The changing political climate in the UK and the US raises the possibility that tariffs and non-tariff barriers to trade may increase for the first time in generations. Both economic theory and the available evidence suggest that tariff increases reduce international trade. But will trade fall due to the mere possibility that tariffs may increase in the future? And if tariffs are increased, how long will it take for the full impact on trade to be felt? This depends on the way in which exporters and potential exporters make decisions.
When data on exports at the firm level first became available, researchers saw that participation in exporting was persistent. They noted that this persistence could potentially be accounted for by the existence of sunk costs of entering export markets (Roberts and Tybout 1997). Estimates of the sunk costs of export-entry necessary to account for the average persistence of exporting were very large indeed (Das et al. 2007). As pointed out by Dixit (1989) and Baldwin and Krugman (1989), when sunk costs are large, existing exporters may not exit even if market conditions deteriorate. At the same time, sunk costs make potential exporters reluctant to enter new markets if the future payoff for doing so is uncertain. The resulting behaviour of entry and exit may reduce trade in response to the mere possibility of tariff increases, while slowing down the impact of tariff increases after they are implemented.
However, more recent research shows that while exporting is persistent on average, export entrants initially face a very high probability of exit as well as having low export sales (Eaton et al. 2014, Ruhl and Willis 2016). These facts are inconsistent with the existence of large sunk costs of entering export markets. Instead, some other mechanism must be driving the behaviour of exporters. Does this mean we should expect swift responses to tariff changes?
In new research, we show that examining the separate contributions of quantities and prices to the dynamics of exports, rather than focusing solely on revenue, is very informative about what drives exporter behaviour (Fitzgerald et al. 2016). Our research points to a key role for the accumulation of a customer base. It also points to a role for exporter learning. Both of these features of exporter behaviour imply that export responses to tariff increases are likely to be gradual, even in the absence of large upfront costs of entering an export market.
Export quantity and price dynamics
We use customs data on Ireland's manufacturing exports over a 19-year period to document the separate contributions of quantities and prices to exporter dynamics. Consistent with the past literature, almost half of export episodes are very short-lived, and involve very low levels of sales. A substantial fraction of export episodes start somewhat bigger, last somewhat longer and exhibit hump-shaped sales dynamics. The most successful export episodes are relatively rare, and are characterised by very high initial sales (four times as big as in the shortest episodes), and also by ongoing growth of sales. In first five years of these episodes, sales quadruple, after which sales growth levels off.
Figure 1 Export sales trajectories
Note: Sales are normalised to equal 1 in export episodes that last 1 year. See Fitzgerald et al. (2016) for details.
Figure 2 Distribution of export episodes by duration
Note: See Fitzgerald et al. (2016) for details.
But perhaps surprisingly, these dynamics of sales are driven entirely by the behaviour of quantities, without any accompanying dynamics in prices. Moreover, while survival in an export market is positively correlated with initial quantities, survival is completely unrelated to initial prices. These facts hold across different manufacturing sectors, across small and large firms, across domestic-owned firms and multinationals, and across export markets that are close by (the EU) and export markets that are more distant (non-EU countries).
Figure 3 Export quantity trajectories
Note: Quantities are normalised to equal 1 in export episodes that last 1 year. See Fitzgerald et al. (2016) for details.
Figure 4 Export price trajectories
Note: Prices are normalised to equal 1 in export episodes that last 1 year. See Fitzgerald et al. (2016) for details.
An important role for the customer base
Economists typically emphasise the importance of prices in allocating quantities in competitive markets. But according to our findings, in successful export episodes, firms quadruple their quantities sold without changing their prices at all. This strongly suggests that accumulation of a customer base plays an important role in export expansion. It also suggests that non-price activities of the firm – such as marketing and advertising – play a key role in acquiring and maintaining that customer base.
Depending on how marketing and advertising activities are defined, they account for between 2% and 8% of GDP (Arkolakis 2010). This is a similar order of magnitude to the share of GDP devoted to research and development. The resources devoted by firms to developing markets for their existing products are at least as large as those devoted to developing new products and improving production processes. This is consistent with our findings and suggests that firms see the customer base as key to their success.
Learning about demand
The large number of short-lived export episodes, together with the fact that higher quantities sold in the initial year forecast longer survival in a market, implies that firms need to actually sell in a market in order to learn whether their demand is strong or weak. Then, if they learn that demand is strong, they remain in the market. If they learn that it is weak, they exit. A number of recent papers (e.g. Berman et al. 2015, Timoshenko 2015, Bastos et al. 2016) highlight this mechanism. Our finding that it is initial quantities rather than initial prices that forecast survival in an export market suggests that firms are not using information from prices to learn whether demand is high or low. This is yet more evidence that the role of prices is not as primal as economists often assume, and is consistent with an important role for the customer base in driving the behaviour of exports.
We go on to examine why exporters accumulate a customer base gradually, rather than all at once, since we see that successful export entrants continue growing for at least five years after entry. We do this by estimating a variety of different models of the exporter's decision problem. Our estimation procedure tries to match the behaviour of quantities and prices, as well as the proportion of export episodes of different length, and judges a model's success by how well it matches these facts. The gradual growth of quantities in the most successful export spells implies that even after learning that their demand is strong, firms spread out their accumulation of their customer base to economise on costly marketing and advertising activities. Our preferred model incorporates this feature, and closely matches the behaviour of quantities and prices in the data.
While designed to match exporter behaviour without reference to tariffs, our model has important implications for how aggregate trade responds to tariff increases. The impact of a reduction in export entry takes time to feed through into aggregates, as entrants are initially small, and successful entrants take time to grow. An accumulated customer base acts like a sunk cost, delaying the exit of marginal incumbent firms. In addition, costs of adjusting marketing and advertising delay the responses of successful incumbents, magnifying the impact on overall trade flows. Simulations based on our model suggest that it may take more than a decade before the full impact of any tariff increases on trade is felt. Simulations looking at the impact of potential future tariff increases suggest modest responses before the changes are actually implemented.
New facts about the evolution of quantities and prices for export entrants suggest an important role for the customer base in explaining exporter behaviour. This implies that aggregate exports are likely to respond slowly to increases in tariffs, and may not respond much in advance of the actual implementation of those increases.
Authors’ note: The views expressed in this column are those of the authors, and not necessarily those of the Federal Reserve Bank of Minneapolis.
Arkolakis, C (2010), “Market Penetration Costs and the New Consumers Margin in International Trade,” Journal of Political Economy, 118, 1151-1199.
Baldwin, R, and P Krugman (1989), “Persistent Trade Effects of Large Exchange Rate Shocks,” Quarterly Journal of Economics, 104, 635-654.
Bastos, P, D A Dias, and O Timoshenko (2016), “Learning, Prices, and Firm Dynamics,” mimeo.
Berman, N, V Rebeyrol, and V Vicard (2015), “Demand Learning and Firm Dynamics: Evidence from Exporters,” mimeo.
Das, M, M Roberts and J Tybout (2007), “Market Entry Costs, Producer Heterogeneity and Export Dynamics,” Econometrica, 75, 837-873.
Dixit, A (1989), “Hysteresis, Import Penetration, and Exchange Rate Pass-Through,” Quarterly Journal of Economics, 104, 205-228.
Eaton, J, M Eslava, D Jinkins, C Krizan, and J Tybout (2014), “A Search and Learning Model of Export Dynamics,” mimeo.
Fitzgerald, D, S Haller and Y Yedid-Levi (2016), “How Exporters Grow,” NBER Working Paper 21935.
Roberts, M, and J Tybout (1997), “The Decision to Export in Colombia: An Emprical Model of Entry with Sunk Costs,” American Economic Review, 87, 545-563.
Ruhl, K, and J Willis (2016), “New Exporter Dynamics,” International Economic Review, forthcoming.
Timoshenko, O A, (2015), “Learning versus Sunk Costs Explanations of Export Persistence,” European Economic Review, 79, 113-128.