Can GDP, CPI, and Treasury Bonds Signal a Recession?

“I knew which shifts in the economic environment caused asset classes to move around, and I knew that those relationships had remained essentially the same for hundreds of years. There were only two big forces to worry about: growth and inflation.” -Bridgewater Founder Ray Dalio

During the 2008 Financial Crash, many economists were uncertain as to whether the United States was in a recession. It wasn’t until a year after the recession began that the NBER, a private group of top economists declared the US was in the midst of an economic recession. They dated the beginning of the recession as December 2007, the turning point of the economy, and the “true peak of economic expansion”. The end of the “Great Recession” was June 2009. In 2001, the recession lasted eight months between March through November. On July 17, 2003, roughly two years later the NBER declared the recession ended November 2001. An article written by the Wall Street Journal titled, It’s Official: U.S Economy Is in a Recession That Began in March, Key Committee Says, was written in November 2001 at the end of the recession. A little late if you ask me.

The NBER looks at four main data points to determine whether the economy is in a recession. In this article, I won’t get into the details of these lagging indicators but for the record, NBER states, “The committee gives relatively little weight to real GDP.”

It is often said we can’t predict the weather. This may be true but at the same time, it is possible to predict the seasons. This first chart depicts the seasonality of the United States economy and instead of using seasons I named these four quadrants ‘Regimes’. Unlike the seasons, these Regimes never need to happen in a particular order. The U.S. Federal Reserve will take predictable actions depending on GDP (a measure of growth) and CPI (a measure of inflation) and whether those indicators are accelerating or decelerating.

Data from St. Louis Federal Reserve Website FRED

In the rate of change terms, we can see that GDP and CPI tend to move throughout the different quadrants based on the FED’s monetary policies. It should also be noted that GDP is not a lagging indicator. It is a coincident indicator meaning that it represents ‘real-time’ facts and the present state of the economy. Regime 4, the current regime, is characterized in economics as a period of deflation. Regime 3, is a period of stagflation. The 1st and 2nd Regime are categorized by growth and the expansion of the economy.

It was a long and winding road to settle on the 10 year Treasury Bond. Originally, I was looking for correlations between the VIX index, the S&P 500, and the 2 year Treasury bond. While I saw correlations by doing PC Analysis and basic correlation tests, the results never preceded a recession. Instead, they were reactionary post-recession. I needed something that was more forward-looking. So instead of a 2-year bond, I turned my sights to using a 10-year bond.

Using the same colors categorizing Regimes 3 and 4 we will apply these quarters to the second graph during periods of the most recent U.S. recessions.

Data from St. Louis Federal Reserve Website FRED

We can see in the above chart that the three most recent recessions have taken place during Regimes 3 and 4. Concomitantly, the 10 year Treasury bond saw a steep fall from its local quarterly highs and possibly giving confirmation of a looming recession. I would like to point out [according to the above chart] that periods of recession are only acknowledged after many months or a few quarters have passed.

To visualize the forward-looking 10-year bond and the short term 3-year bond I calculated the spread of the two bonds’ yields.

Data from St. Louis Federal Reserve Website FRED

Notice how the calculated spread touches or comes extremely close to touching the zero line anywhere between seven and three quarters before a recession. One can also see the divergence between the price of the 10-year and the spread preceding the recession. Three out of three times this has happened before a recession.

The results of this study are not definitive. Only time will tell if the GDP, CPI, and Treasury market will continue to predict a recession. However, the combination of macroeconomics and quantitative data does tell a story in hindsight of outcomes but by that time it is usually too late and the damage has already taken its toll.

For more information regarding past recessions visit www.nber.org.