Family Matters in a Meritocracy: Networks, Civil Service Exams, and Officialdom in the Joseon Dynasty Sok Chul Hong† Christopher Paik‡ Yangkeun Yun§ July 2019 Abstract How do family networks influence social mobility in a meritocracy? Climbing the ladder of success may be fraught with nepotism and corruption, especially in monarchies where connections can trump talent. A merit-based selection of government officials in such context may serve as a remedy to curb these negative outcomes. In this paper, we investigate the effects of family networks on successfully obtaining official positions during the Joseon Dynasty from 1392 to 1897 CE. The Korean kingdom implemented literary examinations intended to fill central official positions based on merit. Its comprehensive records on family ties, exam results and official positions span over 503 years, longer than any other such data under a single dynasty in the world to our knowledge, and offer researchers a unique opportunity to investigate the efficacy of merit- based selections of political elites under a monarchy. We use an individually linked database of successful candidates and their family members from the literary examination rosters and official position information. We find that those from more connected predecessors in the network had significantly higher likelihoods of obtaining high-level rank positions after passing the exams, even when conditioning on age and performance at the examination. In light of the persistent family network influence, we evaluate the efficacy of meritocratic selection of political elites under a monarchy, and changing relevance of family networks as reflective of state performance over time. This study was financially supported by Center for Distributive Justice at Seoul National University (Grant Number: 0405-20180014) † Professor, Department of Economics, Seoul National University; [email protected] ‡ Assistant Professor, Division of Social Science, New York University Abu Dhabi; [email protected] § Ph.D. student, Department of Economics, University of Connecticut; [email protected] 1 1. Introduction Throughout history, monarchies have existed as common forms of governance. Hierarchical order and class divisions have often characterized their structures, in which social mobility was limited and representation in government was reserved for only the powerful few with connections. In order to combat nepotism and corruption from within, the rulers sought out ways to select government officials based on merit. An examination system that screened talented and capable candidates would, in theory, serve as a remedy to curb these negative outcomes. Similar to the civil service examination system in China, the Joseon Dynasty (Korean dynastic kingdom, 1392-1897 CE) was a centralized bureaucratic state which gave opportunities for officialdom to those who succeeded in Joseon’s merit-based examination. Those who passed gwageo, the Civil Service Examination, formed the ruling class. The examination system was the most significant means of recruiting officials for major central and provincial government posts (Wagner, 1974). Among the different types of exams under gwageo, mungwa (the literary examination) in particular was the most selective and accordingly prized as ensuring the elite status in society.1 Did the merit-based selection process actually work? Family connections, particularly in the form of prestigious lineage of mungwa passers, could strongly influence the career paths of elites. According to Kyŏngguk taejŏn (the National Code) and its subsequent laws, a variety of promotion standards were applied (Kim, 2017). The assignment of official position was apparently based on not only the competence of specific candidates but also recommendations from other officials as well as the king, particularly for high-level officials. In other words, the road to eventual officialdom was likely met with subjective factors such as family connections and king’s favors, in addition to scholarly ability (Won, 2007; Kim, 2017). Passing mungwa thus might not have been a sufficient condition to obtain a high-ranking position in the court. In this paper, we investigate the effects of family networks in obtaining official positions 1 Gwageo were composed of four categories: (1) mungwa (literary examination); (2) mugwa (military examination); (3) japgwa (technical examination); and (4) saengwon jinsa (classics and literary licentiate examination). We focus on those who succeeded in mungwa because they represented the ruling class of Joseon as major state officials. Successful candidates of saengwon jinsa became qualified to enroll at Sungkyunkwan, the National Confucian Academy, which trained students for mungwa. They could alternatively be appointed as ninth-ranked junior officials, which were the lowest positions among the court officials. Those who passed mugwa (military examination) or japgwa (technical examination) were regarded as lower-class officials. Table A1 in Appendix A provides the official ranking system during the Joseon Dynasty. 2 during the Joseon Dynasty. We use an individually linked database of successful mungwa passers and their family members from both mungwa rosters and appointment records of government officials. We find that those who are central, i.e. having more connected predecessors in the network, had significantly higher likelihoods of getting high-level rank positions after passing the exams, even when conditioning on age and performance at the examination. Specifically, our centrality score for each successful mungwa passer considers whether the ancestors themselves also passed mungwa. Because only the successful candidates appeared on the exam rosters, the ancestors had records of their own only if they passed the exam. The successful candidates with more ancestors who passed the exams themselves thus would have more connections, and become central by definition. We capture this score by using the eigenvector centrality measure, which accounts for the number of each candidate’s ties as well as the connections of the ties themselves (Bonacich, 1972, 1987; Jackson, 2010). The measure allows one to capture how those connected to the candidate are themselves influential, and is often used to assess prestige and popularity (Cruz et al. 2017; Jackson, 2010). Under the Confucian tradition that dominated the morals of society, scholarly achievement was of utmost importance for the elites in Korea. In our case, a higher score would indicate a more academic, and thus prestigious and influential, family connection. Throughout five centuries of rule in which the merit-based exam was in effect, we find that family connection was a key factor in selection into officialdom. The estimates from our preferred specifications, which controls for family clan, king in rule, pre-exam social status, exam type, year-of-birth, and residence fixed effects among other, suggest that one standard deviation increase in the natural log of eigenvector centrality is associated with approximately a 4 percent point increase in the likelihood of being high-level officials.2 We find this result for the positions that are higher than or equal to the upper senior third rank, the upper echelon of political elites in Joseon. The case of Joseon is of interest for its commonalities with other monarchies, but also for its unique context and comprehensive historical records. Nepotism among political elites likely exists in any governance structure, and arguably remains more prevalent in non-democratic settings where political representation and accountability are limited. Throughout history, monarchies represented the most common types of government. They adopted and institutionalized mechanisms to curb the influence of family connections and sought to recruit talent, while still 2 The probability that a candidate reached high-level officials was 56% in our sample. 3 maintaining strict social hierarchies. One of these institutional inventions was the merit-based selection of government officials. First found in Chinese dynasties, other Asian countries including Korea and Vietnam adopted similar practices. In a broader context, exam-based civil servant selections were also later found in the British government and other European states, as well as the United States. Among the monarchies that adopted the examination system, the Joseon dynasty stands out as an invaluable case study offering comprehensive records of exam passers, their eventual career paths as well as family connections spanning over five centuries. For example, in addition to information on the history of passing exams and the ranks of official positions the ancestors had obtained, the data allow us to analyze family networks through both marriages as well as patrilineal connections.3 Marriage is an important underlying mechanism that influences social status, as it plays a key role in preserving status-relevant family groups (Clark, 2014; Shiue, 2016). We have information on the marriage relations of each candidate who passed mungwa, with profile information of the candidate’s maternal grandfather and father-in-law. To our knowledge, these records comprise the world’s longest continual data of such kind under single dynasty. Our paper is also the first study examining the eventual political ascension of successful examination candidates, and the importance of family networks that could compromise the principles of meritocracy.4 Our work relates to several strands of the literature. First,