The Empirical Features and Economic Implications of Biased Expectations

This podcast episode features Gemma, an assistant professor at the University of Chicago Booth School of Business, who discusses her research on expectations formation and its empirical features. She highlights the limitations of the traditional rational expectations benchmark, which assumes people process information frictionlessly and form unbiased beliefs. Instead, her work, rooted in behavioral economics, focuses on understanding how real-world expectations deviate from this ideal, utilizing extensive datasets from financial markets, corporate investments, and macroeconomic outcomes, as well as controlled experiments.

A central finding of Gemma's research is the pervasive tendency for people to "over-extrapolate" past trends and shocks, meaning they expect recent patterns to continue more than they actually do. This bias is particularly pronounced when the underlying economic processes exhibit low persistence (e.g., corporate earnings, stock market returns) and in environments characterized by high volatility, where the potential for significant forecasting errors is greater. This over-extrapolation is observed across various domains, from firms' expectations of future earnings to investors' beliefs about stock returns, often leading to predictions opposite to what rational models would suggest during market booms.

The implications of these systematic biases are significant, contributing to phenomena such as fluctuations in asset prices, the formation of market booms and bubbles, and cycles in credit, investment, and house prices. From a policy perspective, recognizing these inherent limitations in forecasting the future, both for market participants and policymakers, underscores the importance of building robust financial structures. The concept of stress tests for systematically important financial institutions, developed post-crisis, is presented as a crucial development to ensure resilience against unexpected large negative shocks, regardless of prevailing beliefs.

Beyond economic applications, the discussion delves into the deeper foundations of biased belief formation, exploring interdisciplinary connections. Gemma suggests that tendencies like over-extrapolation might be linked to cognitive mechanisms such as associative memory, where easily recalled associations (e.g., past high returns leading to expectations of future high returns) can shape beliefs. Furthermore, she connects this to the perpetuation of stereotypes by social groups, where exaggerating certain features or abilities can lead to biased societal beliefs, highlighting the exciting potential for learning from psychology and neurobiology to understand these foundational questions in belief formation across social sciences.

I think one of the important lessons behavior economics teaches us is to recognize our limitations and to recognize that there are many imperfections and for example belief formation and to be modest about the capacity to forecast the future

for the past few decades the benchmark in research is generally rational expectations which postulate that people can process information friction sleeve they know the right set of information to use and they can form rational and biased beliefs about the future

one of the features that does stand out is people tend to over extrapolate and in particular that means that people think past trends where sharks are likely to continue more than they actually do

these features are particularly pronounced when the actual process has low persistence

biases and expectations tend to have a bigger impact when there is more volatility

rational models generally predict that there are times of stock market booms then the expected returns going forward which should be low because current prices are high whereas when you look at beliefs there they are unbold from investor surveys when actually ask investors to stock market boom so would say they expect stock returns to go up

recognizing the limitations for both the markets and policymakers to forecast the future is to build financial structures that are robust to unexpected large declines in asset prices or two recessions

there is some potential that exaggerating these features in the characteristics of certain groups and exaggerating the possibility of future stock returns being high maybe there is something that is connected for example it's much easier for us to think about things to come to mind more easily

some of these tendencies can be used by social groups for certain purposes strategically to perpetuate some norms in the context of gender

those are very exciting interdisciplinary areas where we can learn a lot from psychologists from neurobiologists and ultimately understand some of the very foundational questions in belief formation in financial markets and in other areas of social science and economics and finance