Aggregate exports are highly concentrated in a small group of firms, which are generally more productive than other firms. This column shows that productivity shocks generate different responses of aggregate exports depending on the characteristics of the whole productivity distribution of active firms. Most of the action is driven by superstar firms. Aggregate exports grow considerably more if productivity shocks are concentrated among top performers than across the whole population of firms. This has important implications for the design of export support and industrial policies.
Aggregate exports from any country are highly concentrated in a small group of ‘superstar exporters’ (Freund and Pierola 2015, Gaubert and Itskhoki 2020). These large exporters are generally more productive than other firms (Bernard et al. 2007, Melitz and Redding 2014). The literature on trade with heterogeneous firms has helped us rethink the role of individual companies in export markets (Breinlich et al. 2020), introducing concepts such as the extensive margin (how many firms export in a given market) and the intensive margin (how much each firm exports in a given market on average).
Yet, it is not clear how the shapes of productivity distributions of the whole population of firms affect aggregate exports. Given two distributions with identical average productivity, we do not know if and how the variance or the skewness of such distributions, i.e. their second and third moments, which may differ, influence aggregate exports.
This is not just a statistical question. These higher moments provide important information that is too easily dismissed. The variance measures the distance between the most and the least productive firms. The skewness measures the asymmetry in the distribution whereby, for example, many firms could be concentrated at low levels of productivity and a few superstar champions achieve far superior performances. By studying the role of these higher moments, it is possible to show both theoretically and empirically that the aggregate export response to trade and productivity shocks is to a large extent driven by few large superstar exporters.
Looking at the shape of productivity distributions implies inquiring into which firm characteristics affect their export performance, and consequently how such micro-level characteristics combine to generate aggregate exports and aggregate export growth. Productivity is of course the crucial parameter, as substantial empirical evidence supports its positive relationship to export performance. But what matters to relate microeconomic characteristics to macro effects is not just the features of some firms, but rather the features of the whole productivity distribution of producers.
Two cases are especially worth studying within this framework. The first one is the elasticity of exports to trade costs. The second one is the impact of microeconomic shocks on productivity. Both have important implications for policy and welfare. Think first of policies that reduce trading costs, such as, for example, a possible reversal of the recent rise in tariffs and non-tariff trade barriers induced by the ‘trade war’ with China, or a reduction of the post-Brexit barriers between the EU and the UK. Their impact will depend on trade elasticities, i.e. how trade flows respond to such cost reductions. Given the heterogeneity of firm level responses to trade costs, we may well expect aggregate elasticities to be affected by the shape of productivity distributions, for example the number of superstars and their distance from other less productive firms.
Equally, the resurgence of industrial policy to enhance the green transition or technological sovereignty, policies such as the US Inflation Reduction Act (IRA) or the CHIPS Act, though not directly targeted to the export market, also aim at gaining competitiveness in targeted activities. Such policies cause microeconomic shocks that affect firm level productivity, for example by supporting new investment or the adoption of new technologies. How do these shocks boost export? Here again we expect the shape of productivity distributions and how these shocks alter different moments to play an important role. Policy wise, consequently, here is the difficult question: is it more effective to aim at industry-wide efficiency, or just focus on a few superstar champions?
In a recent paper (Barba Navaretti et al. 2024), we try to extend this knowledge by studying how different moments of the distribution of producers’ productivity affect the trade elasticity, and in turn how microeconomic shocks that alter those moments in different ways have different impacts on aggregate exports.
Gravity models are the most popular tool used to explain and predict international trade flows. The importance of the specific parametrisation of the distribution of producers’ productivity for gravity estimation within the canonical framework by Melitz (2003) is well-known since initial applications by Chaney (2008) and Helpman et al. (2008). Compared to these earlier micro-founded gravity models, our study enriches the analytical framework to account not only for the share of producers that export (the extensive margin), but also for the productivity premium of exporters relative to all producers. We show that by taking into full consideration the occurrence of ‘superstar exporters’ we can more precisely predict the response of aggregate exports to changes in trade costs and different productivity shocks. As for the former, we find that the smaller the share of exporters and the larger their productivity premium versus non-exporters, the more aggregate exports are driven by few overperforming ‘superstar exporters’.
As for the impact of shocks to the distribution of producers’ productivity on aggregate trade flows, we show how various moments of the distribution matter for aggregate exports. By perturbing the values of the distribution’s parameters, we study how different changes in its moments translate into different responses of aggregate exports and their margins. All in all, the simulation results highlight that identical increases in average productivity can result in very different increases in exports depending on the exact parameters (and the associated moments) of the productivity distributions causing them. In particular, exports rise more if the density on the upper tail grows and the initial share of exporters is small, which again highlights the essential role of superstar firms located in the upper tail of the distribution.
This finding is represented in Figure 1 below, which illustrates the impact of variations in productivity on total exports, differentiated by their source. Across all sources, simulations indicate that a 1% increase in average productivity leads to a rise in total exports. However, the same 1% increase in average productivity could stem from different sources: an efficiency boost among firms at the technology frontier, a shift in median productivity, or improvements in productivity among lagging firms.
Figure 1 Impact of 1% increase in mean productivity on total exports by source of the shock
Source: Authors’ calculation.
Note: Sector: 28 (NACE Rev.2); year: 2010; country: Italy.
The specific source significantly affects the trade response, even if the aggregate productivity change is the same. First, when average productivity growth stems from the most efficient firms becoming more productive, total exports rise by 6.6%. This increase at the technology frontier corresponds to an upward shift in the upper bound of the productivity distribution, or equivalently an increase in its right-wing (r.w.) skewness. Second, when the 1% average productivity growth is driven by rising productivity among the least efficient (laggard) firms, exports rise by 2%. In this case, there is an upward shift in the lower bound of the productivity distribution, or equivalently an increase in its left-wing (l.w.) skewness. Finally, when the 1% increase in average productivity is the consequence of an increase in the productivity of the median firm, total exports rise by 2.5%.
To sum up, the extent to which a given increase in a sector’s average productivity leads to more exports depends on which moment of the productivity distribution drives it. The biggest boost to exports materialises when the given increase in sectoral productivity is driven by more efficient superstar exporters.
Source : VOXeu
