financialengineerlab/yield-curve

A Python/Jupyter notebook project to understand the Yield Curve and its potential for forecasting a recession

YIELD CURVE

A Python/Jupyter notebook project to understand the Yield Curve and its potential for forecasting a recession. Yield curve rates between 1990 and present are from the U.S. Department of the Treasury

Getting Started

• Git clone the repository
• Create a virtual environment and run $ pip install -r requirements.txt
• Update the docker-compose.yml volumes to point to your local instance and cd to where the repo is stored
• Run the following:

  $ docker-compose build
  $ docker-compose up
  

You can access the project at http://localhost:8888/notebooks/yield%20curve.ipynb and you'll need to provide a token that appears in the docker logs. To bring down the project, run $ docker-compose down and to destroy the containers, run $ docker-compose kill.

Forecasting a Recession

A deeper dive into U.S. Department of the Treasury Yield Curve data and it's predictive capaiblities.

In an efficiently performing market, long-term bonds have higher bond yield rates than shorter-term bonds, T-notes, and T-bills as the market expects greater risk in investing in long-term bonds (a lot can happen in 30 years). However, when the yield curve inverts, the bond yield rates for shorter-term bonds are higher than long-term bond yield rates. An Inverted Yield Curve is used as one predictor of a recession as it captures the nervousness of investors about the near term market outlook.

In my analysis, an Inverted Yield Curve occurs when the ratio of long-term bond rates (i.e. 30 years, 10 years) versus short-term bonds (6 months, 1 year, 3 years, etc.) is between 0 and 1. The yield curve last inverted between 2006 and 2007.

Yield Curve Ratios for Years 1990 to Present

Yield Curve Ratios Demonstrating Inversion from 2006 to 2007

Yield Curve Ratios for 2019

The ratio of 10 year bonds/6 month bonds and 10 year/1 year bonds inverted in May 2019