecting employment dynamics drive the residual TTBs dynamics.

We consider two alternative and complementary approaches to identify trade-policy shocks. One focuses on within-industry time-series variation, while the other one uses the panel dimension of the data. In both cases, we exploit regulation-induced lags in the opening of an investigation to impose short-run restrictions—TTBs cannot react to economic shocks within a month. We then control for past economic conditions and measures that capture expected economic outcomes. In particular, using firm-level data, we construct for each industry a benchmark measure of expected returns used in the finance literature, the market-to-book ratio (e.g., Pontiff and Schall, 1998, and subsequent literature). With panel data, we also include industry and time fixed effects, stronger controls for unobserved heterogeneity and common shocks.

Our analysis yields three main results. First, protectionism has small and short-lived beneficial effects on industry employment. Across specifications, an increase in the share of imports subject to TTBs equal to 2 percentage points (the average import share affected by TTBs in the episodes we analyze) leads to an employment increase at most equal to 0.15 percentage points on average. The response turns negative after approximately one year. The effects are in general statistically insignificant. This finding is consistent with different explanations, including the fact that TTBs affect profits (e.g., markups) rather than output in protected industries, heterogenous responses across producers (e.g., a different exposure to products covered by TTBs), as well as the presence of offsetting forces determining industry’s output demand (e.g., expenditure switching versus negative income effects).

Second, protectionism has negative, persistent, and statistically-significant effects on employment in downstream industries. Our estimates imply that a uniform 2 percentage-point increase in the share of imports subject to TTBs in upstream industries generates an average employment decline up to 1 percentage point after two years. Third, the downstream employment loss is accompanied by a statistically significant increase in both intermediate-input and final producer prices. From a timing perspective, the peak of the price increase precedes the peak of the employment decline, suggesting that it is indeed a loss of 3 competitiveness that causes employment losses. In addition, employment falls more on average when goods are more substitutable

Prices

There exist alternative possible explanations for the negative effects of protectionism on downstream employment. For instance, when an intermediate input is subject to TTBs, downstream producers may find it hard to quickly replace it, ending up paying a higher price. Alternatively, producers may switch to potentially less-efficient suppliers, facing higher prices relative to the pre-TTBs scenario. While these two scenarios have different implications for the response of imports, marginal costs and final-producer prices in downstream industries are predicted to increase in both cases. In turn, higher prices reduce competitiveness, lowering demand and employment.