ID: 66517
Authors:
Lucélia Viviane Vaz, Rodrigo Jardim Raad.
Source:
Revista Brasileira de Finanças, v. 20, n. 1, p. 1-23, January-March, 2022. 23 page(s).
Keyword:
Functional data analysis , Functional principal component analysis , Term structure of interest rate
Document type: Article (English)
Show Abstract
This paper analyzes Brazilian nominal yield curves based on a functional data analysis framework. Specifically, we use functional principal component analysis to describe sources of variability in yield curves and their related level, slope, and curvature. We also present a functional linear regression model to investigate macroeconomic determinants of the yield curves. We conclude that level shocks strongly explain variability in interest rate curves. Slope changes are the second-largest source of variability. The slope of the yield curve is negatively affected by the nominal exchange rate and Selic reference rate, and positively affected by Brazil’s risk and industrial capacity utilization. We also infer that the following explanatory variables: expected inflation, Selic reference rate, Brazil risk, and industrial capacity utilization, all have positive effects on the level of the yield curves. The variables Selic, Brazil risk, and the nominal exchange rate positively impact curvature. The yield curve is negatively impacted by industrial capacity utilization and expected inflation.
