Novel research on expectations formation

A global study of consumer economic expectations

Francesco D’Acunto and Michael Weber explore how consumers form macroeconomic expectations and the role of information sources in this process. The study uses a survey conducted in April 2023, covering 47,000 consumers from 47 countries, representing 90% of global GDP. The survey was designed to be homogeneous across countries to ensure comparability of subjective macroeconomic expectations and included questions on demographic characteristics, financial status, macroeconomic perceptions and expectations (e.g., inflation, interest rates, house prices), as well as sources of economic information. It also assessed respondents’ trust in economic institutions. Finally, the authors distinguish between local economic environments (e.g., utility bills, shopping experiences) and aggregate sources (e.g., government reports, official statistics) as bases for forming macroeconomic expectations.

Main findings:

• Bias in Expectations: Consumers’ macroeconomic expectations are systematically biased due to reliance on local economic signals that are not representative of aggregate economic variables. This leads to inaccurate expectations about inflation, interest rates, and house prices.
• Heterogeneity Across Countries: There is significant variation in how consumers seek economic information across countries. In countries with higher historical inflation, consumers tend to seek more economic information but still form inaccurate expectations because they rely on local signals.
• Role of Trust: Trust in governments and central banks influences the choice of information sources. Higher trust in these institutions leads consumers to rely more on aggregate signals, resulting in smaller forecast errors.
• Implications for Models: The findings suggest that existing rational-inattention models need adaptation to account for systematic biases in consumer expectations. Trust in information sources should be considered a key factor influencing the accuracy of macroeconomic beliefs.

Information and Macroeconomic Expectations: Global Evidence
Authors: Francesco D’Acunto, Michael Weber
From: Georgetown University, University of Chicago

Cognitive constraints and macroeconomic expectations

Yeji Sung, in this paper, introduces a novel model of expectations formation that incorporates cognitive constraints in processing both external and internal information.

I introduce a new model where imperfect attention and memory hinder the seamless utilization of both types of information. This framework introduces a new mechanism: the inability to seamlessly integrate past knowledge into new forecasts. Consequently, prior beliefs play a diminished role in resolving economic uncertainty, leading forecasters to overemphasize recent news.

The key components of the methodology include:

Cognitive Process Model: Sung proposes a dual-system cognitive process involving attention and memory. The attention system focuses on processing external information while the memory system handles internal information (accumulated knowledge).
Mathematical Framework: Sung uses linear-Gaussian filters to model mental representations of information and employs Shannon mutual information to quantify representation accuracy. She introduces constraints on the precision of mental representations

Key findings:

• Horizon-Dependent Forecast Sensitivity: Short-term forecasts tend to underreact to new information while long-term forecasts are prone to overreacting compared to frictionless models.
• Estimation of Information Frictions: Estimates of attention constraints are approximately 50% larger than in models assuming perfect memory and indicate that conventional models may underestimate the severity of information frictions.
• Long-Run Expectations: Sung’s model predicts fluctuations in long-run expectations even with a fixed central bank inflation target and attributes this to memory frictions hindering accurate learning of long-run economic states.

The paper concludes that cognitive noise alone can explain various forecast biases observed in survey data, offering a parsimonious framework for understanding expectation formation in macroeconomics.
Macroeconomic Expectations and Cognitive Noise
Author: Yeji Sung
From: Federal Reserve Bank of San Francisco