The Determinants of the U.S. Consumer Sentiment: Linear and Nonlinear Models
Open Access
- 1 July 2020
- journal article
- research article
- Published by MDPI AG in International Journal of Financial Studies
- Vol. 8 (3) , 38
Abstract
We examined the determinants of the U.S. consumer sentiment by applying linear and nonlinear models. The data are monthly from 2009 to 2019, covering a large set of financial and nonfinancial variables related to the stock market, personal income, confidence, education, environment, sustainability, and innovation freedom. We show that more than 8.3% of the total of eigenvalues deviate from the Random Matrix Theory (RMT) and might contain pertinent information. Results from linear models show that variables related to the stock market, confidence, personal income, and unemployment explain the U.S. consumer sentiment. To capture nonlinearity, we applied the switching regime model and showed a switch towards a more positive sentiment regarding energy efficiency, unemployment rate, student loan, sustainability, and business confidence. We additionally applied the Gradient Descent Algorithm to compare the errors obtained in linear and nonlinear models, and the results imply a better model with a high predictive power.Keywords
This publication has 30 references indexed in Scilit:
- When does investor sentiment predict stock returns?Journal of Empirical Finance, 2012
- Absenteeism, Performance and Occupant Satisfaction with the Indoor Environment of Green Toronto SchoolsIndoor and Built Environment, 2011
- The power of bad: The negativity bias in Australian consumer sentiment announcements on stock returnsJournal of Banking & Finance, 2011
- Investor Sentiment, Trading Behavior and Informational Efficiency in Index Futures MarketsFinancial Review, 2008
- Investor Sentiment in the Stock MarketJournal of Economic Perspectives, 2007
- Investor Sentiment and the Cross‐Section of Stock ReturnsThe Journal of Finance, 2006
- Random matrix approach to cross correlations in financial dataPhysical Review E, 2002
- Investor Sentiment and Stock ReturnsCFA Magazine, 2000
- Noise Dressing of Financial Correlation MatricesPhysical Review Letters, 1999
- Efficient tests for normality, homoscedasticity and serial independence of regression residualsEconomics Letters, 1980