Ceo Behavior and Firm Performance

Abstract : We measure the behavior of 1,114 CEOs in six countries parsing granular CEO diary data through an unsupervised machine learning algorithm. The algorithm uncovers two distinct behavioral types: “leaders” and “managers”. Leaders focus on multi-function, high-level meetings, while managers focus on one-to-one meetings with core functions. Firms with leader CEOs are on average more productive, and this di↵erence arises only after the CEO is hired. The data is consistent with horizontal di↵erentiation of CEO behavioral types, and firm-CEO matching frictions. We estimate that 17% of sample CEOs are mismatched, and that mismatches are associated with significant productivity losses.

---

A Structural Topic Model of the Features and the Cultural Origins of Bacon's Ideas

Abstract : We use machine-learning methods to study the features and origins of the ideas of Francis Bacon, a key figure who provided the intellectual roots for a cultural paradigm that spurred modern economic development. We estimate a structural topic model, a state-of-the-art methodology for analysis of text corpora. The estimates uncover sixteen topics prominent in Bacon's opus. Two are key elements of the ideas usually associated with Bacon—inductive epistemology and factseeking. We provide the first quantitative evidence that the genesis of Bacon's epistemology lies in his common-law jurisprudence, a conclusion that has not been prominent in the conventional text-analysis literature. Fact-seeking is more isolated from Bacon's other intellectual pursuits. The utilitarian promise of science and the centralized organization of the scientific quest, embraced by Bacon's followers, were not emphasized by him. Bacon's use of different topics varies notably with intended audience and chosen medium. Our results have direct implications for the interpretation of the determinants of political and economic development in 17th-century England.

---

Using Machine Learning to Predict Reasonable Notice Awards

Abstract : Under Canadian employment law, when an employee is terminated without cause, they are entitled to a reasonable notice period from their employer – or payment in lieu of this notice period. What is “reasonable” for any given worker is, however, determined on a case-by-case basis. Employers and lawyers currently tend to rely heavily on length of service when seeking to predict how a court might decide reasonable notice cases. These rules of thumb are very weak predictors. The average prediction error generated by these simple heuristics is large. Machine learning techniques vastly outperform these simple heuristics in terms of predicting the outcome of reasonable notice cases. I use machine learning algorithms that have been developed for the purpose of prediction. Familiar estimation techniques, such as ordinary least squares, do (of course) offer simple ways to generate predictions. But they do not offer the best means of predicting out-of-sample. I explore the power of boosted decision trees, random forests, and neural net regression to predict how a court will decide cases not included in my training data. Boosted decision trees, tuned by depth, offer the best means of predicting court outcomes in my sample. In terms of explanatory power, boosted decision trees outperform the linear prediction model by as much as 12% and reduce the prediction error by as much as 15%. Random forest algorithms also outperform the linear prediction model, but the performance is weaker than with boosted decision trees. Neural net regression algorithms do not outperform the linear model, which is not surprising, given the size of our dataset.

---