Research

You can find an overview of key themes in my current and future research here: Research Statement, Oberg

Should Consumer Subsidies Be More Flexible? A Structural Analysis and Case Study of the Food Stamps Program. (Job Market Paper) [Link to latest draft]

Abstract: Consumer subsidies reduce the price of goods and services to the consumer. For example, many government funded welfare programs provide its recipients in-kind transfers which are subject to different spending restrictions. In this paper, we study how consumers respond to subsidy designs with different spending restrictions, with implications to optimal design of welfare subsidies. In the context of SNAP (commonly known as the Food Stamps Program), we analyze the effect of food stamp benefits from the perspective of recipients (who care about overall consumer welfare) and the policymaker (who prefers that funds are used to buy food). To simulate consumer behavior under different subsidies, we develop a structural model of consumer demand which integrates consumer decisions for brands, categories, and stores. Our main finding is that expanding food stamp benefits to include household goods would be preferred by both benefit recipients and the policymaker. The mechanism driving this result is that flexible benefits give access to a wider selection of items which provides greater incentives to visit stores. In addition, we quantify trade-offs between different benefit designs and study the effect of banning benefit use on sweetened soda. Finally, our model of brand, category and store choice makes a technical contribution that could be interesting beyond the application considered in this paper.

Keywords: structural modeling, supermarket demand, food policy.

Tackling External Validity in Marketing Decisions: The case of Media Planning in Online Display Advertising.

Abstract: In this paper, we propose two ideas: (i) a method that measures ad effectiveness leveraging quasi-experimental variation in auction outcome data, and (ii) a model of optimization that helps decision makers account for parameter variability when setting advertising strategy. The two methods ((i) and (ii)) are integrated to develop a decision aid for advertisers tasked with setting campaign strategy, with a focus on out-of-sample performance. In the application, we examine parameter variability in the context of campaign-level decision variables (e.g., choice of audience segments, ad visuals, and ad placement) and study implications for optimal firm spending across publishers when launching a new campaign with design choices (e.g., new choice of audience segments, new ad visuals, new ad placement) that are not present in observed data.

Keywords: attribution modeling, multichannel marketing, causal inference, selection bias, robust optimization.

Consumption and Generation of Reviews in Online Platform Markets — A Structural Model.

Abstract: In markets with many products – e.g., vacation rentals, consumer electronics, restaurants – consumers rarely have access to vertical quality measures (like hotel star ratings) and often rely on reviews written by other consumers to choose among products. Learning from reviews is complicated because review information is a mix of vertical quality signals and horizontal preferences of the respective reviewers who are often unknown to the customer reading reviews. As a result, review scores and content hide two types of quality signals, creating a challenging problem for the platform to order search alternatives and review content to match each customer’s tastes. In this paper, we develop a model that allows us to decompose review scores along two dimensions, corresponding to vertical and horizontal preferences. We use data from an online marketplace for vacation rentals and study how ordering search alternatives and review content based on two preference dimensions instead of one aggregated dimension affects search and choice, and the attendant implications for customer satisfaction and platform revenue.

Keywords: structural modeling, consumer reviews, platform design.