Leader's Baggage? A Latent Variable Model and New Data on the Experience of Political Leaders

There is a well-developed political selection literature, which finds that individuals sharing particular traits are more likely to be selected for leadership. We know that democracies seem to attract more educated leaders (Besley and Reynal-Querol 2011). We also know that democracies tend to have leaders with a legal background and non-democratic, particularly military, regimes, those with military backgrounds (Blondel, 1987; Eulau and Czudnowski, 1972). This is because individuals with traits conforming to the norms of existing political regimes find it easier to progress their political careers. While it is debatable whether education or competence matters for governance (e.g., Bienen and van de Walle, 1991; Carnes and Lupu, 2015), the view that democracies are steered by more experienced leaders than dictatorships are, has arguably attained the status of a conventional wisdom. Given recent watershed victories of political novices such as Donald Trump or Emmanuel Macron over more experienced opponents in the US and France as well as the rise across many democracies of political outsiders to leadership, in sharp contrast to a more experienced leadership in contemporary dictatorships of China, Russia or Vietnam, there are grounds to re-examine whether democracies continue to select more experienced leaders. Drawing from a new data set of prior political experience of 2,025 national political leaders across the world from 1950-2017, we propose a new measurement model of a leader's experience. We argue that the experience is a multifaceted concept and that we need to account for a range of factors in assessing an individual's level of experience. We include the duration in formal and informal politics prior to assuming the highest office; the quality of experience - political offices and ministerial portfolios occupied in the past; as well as non-political experience, particularly in business management and public service that may substitute for political experience. In contrast to the existing literature that relies on particular proxies for experience, such as age, education or military rank, we develop a latent variable model using all relevant, multidimensional data. Using a Bayesian latent variable model on a range of experience-related variables, while accounting for age, we develop a new measurement tool of political experience: the PolEx. This measure can be applied across both democracies and non-democracies and provides an estimate of a leader's level of prior experience, as well as associated measures of uncertainty. The PolEx allows for a cross-national and longitudinal investigation of the political selection, as well as impact of prior experience on national leadership. We find support for the argument that democracies generally select more experienced leaders over time. However, because in the last two decades democratic leaders have become less experienced on all measured dimensions, while non-democratic leaders more so, there is a significant convergence among regime types. We also find support for the Linzean argument that presidential democracies select less experienced leaders than parliamentary regimes. Non-democratic regimes also vary in terms of experience of their political leaders, depending on the regime type.