As Romer and Weitzman emphasized in the 1990s, new ideas are often combinations of existing ideas, an insight absent from recent models. In Kortum's research around the same time, ideas are draws from a probability distribution, and Pareto distributions play a crucial role. Why are combinations missing, and do we really need such strong distributional assumptions to get exponential growth? This paper demonstrates that combinatorially growing draws from standard thin-tailed distributions lead to exponential growth; Pareto is not required. More generally, it presents a theorem linking the max extreme value to the number of draws and the shape of the upper tail for probability distributions.

By the end of 2019, U.S. output was 14% below the level predicted by its pre-2008 trend. To understand why, I develop and estimate a model of the United States with demographics, real and monetary shocks, and the occasionally binding zero lower bound on nominal rates. Demographic shocks generate slow-moving trends in interest rates, employment, and productivity. Demographics alone can explain about 40% of the gap between log output per capita and its linear trend by 2019. By lowering interest rates, demographic changes caused the zero lower bound to bind after the Great Recession, contributing to the slow recovery.

Extensive evidence suggests that managers strategically choose the complexity of their descriptive disclosures. However, their motives in doing so appear mixed, as complex disclosures are used to obfuscate in some cases and to provide information in others. Building on these observations, we first identify a novel stylized fact: disclosure complexity is non-monotonic in firm performance. We develop a model of disclosure complexity that incorporates the dual roles of complexity and can explain this stylized fact. In the model, a manager discloses to investors of heterogeneous sophistication and can adjust the complexity of the disclosure to either provide more precise information or to obfuscate. When the firm's investor base is largely unsophisticated, the manager issues a complex disclosure only upon observing negative news. In contrast, when the firm's investor base is more sophisticated, the manager issues a complex disclosure upon observing either highly positive or negative news. As a result, the market may react more positively to complex information releases than to simple releases, complex disclosures generate heightened return volatility, and firms with more inherently complex information are more likely to use their discretion to simplify their disclosures.

This study uses machine‐learning methods to predict next‐period change in profitability based on a model proposed by Penman and Zhang (2004, Working paper, Columbia University and University of California, Berkeley; “PZ”). We find that new machine‐learning methods predict out of sample substantially better than traditional regression methods and provide richer interpretations about the role and impact of different predictor variables through their nonlinear relationships and interaction effects. For example, our results contrast with previous research by showing that both components of the DuPont decomposition (change in profit margin and change in asset turnover) are informative of next‐period changes in profitability. Our results are robust across different performance metrics, alternative machine‐learning models, and software. Furthermore, an unconstrained machine‐learning model using a larger feature space could not significantly improve the performance of the PZ model. PZ variables alone accounted for most of the explanatory power of the unconstrained model, suggesting the PZ model is both well specified (in terms of feature selection) and robust in higher dimensional settings. With respect to the economic significance of this information, we find mixed results. The market appears to adjust its expectations more in line with the machine‐learning predictions relative to the PZ model but the portfolio returns are not significantly different.

We investigate how sizes of professional networks affect the probability of appointment to a supervisory board and whether the effect is gendered. Using an employer-employee data set of the Danish labor market, 1995–2011, we find larger networks to associate with a higher probability of becoming a first-time director. The effect is larger for men. One explanation is that men, compared with women, have more connections to larger and listed firms and to other males—attributes that increase the appointment probability. Women who have connections to incumbent directors before being appointed director have more labor market experience than other directors.

We examine the relationship between leverage (the assets-to-equity ratio) and the weighted-average cost of capital (WACC) for an unbalanced panel of U.S. bank holding companies (BHCs) over the period spanning 1996 through 2019. We assess the extent to which changes in capital requirements affect the WACCs of these institutions and find significant differences across three, mutually exclusive asset-size classes of BHCs. In particular, the WACCs of small and medium-sized BHCs are found to be more sensitive to changes in capital requirements than the WACCs of large BHCs. For a hypothetical doubling of Tier-1 capital, we find that the largest BHCs saw their WACCs increase by 42 basis points, while the WACCS of small and medium-sized BHCs are estimated to have increased by 62 and 76 basis points, respectively. We attribute differences between these results to differences in the strength and scope of government guarantees and in the composition of BHCs’ assets and liabilities as reported on their balance sheets.

In this paper, I develop a measure of the difference in the amount of information that investors expect a forthcoming disclosure to contain should it reveal good news versus bad news (the disclosure’s “asymmetry”). To do so, I first show that this asymmetry is linked to the skewness of returns that the disclosure creates. I then show that this skewness can be measured using a weighted change in option-implied return skewness leading up to the disclosure’s release. The measure’s ability to capture investors’ prior beliefs regarding asymmetry is advantageous when studying ex ante decisions including contracting and information acquisition choices. I implement it on a sample of large firms’ quarterly earnings announcements, finding evidence that investors anticipate cross-sectional but not time-series variation in earnings’ asymmetry.

Children’s indirect exposure to the justice system through biological parents or coresident adults is both a marker of their own vulnerability and a measure of the justice system’s expansive reach in society. Estimating the size of this population for the United States has historically been hampered by inadequate data resources, including the inability to observe nonincarceration events, follow children throughout their childhood, and measure adult nonbiological parent cohabitants. To overcome these challenges, we leverage billions of restricted administrative and survey records linked with Criminal Justice Administrative Records System data and find substantially larger exposure rates than previously reported: prison, 9% of children born between 1999–2005; felony conviction, 18%; and any criminal charge, 39%. Charge exposure rates exceed 60% for Black, American Indian, and low-income children. While broader definitions reach a more expansive population, strong and consistently negative correlations with childhood well-being suggest that these remain valuable predictors of vulnerability. Finally, we document substantial geographic variation in exposure, which we leverage in a movers design to estimate the effect of living in a high-exposure county during childhood. We find that children moving into high-exposure counties are more likely to experience postmove exposure events and exhibit significantly worse outcomes by age 26 on multiple dimensions (earnings, criminal activity, teen parenthood, mortality); effects are strongest for those who moved at earlier ages.

This article provides a comprehensive history of default risk for newly originated home mortgages in the USA over the past quarter century. The loan-level source data include the entire guarantee book for Fannie Mae and Freddie Mac. We track many loan characteristics and produce a summary measure of risk. Among our many results, we show that mortgage risk had already risen in the 1990s, planting seeds of the financial crisis well before the actual event. Our results also cast doubt on explanations of the crisis that focus on borrowers with low credit scores. The aggregate series are available for download at https://www.fhfa.gov/papers/wp1902.aspx.