International trade benefits consumers by lowering the prices of the goods they consume. An important distinction is that between trade in final goods and trade in intermediate inputs. While the former benefits consumers directly (see Broda and Weinstein (2006) for a quantification), the latter operates only indirectly. By allowing firms to access novel, cheaper or higher quality inputs, input trade reduces firms’ production costs and thus the prices of locally produced goods. In our recent work (Blaum, Lelarge and Peters 2015), we quantify this indirect channel, which we refer to as the gains from input trade. In contrast to existing work which uses only aggregate information, we rely on firm-level data and show that doing so is quantitatively important.
A recent body of work has incorporated input trade into quantitative trade models – see e.g. Eaton, Kortum and Kramarz (2011), Caliendo and Parro (2015) and Costinot and Rodriguez-Clare (2014). These frameworks have the convenient implication that the welfare consequences of input trade are fully determined from aggregate data. This property, however, only holds when firms’ import intensities are equalised – a feature that is at odds with the data. Figure 1 depicts the cross-sectional distribution of French manufacturing firms’ import intensities, defined as the share of material spending allocated to foreign varieties. We find substantial heterogeneity. While most firms spend less than 10% of their material spending in foreign inputs, some firms are heavy importers with shares exceeding 50%. Does this micro-level heterogeneity affect our understanding of the aggregate gains from input trade? Our research shows that it does.
Figure 1. The distribution of French manufacturing firms’ import intensities (share of material spending allocated to foreign varieties)
We focus on a widely used class of models of importing with heterogeneous firms, where firms’ demand system between domestic and foreign inputs has a constant elasticity of substitution (CES) but we leave other aspects of the import environment unrestricted.1 This class includes recent contributions such as Gopinath and Neiman (2014), Antras, Fort and Tintelnot (2014) and Halpern, Koren and Szeidl (2015). In this context, we show that the effect of any shock to the input trading environment (e.g. a decline in trade costs or an improvement in foreign technology) on domestic consumer prices can be measured from firm-level data on import intensity and value added. In particular, knowledge of the joint distribution of firm size and the change in import intensities resulting from the shock is sufficient to measure the change in consumer prices of locally produced goods. While this result can be used to study any counterfactual, a focal point for the analysis is the case of a reversal to input autarky – a situation where firms are forced to buy their inputs domestically. Because in such case counterfactual import intensities are zero for all firms, the change in consumer prices between the observed equilibrium and input autarky can be directly measured from the observed micro data.
Why does the micro data matter?
Import intensities are important because they summarise how production costs are affected by the use of imported inputs. A high material import share indicates that the firm benefits substantially from input trade – that is, it would see its unit cost increase substantially in the absence of trade. In this sense, Figure 1 shows the dispersion in the gains from trade at the micro level. To aggregate these firm-level gains into the aggregate gains for local consumers, one needs to know each firm’s importance in the economy. In particular, when relatively large firms tend to have high import shares, the aggregate gains from input trade will turn out to be large. The extent to which this is the case in France is depicted in Figure 2. Surprisingly, we find that the relationship between import intensity and firm size is essentially flat and that there is substantial dispersion in import shares conditional on size. It is this data which crucially determines the magnitude of the gains from input trade.
Figure 2. Firm size and imported inputs
Quantifying the gains from input trade
We apply our methodology to quantify the gains from input trade relative to autarky to the population of manufacturing firms in France. We allow for multiple sectors, with a rich structure of input-output linkages. We find that input trade reduces consumer prices of manufacturing products by 27.5%. That is, the prices of manufacturing goods produced in France would be 27.5% higher if French firms were not allowed to source their inputs from abroad. Relying on aggregate data would result in a significantly different estimate. First, for given parameters, we find that ignoring the dispersion in firms’ import intensities would result in overestimating the price index by 10%. Second, an important parameter required by our methodology is the elasticity of substitution between domestic and foreign varieties. In our work, we identify it from firm-level data with a methodology akin to production function estimation. We find a value for this elasticity of about 2, which is substantially smaller than the estimates based on aggregate data (usually above 4). This effect leads to underestimating the price index gains by about 50%. Hence, relying on aggregate data can lead to substantial biases in the estimates of the gains from input trade.
The effects of a devaluation
Finally, we consider unobserved conterfactuals and focus on a currency devaluation which makes imported inputs relatively more expensive. Because firms’ import intensities after the shock are unobservable, we need to specify a full model of importing to predict them. We consider a model of fixed costs and calibrate it to salient features of the micro data depicted in Figures 1 and 2. We find that a devaluation that decreases the aggregate trade share by 20% leads to an increase in the consumer price index of 7%. A parametrisation of the model that is not disciplined by the micro data predicts a change in consumer prices that is almost 20% higher. We conclude that the observed distribution of import intensities and size is crucial to quantify the normative consequences of changes in firms’ ability to source foreign inputs.
In this column, we argued that access to firm-level data on import behaviour is important to accurately measure the normative consequences of input trade. We showed that importing patterns display substantial heterogeneity at the firm level and that taking this dispersion into account has a quantitatively meaningful impact in the estimates of the gains from input trade and the effects of counterfactuals. These findings are relevant for the further development of quantitative trade models and provide guidance to policymakers in the field of trade policy.
Antras, Pol, Theresa Fort and Felix Tintelnot (2014), “The Margins of Global Sourcing: Theory and Evidence from U.S. firms”, working paper
Blaum, Joaquin, Claire Lelarge and Michael Peters (2015) “The Gains from Input Trade in Firm-Based Models of Importing”, NBER working paper 21504
Broda, Christian and David Weinstein (2006), "Globalization and the Gains from Variety," Quarterly Journal of Economics, vol. 121(2), p.541-585
Costinot, Arnaud and Andres Rodriguez Clare (2014), “Trade Theory with Numbers: Quantifying the Consequences of Globalization”, Handbook of International Economics
Caliendo, Lorenzo and Fernando Parro (2015), “Estimates of the Trade and Welfare Effects of NAFTA”, The Review of Economic Studies 82(1), p.1-44
Eaton, Jonathan, Samuel Kortum and Francis Kramarz (2011), “An Anatomy of International Trade: Evidence from French Firms”, Econometrica 79(5), p.1453-1498
Gopinath, Gita and Brent Neiman (2014), “Trade Adjustment and Productivity in Large Crisis”, American Economic Review, 2014, 104(3), p.793-831.
1 For example, we do not specify how prices or qualities vary by country, or how firms determine their set of trading partners (the extensive margin of trade).