Knowledge that Transforms

Abstract We propose a new measure of underlying inflation that informs, in real time, about asymmetric risks on the outlook of inflationary pressures. The asymmetries are generated through nonlinearities induced by economic activity. The new indicator is based on a multivariate regime-switching framework jointly estimated on disaggregated sub-components of the euro area HICP and has several additional advantages. First, it is able to swiftly infer abrupt changes in underlying inflation. Second, it helps to timely track turning points in underlying inflation. Third, the proposed indicator also has a satisfactory performance with respect to various criteria relevant for inflation monitoring.

Abstract Do firms value court enforceability of their workers' noncompete agreements (NCAs)? We leverage a 2020 Washington law that made NCAs unenforceable for workers earning less than $100k per year. If firms value the ability to enforce NCAs in court, then they should give just-below threshold workers raises to reach the threshold, resulting in excess mass just above the threshold. Using administrative data, we find no evidence of bunching, even where efficiency arguments are most plausible. A survey of Washington employment attorneys suggests little bunching because firms rarely need to enforce NCAs and because firms can use other, less restrictive alternatives.

Abstract We study how peer beliefs shape individual attitudes toward maternal labor supply using hypothetical scenarios that elicit recommendations on the labor supply choices of a mother with a young child and an information treatment embedded within geographically representative surveys of the US population. Across scenarios, we find that individuals are systematically misinformed about the extent of gender conservativeness of the people around them. Exposure to information on peer beliefs leads to a shift in recommendations, driven largely by information-based belief updating. The information treatment also increases (intended and actual) donations to a non-profit organization advocating for women in the workplace.

Abstract International transactions are costly because they require investments in logistics, contracts, and the acquisition of local institutional knowledge. We posit that a portion of the fixed cost of entering a specific export market can be used toward covering the cost of acquiring imported inputs from that same market, and vice versa. Using dis-aggregated transactions data for Chinese firms from 2000 to 2015, we document firm-level trading patterns suggesting such bilateral economies of scope. Through a structural model, we estimate that the simultaneous export and import in a given country reduce export and import fixed costs by around 42 and 35 percent, respectively.

Abstract We evaluate the nationwide impact of the Washington State Attorney General's 2018-20 enforcement campaign against no-poach clauses in franchising contracts, which prohibited worker movement across locations within a chain. Implementing a staggered difference-in-differences research design using Burning Glass Technologies job vacancies and Glassdoor salary reports from numerous industries, we estimate a 6% increase in posted annual earnings from the job vacancy data and a 4% increase in worker-reported earnings.

Abstract We study how a market uses temporary workers to accommodate extraordinary demand shocks. When COVID-19 surges, hospitals need additional nurses—especially in specialties central to COVID-19 care. By comparing markets for COVID-relevant and other specialties, we show that the market for travel nurses expands dramatically and estimate travel nurse labor supply across space. Supply is quite elastic, as workers can choose to travel where they are needed. Workers travel longer distances to temporary jobs when payment increases, suggesting that an integrated national market facilitates reallocation when demand spikes. But when national cases peak, travel distance is less responsive to local demand.

Abstract We study communication frictions within multinationals (MNCs), hypothesizing that language barriers reduce management knowledge transfers within the organization. A distinct feature of such MNCs is a three-tier hierarchy: foreign managers (FMs) supervise domestic managers (DMs) who supervise production workers. Tailored surveys from our setting – MNCs in Myanmar – reveal that language barriers impede interactions between FMs and DMs. A first experimental protocol offers DMs free English courses and confirms that lowering communications costs increases their interactions with FMs. A second experimental protocol that asks human-resource managers at domestic firms to rate hypothetical resumes reveals that multinational experience and, specifically, DM-FM interactions are valued in the domestic labor market. Together, these results suggest that reducing language barriers can improve transfers of management knowledge, an interpretation supported by improvements in soft skills among treatment DMs in the first experiment. A model in which communication within MNCs is non-contractible – a realistic feature of workplace life – reveals that the experimental results are consistent with underinvestment in language training and provide a rationale for policy intervention.

Abstract How substantial are the economic benefits from democratic regime change? We argue that democratisation is often not a discrete event but a two-stage process: autocracies enter into ‘episodes’ of political liberalisation which eventually culminate in regime change or not. To account for this chronology and the implicit counterfactual groups, we introduce a repeated-treatment difference-in-difference implementation capturing non-parallel trends and selection into treatment. We find that modelling regime change in two stages rather than a single event yields stronger long-run growth effects. Among democratizers, experiencing repeated episodes without regime change reduces growth in democracy whereas length of episode does not.

Abstract In nonlinear panel data models, fixed-effects methods are often criticized because they cannot identify average marginal effects (AMEs) in short panels. In contrast with that criticism, we prove the point identification of different AMEs, including causal effects of changes in the lagged dependent variable or the last choice's duration, in a panel dynamic logit model for T as small as three. Our proofs are constructive and provide simple closed-form expressions for the AMEs in terms of probabilities of choice histories. We illustrate our results using Monte Carlo experiments and with an empirical application of a dynamic model of consumer brand choice.

Abstract Exponential regressions are frequently used when outcomes are non-negative. They are attractive because they are easy to interpret and to estimate, using pseudo maximum likelihood (PML). However, the validity of these methods depends on the correct specification of the conditional expectation, and little is known regarding their properties when the conditional expectation is misspecified. We show that PML estimators of misspecified exponential models provide optimal approximations to the conditional expectation, in a weighted mean squared error sense, and we give conditions under which their Poisson PML estimator identifies average marginal effects.