Information Design and Data Science

The rules, information structure, and types of economic behavior in online markets impact our ability to estimate and learn about participants' preferences and market outcomes. In turn, learning about preferences and behavioral biases allows us to improve the ways in which markets operate, reducing market frictions and otherwise improving market performance. This is the promise of empirical market design.

One research direction involves estimating preferences and outcomes in two-sided markets using structural models, as well as the design of experiments (the two-sidedness as well as the dynamic nature of these markets provide technical and practical challenges). Similar challenges arise in online markets, where machine learning is used by market participants, as well as being central to the algorithms that run these markets. At the same time, digital markets provide new opportunities for rapid iteration through large-scale, randomized experiments.

A second research direction considers markets with participants that follow behavioral models of decision making, perhaps with simple adaptive behaviors, though not much is known about inference in matching environments with externalities. Here too, it will be interesting to couple the ability to learn about preferences and about systematic deviations from optimal choices with questions about design, including whether through careful design we can enable better inference about participants and reach more desirable outcomes.

There is also emerging theoretical and experimental work on information design for markets and mechanisms—which participants should receive which information, and when—along with the design of the methods of information elicitation. Both information design and the methods of information elicitation are major tools in reducing marketplace frictions.

This workshop will emphasize research in these three directions. We anticipate participation by leading scholars from economics, computer science, and operations research, in the evolving areas of empirical market design, machine learning and inference, and information acquisition and information design aspects of mechanism design and matching market design.

As with the second workshop, we also plan for participation by experts from industry, to add to the discussions. The goal of this workshop is to learn about research frontiers and create discussion between the scholars in these areas, aimed at studying challenges, both in the present and in the future.