Following a 2018 ruling of the U.S. Supreme Court, 38 states have legalized sports gambling. We study how this policy has impacted consumer financial health using a large and comprehensive dataset on consumer financial outcomes. We use data from the University of California Consumer Credit Panel, containing credit rating agency data for a representative sample of roughly 7 million U.S. consumers. We exploit the staggered rollout of legal sports betting across U.S. states and evaluate two treatment effects: the presence of any legal sports betting in a state and the specific presence of online or mobile access to betting. Our main finding is that overall consumers' financial health is modestly deteriorating as the average credit score in states with legalized sports gambling decreases by roughly 0.7 points. When states introduce online sports gambling, declines in average credit scores are much larger (12.04 points). The decline in credit score is associated with changes in indicators of excessive debt. We find a significant increase in average bankruptcy rates, debt sent to collections, auto loan delinquencies, and credit card delinquencies. Together, these results indicate that the ease of access to sports gambling is harming consumer financial health by increasing their level of debt. 

Policymakers have been quick to legalize access to online sports gambling (OSG) platforms — 30 states and the District of Columbia now have some form of OSG. We assess the state fiscal benefits of this policy by examining how OSG impacts demand for state lotteries, an alternative gambling market, and a crucial source of funding for public programs. Leveraging the staggered adoption of OSG across states, we analyze changes in lottery demand using a county-level lottery sales dataset across 17 states plus DC and a separate transaction-level dataset covering 80% of all independent convenience stores in all 50 US states. We find consistent evidence that OSG significantly reduces state lottery demand, with sales declining by 5.4% to 6.2% in the 16 months following implementation. Because of reduced demand and the initial sports gambling tax policy, we find that our set of 17 states plus DC incurred a cumulative loss of -$112 million (95% CI = [-$12M, -$213M]) in state revenue following OSG implementation. Analysis of individual states shows that early adopters (New Jersey and Pennsylvania) experience significant revenue losses, while most other states break even. We present evidence that the reduced demand is likely due to lower lottery purchasing intensity, as consumers shift a larger share of their budgets toward OSG. For state policymakers, these results challenge the notion that OSG is a fiscal windfall opportunity. We conclude with a discussion of tax policy implications for state lawmakers.

Advertisers face an increasingly cluttered digital ecosystem where customer attention matters. Despite this, the extent to which digital advertising clutter influences ad viewership remains unclear. This paper presents three separate eye-tracking experiments (subjects: N = 99, N = 250, and N = 575) using over a dozen custom-built web pages to better understand how digital ad clutter affects consumers' viewing of digital ads. Across experiments, we show that increased clutter led consumers to allocate more attention to the overall advertising space and adjust how they distribute viewership across ad positions. A complementary survey revealed that clutter also reduced ad memory and recall for individual ads. We discuss the implications for advertisers and publishers regarding ad pricing and placement in crowded digital environments.

Natural language processing algorithms interpret search queries and support the real-time matching between advertisers and sponsored search ad auction opportunities on search engines. Improvements in interpretation capabilities can reshape how advertisers are ranked for ad space and what they pay. We investigate how improvements to query interpretation algorithms in sponsored search markets affect short-term cost-per-click (CPC) and the number of advertisers bidding for ad space (competition). We develop a conceptual framework that considers the types of information present within search queries. We identify two channels through which improved query interpretation can affect auction outcomes: topic understanding expands markets by enabling the platform to identify relevant advertisers with greater confidence, while context understanding creates dispersion in relevance scores that can lower prices for context-rich queries. This can lead to auctions with more bidders yet lower final clearing prices. We empirically test predictions from this framework using monthly data on the number of bidders and CPC for 12,000 queries, exploiting Google's 2019 rollout of Bidirectional Encoder Representations from Transformers (BERT) as a natural experiment. Using a year-over-year difference-in-differences strategy, we find that the number of bidders competing for ad space increases across queries, CPC rises for short, context-poor queries, and falls for longer, context-rich queries. The finding that CPC declines for long queries despite increased competition is inconsistent with alternative platform changes. Our results illustrate how improved interpretation algorithms can reshape prices and competition in the sponsored search market.

As consumers increasingly delegate purchasing decisions to AI agents, firms face a new question: how does consumer heterogeneity translate into agent-mediated demand? Human and model preferences need not coincide, preferences may differ across models, and consumers nonrandomly select into agent use. We study this problem through delegated preference transmission: the extent to which consumer-level preferences and behavioral heuristics are preserved when an AI agent makes choices on a consumer's behalf. We develop a framework separating two necessary conditions for transmission: identifiability of the consumer parameter in the prompt and model responsiveness to that signal. Using a conjoint experiment and an incentivized online bookstore adoption task, we study attribute preferences as well as left-digit bias (LDB), a behavioral heuristic in price perception. Attribute preferences are meaningfully reflected in prompts and transmitted to agents. In contrast, high-LDB consumers write shorter prompts with less price information, but current LLMs do not meaningfully adjust their LDB in response; instead, agents pool consumers towards model-specific priors. Consumers with higher LDB are also more likely to state intent to use agents; revealed-choice evidence in a follow-up bookstore experiment is directionally consistent but does not reach conventional significance. A stylized pricing simulation shows that these forces change optimal markups, increase the use of X.99 pricing, and make market composition central to pricing decisions. These findings imply that agentic markets do not eliminate consumer heterogeneity; they transform it across consumers, prompts, models, and adoption decisions.