Die another day: duration in German import trade

Abstract

International trade patterns at the product level are surprisingly dynamic. The majority of trade relationships exist for just a few, often only one to three, years. In this paper, I examine empirically the duration in German import trade at the 8-digit product level from 1995 to 2005. I find that survival probabilities are affected by exporter characteristics, product type and market structure. Specifically, I show that the duration of exporting a product to Germany is longer for products obtained from countries that are economically large and geographically close to Germany; for products with large trade value and a low elasticity of substitution; and for trade pairs that command a large share of the German import market and are characterized by two-way trade.JEL Code: F14, F19, C14, C41.

What Makes a Successful Exporter?

“…As this figure makes clear the hazard reaches its maximum between 3 or 4 years, and then starts to decrease rapidly after six years or so. This length of time is close to those estimated from the duration analysis on imports to the US and to Germany using disaggregated bilateral trade data by Besedes and Prusa (2006) and Nitsch (2007). This finding may be interpreted in different ways.…”

“…There is also a somewhat related field in international economics examining the duration of imports (Besedes and Prusa, 2006a,b;Nitsch, 2007). This literature uses only information on imports of products, not at the firm-product level as in our paper.…”

Exploring the duration of EU imports

“…However, since we, unlike, for example, Besedeš and Prusa (2006a,b) and Nitsch (2009), have both importer and exporter variation, we are able to control for both importer and exporter characteristics. Based on the findings in the descriptive analysis, we also include some variables in our analysis that have not previously been used.…”

“…Besedeš (2008) builds on this work, and investigates further the implications of the model by Rauch and Watson (2003). Nitsch (2009) examines the duration of German imports at the 8-digit (Combined Nomenclature) product level, using data from Eurostat for 1995-2005. Employing a stratified Cox proportional hazards model, he investigates the effects of numerous regressors on the hazard rate.…”

“…Besedeš and Prusa (2006a,b), for example, find that the median duration for US imports (at the 4-digit SITC level of aggregation) is merely two years. Nitsch (2009), using much more detailed data, finds that the median duration of German imports is two years. In other words, the conclusion from empirical studies so far is that most trade flows do not survive for very long.…”