Modeling Misspecification: Incorporating Information Processing Biases into Economic Models

Differences in how individuals interpret the same information have proven to play an important role in societal decision making.

Yet, most economic models assume individuals not only interpret information correctly but fundamentally agree on how information should be interpreted.

In ModelMisspec, I will develop a unified framework that links two approaches to modeling information processing biases.

I plan to (i) develop a unified framework to model information processing biases, (ii) develop tools, both theoretical and empirical, that can be applied to the analysis of problems this framework, and (iii) bring this framework to bear on applied models to study to the role of biases in price dynamics, learning, reputation and other applications.The proposed project has theoretical, empirical and applied components.

In the theory part, I will develop a unified framework that bridges two common approaches to modeling biases. This approach provides a transparent and general framework for modeling information processing biases. I'll then use this framework to explore the implications of information processing biases for decision making.

In the second part, I study misspecification empirically. In the first empirical component, I will experimentally elicit a key part of this framework, the forecast. This object is an key piece of decision making, that has gone unstudied.

I also will study misspecification in the field, through a survey of Finnish drivers' beliefs about the penalty schedule for speeding.

This section will focus on both developing tools for studying misspecification empirically, and its implications for decision making.Finally, I plan to explore the economic applications of this framework, including timing games, social learning, reputation, markets, etc.