We examine the effects of the big five personalities of CEOs (openness, conscientiousness, extroversion, agreeableness, and neuroticism) on their annual compensation. We hand-collect the tweets of S&P 1500 CEOs and use IBM's Watson Personality Insights to measure their personalities. CEOs with high ratings of agreeableness and conscientiousness get more compensation. We further find that the firms with these CEOs outperform their peers due to better investment efficiency. Firms are willing to pay higher compensation for talent, especially for firms with better operations, located in states with higher labor unionization, or facing higher competition in the product market. Overall, CEO personality is a valid predictor of CEOs' compensation.

This field experiment investigates the causal impact of mothers’ perceptions of gender norms on their employment attitudes and labor-supply expectations. We provide mothers of young children in Germany with information about the prevailing gender norm regarding maternal employment in their city. At baseline, over 70% of mothers incorrectly perceive this gender norm as too conservative – the most pronounced misperception among the various gender norms we examine. Our randomized information treatment improves the accuracy of these perceptions, significantly reducing the share of mothers who perceive gender norms as overly conservative. The treatment also shifts mothers’ own labormarket attitudes in a more liberal direction. Leveraging the fact that we assessed attitudes in a prior survey, we show that specifically the shifted attitude is a strong predictor of mothers’ future labor-market participation. Consistently, treated mothers are more likely to plan an increase in their working hours, particularly those with existing support to facilitate their employment.

This paper provides a comprehensive analysis of the forecastability of the real price of natural gas in the United States at the monthly frequency considering a universe of models that differ in their complexity and economic content. Our key finding is that considerable reductions in mean-squared prediction error relative to a random walk benchmark can be achieved in real time for forecast horizons of up to two years. A particularly promising model is a six-variable Bayesian vector autoregressive model that includes the fundamental determinants of the supply and demand for natural gas. To capture real-time data constraints of these and other predictor variables, we assemble a rich database of historical vintages from multiple sources. We also compare our model-based forecasts to readily available model-free forecasts provided by experts and futures markets. Given that no single forecasting method dominates all others, we explore the usefulness of pooling forecasts and find that combining forecasts from individual models selected in real time based on their most recent performance delivers the most accurate forecasts.