The growing importance of citation-based bibliometric indicators in shaping the prospects of academic careers incentivizes scientists to boost the numbers of citations they receive. Whereas the exploitation of self-citations has been extensively documented, the impact of re-ciprocated citations has not yet been studied. We study reciprocity in a citation network of authors, and compare it with the average reciprocity computed in an ensemble of null network models. We show that obtaining citations through reciprocity correlates negatively with a successful career in the long term. Nevertheless, at the aggregate level we show evidence of a steady increase in reciprocity over the years, largely fuelled by the exchange of citations between coauthors. Our results characterize the structure of author networks in a time of increasing emphasis on citation-based indicators, and we discuss their implzications towards a fairer assessment of academic impact.The majority of measures that evaluate academic impact are based on citations of publications. Ranging from crude citation counts to the well known journal impact factor [1], the Eigenfactor metrics for journal rankings [2] and the h-index [3] for authors, such measures are increasingly relied upon to inform all aspects of academic decision-making, including faculty recruiting, grant attribution, and the formation of collaborations [4]. The onset of such a trend dates back to the 1950s, when the earliest citation-based indices to assess academic impact were put forward [5]. The following decades saw the proliferation and ever-increasing adoption of such indicators in all fields of science [6,7], which eventually led to the systematic analysis of academic citations, to the emergence of bibliometrics as a research field [8], and, more recently, to the rise of a "science of science" devoted to understanding the determinants of scientific success [9][10][11].Given the importance that citations play nowadays in shaping the prospects of an academic career, it is certainly not surprising to see an increased attention to the study of citation patterns and of the publication strategies that can attract a larger number of citations. While an author's productivity and the quality of her work are obvious determinants of academic success [12], other less tangible aspects have recently been identified as key contributors to success. For example, the social network a scientist is embedded in has recently been shown to play a relevant role in determining her future chances of success [13].Other social and behavioural considerations also play relevant roles in modern academic reputation systems, especially concerning the proliferation of self-citations and citations between close collaborators [14-17]. Self-citations and their role in inflating bibliometric indicators have been studied extensively [18]. Several scholars have proposed revised metrics to mitigate their distorting impact [19], and indeed some of the most popular citation indexing services (e.g., Web of Science) and social ne...