Do digital technologies reinforce managerial hierarchies, or instead make them less relevant? We propose that the answer to this question depends on the nature of the technology, specifically its relative impact on managers’ capacity to supervise and on subordinates’ need for supervision. Applying this framework to collaborative work management (CWM) technologies that facilitate real-time collaboration, communication, and task coordination, we predict that the adoption of such technologies should reduce managerial intensity and increase decentralization in organizations. To test this prediction, we use a difference-in-differences design on a novel dataset built from over 26 million job listings (Lightcast) and over 20 million social profiles (Revelio) matched to 3,017 US public firms in Compustat, which we track over the period 2010-2019. We find that over the observation window, CWM technology adopters show a 3% reduction in managerial intensity and a 5-7% increase in non-managerial skills linked to decentralization in their job postings in the years following adoption. The pattern of results is robust to a battery of validations, alternative measures, and specifications, and strongly supports the idea that these technologies enable collaboration and make organizations less hierarchical along the dimensions we studied.

To manage post-acquisition operating performance, acquirers must balance achieving synergies with minimizing disruption. Prior work emphasizes organization-level factors, such as knowledge similarity and cultural alignment, that help achieve this balance but are not easily modifiable in the short term. In this paper, I shift focus to individual-level factors, specifically employee social skills, that are more easily adjusted through hiring. I argue that increased social skills, such as communication, negotiation, and teamwork, mitigate disruption by enabling improved inter-unit collaboration between the acquirer and target units. I test this using a stacked difference-in-differences design applied to data from 26 million job postings between 2010-2020, covering 5,303 acquisitions, merged with firm-level Boardex and Compustat data. I find that a 25 percent increase in social skills in job postings correlates with a 3.1 percentage point improvement in operating performance post-acquisition close. Estimated improvement is higher – 11.5 points – for deals with a greater increase in distinct organizational units, a proxy for higher inter-unit collaboration needs, lending support to the proposed mechanism. I validate these findings using multiple measure operationalizations and empirical designs, as well as, examine acquirers, deals, and roles where the estimated relationships are strongest. These findings help advance research on post-acquisition performance by highlighting the role of structure-human capital dynamics in enabling organizational collaboration.

A widely held view is that Generative AI (GenAI) automates cognitive tasks, allowing employees to focus on interpersonal coordination. In this paper, we develop a framework showing that this is only one of several plausible trajectories. Analyzing job postings, we find preliminary evidence that a reverse trajectory may instead be unfolding among 596 early GenAI-adopting U.S. public firms. Since ChatGPT's launch, the demand for social skills within GenAI-adopting roles has declined by 4.3 percent, while the demand for cognitive skills has remained unchanged. This pattern suggests that GenAI may be modularizing work - reducing reliance on interpersonal coordination rather than amplifying it.

Remote work intensifies communication silos, which is a problem for organizations as it is rarely feasible to perfectly organize interdependencies within them. In this paper, I explore solutions and ask: what organizations can do to overcome silo intensification challenges? I leverage open-source software communities as the context–these are all-remote organizations where contributors carry out software development. By using email metadata and text from the Linux-kernel open-source project, I find that motivational differences exist among open-source contributors that are also correlated with their propensity to undertake distant interactions. Overall, the Linux-kernel organization exhibits the existence of timezone silos–with a 19.3 percent reduction in interactions with those in distant timezones versus those nearby. However, volunteers, unpaid for their work and motivated by recognition and knowledge, show a lower drop in distant interactions, by 12.7 percentage points, as compared to firm-sponsored contributors. Moreover, the estimated difference is even higher for volunteers with lower recognition and knowledge–consistent with the March (1991) idea that those with low accomplishments are more likely to explore, strengthening belief in the motivational mechanism. Broadly, these differences, observed in the behavior of volunteers and firm-sponsored contributors, help understand the characteristics of workers that make them more likely to span internal organizational boundaries.