Business cycles

: the predictive power of the yield curve for recessions in the G7 countries

Abstract

Economic recessions and Business Cycles have been studied over the years because of their serious financial impacts, which tends to spread all over the economy, causing unemployment, poverty, bankruptcies, etc. This Dissertation addresses the reliability of the yield curve as recession’s predictor nowadays and aims to test the forecasting power of the yield curve in countries beyond the United States. By employing Panel Data estimations, an empirical analysis of the ability of the yield curve and other macroeconomic variables to predict recessions is conducted, using indicators of economic activity across the G7 countries as a whole and individually, based on monthly data from 1995 to 2023. Empirical evidence shows that the predictive power for recessions is not uniquely a consequence of the yield curve as predictor, but also a result of the impact of the lags of the dependent variable itself as the current value of an economic activity indicator is deeply influenced by its recent past values. Furthermore, the third predictive equation presented for each predicted variable has the greatest predictive power for recessions, suggesting that the composite leading indicator is a better predictor of the economic activity than inflation and money supply together. Findings suggest that an efficient predictive power of the yield curve and lagged variables specially for Industrial production and Retail sales as dependent variables. Finally, robustness checks are conducted as the regression framework was restricted to a given regional area (America and Europe) and to specific countries, suggesting the stability of the estimations performed.

Date of Award	27 Jun 2024
Original language	English
Awarding Institution	Universidade Católica Portuguesa
Supervisor	Diptes Chandrakante Prabhudas Bhimjee (Supervisor)