Testing for path dependence in economic history publications

This paper asks whether the concept of path dependence applies to economic history’s own publication record. Using annual citation histories for 1262 articles published between 2012 and 2021 in four economic history journals, I document strong persistence: conditional on journal and publication year, the elasticity of long-run citations with respect to age-two citations is about 0.78. The relationship holds when the outcome is restricted to citations accumulated strictly after the early window, and it is stable across journals, topics and publication cohorts. Observable author and article characteristics explain about a fifth of the variation in early citations; yet, the unexplained component predicts long-run citations almost as strongly as the explained component. Consistent with a visibility channel, articles placed earlier in a journal issue receive more early and more long-run citations. Together these patterns are difficult to attribute to quality differences alone, although unobserved quality cannot be ruled out. Half of the classified citation links in the available network sample come from economics, and two fifths come from formats other than journal articles. The cross-sectional association between conference presentation and long-run citations is imprecise conditional on early citations. The findings are relevant to the use of early citation counts in research evaluation.