Abstract
Cross-industry cross-country models are applied widely in economics. For example, to investigate the effect of financial development on economic growth or the effect of institutional quality on international trade. The literature estimates the effect of interest by examining the interaction between country characteristics|for example, financial development or institutional quality -and theoretically relevant technological industry characteristics- for example, dependence on external finance or relationship-specific inputs. As the relevant industry characteristics are unobservable for most countries, they are proxied by industry characteristics of a benchmark country. We analyze this approach when there is cross-country heterogeneity in technological industry characteristics. First, we show that the estimation approach in the literature is biased and that the bias cannot be signed if technologically similar countries are similar in terms of other characteristics. Second, we derive necessary and sufficient conditions for identification of the effect of interest. Third, we use the new identification approach to reestimate the impact of institutional quality on comparative advantage in industries that rely on relationship-specific inputs.