E-mail: aespin@umass.edu
Research
My research interests lie in the evaluation of policy interventions and their impact on society, with a particular emphasis on the ways in which technology can shape different aspects of our lives. My dissertation focuses on various aspects of how information and telecommunications technologies (ICTs) are transforming our society, reflecting my passion for this topic. To examine the complex relationships between policy, technology, and society, I use a diverse range of econometric techniques in my research, including reduced-form approaches like difference-in-differences and structural methods such as discrete choice demand models and counterfactual simulations. Moreover, I rely on machine learning techniques to uncover hidden patterns and insights that may be missed using traditional statistical analysis alone. My work seeks to offer new perspectives on how policy and technology interact to shape society, while also identifying opportunities for leveraging policy to promote positive change.
Published:
The internet plays a vital role in everyday life across the world. The US, however, has seen a slowdown in household broadband adoption since 2010, creating a gap between connected and unconnected households usually referred to as the “digital divide.” While prior studies have documented how the digital divide is related to income, demographics, and geographic location, this paper takes a different approach and focuses on the mechanisms that could help bridge this gap. To this end, we use a two-stage approach. First, we construct a comprehensive and detailed dataset on household internet usage and prices to estimate broadband demand. Second, we employ the estimated income-dependent demand elasticities to assess multiple counterfactuals aimed at evaluating a number of public policy initiatives designed to reduce the digital divide. Central to our analysis are policies recently approved in the 2021 Biden Infrastructure Act. We contrast the effectiveness of the policies on three metrics: a) policy costs, b) reduction of the digital divide, and c) increases in consumer surplus. We find that affordability policies (i.e., subsidies) can have a larger impact on decreasing the gap vis-à-vis infrastructure deployment policies (i.e., increased coverage or greater bandwidth). We discuss how income-varying subsidies can be particularly effective at reducing disparities in broadband access across the income distribution.
Download paperNetwork neutrality mandates have been made out either as necessary to ensure a level playing field in online markets or, alternatively, as overly restrictive regulation preventing innovation and investment. However, there is little empirical research on the consequences of data throttling, which becomes legal without network neutrality regulations. We combine throughput levels measured for mobile ISPs in the United States with usage data to explore how sensitive users are to such practices. We find no evidence that users change their behavior when faced with throttled data rates.
Download paperWe evaluate the impact of the COVID-19 pandemic on the volume and quality of firms' daily usage of remote (video) meeting technologies. While per-firm daily meeting volume (minutes, number of meetings, and total participants) increase significantly (between 15% and 48%), the average meeting is more crowded (+15%), shorter (-30%, or 10 minutes) and of significantly poorer (video/audio) quality (-59%). Firms in the service sector experience the most notable increases in volume usage, while effects on the duration, size and quality of meetings is experienced by firms in all industries.
Working papers:
We use barcode-level data in the US between 2012 and 2020 to document the evolution of trans fat content in manufactured food products before and after the FDA’s 2015 determination that trans fats are not safe for human consumption (effectively a “trans-fat ban”). First, we document how the number of products containing trans-fat has declined over time. Second, using detailed information on household food purchases in retail stores, we quantify the corresponding reduction in total trans-fat purchases. We find that although the number of trans fat products on the market has declined over time, products with a positive trans-fat content continued to exist until 2021. We find a larger presence of trans fat products in small-scale retailers (dollar and convenience stores) than in larger outlets (mass merchandisers and warehouse clubs). Similarly, trans fat products are relatively more prevalent in private label products and in products with a local presence. While trans-fat purchases have declined over time, lower income households purchase product baskets with significantly larger amounts of trans fat. The findings can be used to guide more effective and efficient enforcement efforts.
Download paper