Forecasting the Leading Indicator of a Recession: The 10-Year minus 3-Month Treasury Yield Spread

TL;DR

This paper compares econometric and machine learning models to forecast the yield spread, a key recession indicator, finding that simple ARIMA models perform as well as complex LSTM models.

Contribution

It introduces a comprehensive comparison of traditional econometric and modern machine learning models for yield spread forecasting, highlighting the effectiveness of simple models.

Findings

ARIMA outperforms VAR in accuracy

LSTM performs comparably to ARIMA

Yield spread shocks significantly impact forecasts

Abstract

In this research paper, I have applied various econometric time series and two machine learning models to forecast the daily data on the yield spread. First, I decomposed the yield curve into its principal components, then simulated various paths of the yield spread using the Vasicek model. After constructing univariate ARIMA models, and multivariate models such as ARIMAX, VAR, and Long Short Term Memory, I calibrated the root mean squared error to measure how far the results deviate from the current values. Through impulse response functions, I measured the impact of various shocks on the difference yield spread. The results indicate that the parsimonious univariate ARIMA model outperforms the richly parameterized VAR method, and the complex LSTM with multivariate data performs equally well as the simple ARIMA model.