Exports and productivity selection effects for Dutch firms

The paper investigates whether the self-selection hypothesis and other predictions from the heterogeneous-firms trade models can explain the export participation patterns for Dutch firms in manufacturing and services. The results provide strong support for the self-selection hypothesis, according to which firms need higher productivity performance to compensate for sunk entry costs in export markets. After controlling for many firm and market characteristics we robustly find higher productivity levels for exporters. The paper also tests for the reverse causality (learning-by-exporting), but finds no empirical support for it, not even after controlling for the firm's distance to a constructed international productivity frontier. This latter result may be important for the motivation of future export promotion policies. The empirical estimates are achieved by probit regressions at the plant level and at the firm level. As a robustness test we also applied the more standard OLS panel regression estimates, which provided similar results. The paper also tested whether the productivity-export link is conditional on the sectoral market structure and multinational affiliation. Services sectors with high competition and a lower degree of product differentiation have significantly higher export productivity premia than services firms in less competitive sectors. Such differences are not found in the manufacturing sector.