What's the bond yield curve?

If you even glance at the business news, I am sure that by now you must have heard about the ‘Inverted Yield Curve’. I saw it covered in a variety of newspapers and websites. Most of the coverage was headline grabbing: “bond yield curve has inversed so there is a recession coming”. I realized that the challenge with concepts like these is that if there is not a fair understanding then it can lead to people jumping to wrong conclusions or panic. While I am by no means an expert, I thought it could be useful to write a quick guide to the bond yield curve and what it means and some cautionary notes on its interpretation.

First, it’s very critical to understand that a bond yield and implications of its inversion are applicable ONLY to the US treasuries/bonds and US economy. There have been times, in UK for example, where bond yields have been inverted and it was not followed by a recession. Second, it’s also important to understand that the inversion does not point to timing of the recession – in some cases, recession has followed inversion after 7 months and in some cases, recession came in 23 months or almost 2 years later. 

Also, the explanation is simplified to make it easier to understand. The mechanics of bond markets in US are extremely complex and have hundreds of variables influencing them. Finally, US government issues bonds ranging in duration from 3 months to 30 years. Again, for sake of simplicity, I am only going to talk about bonds ranging from 3 months to 10 years. These bond duration's are most used when looking at inversion and the price complexity gets significantly higher for even longer duration bonds.

OK, now that we have these out of the way, let’s understand what the yield curve is and what is its inversion. I am assuming that you already know what a bond is!!

First, let’s understand what a bond yield is – its often confused with interest on the bond, which it is not. Let’s assume that US government borrows money by issuing a bond with a duration for say, 1 year. This bond pays, say a 2% interest rate and has a face value of $100 (i.e. interest earned at end of the year is $2). After the US government issues a bond, it starts trading (like a stock) on the bond market. Now let’s assume that the economy is doing really well, which means that investors have other, higher return places to invest – like equity markets, real estate, and so on. As a result, the demand of these bonds falls and so traded price comes down to say, $95. The absolute amount earned on this bond remains at $2 – but now the $2 can be earned by investing only $95 instead of $100. Now the return on this bond is no longer 2% ($2 on $100), it’s 2.1% ($2/$95)- the 2.1% is the yield!  

The bond yield curve is simply a chart with duration of bonds on X axis and yields on Y axis. over a period. As you can imagine, the price of bonds, and hence yields, changes all the time - and so does the shape of this curve.

Usually, the chart or curve looks like the below (Yield curve inAugust 2017):

Something like the above is the ‘normal’ shape of the chart. The bonds with shortest duration have the least interest rate risk. For example, let’s say you own a 1 year bond that pays 2% interest. Now let’s say the stock market starts to boom and you feel confident that you can make much more than 2% on your money. Now you have an opportunity cost here – if you stay invested, you are making just 2% on your money while others may make more. However, since it’s only a 1 year bond, you will miss out on these other opportunities only for a short time – a year at most. This risk is even lower if you own 3 month or 6 month bonds. In addition, it’s highly unlikely that any material changes will happen in a short period in an economy as large and complex as the US. Because of this low risk, demand for these bonds is high and so is their price. That in turn means they have the lowest yield. Remember this rule: high price leads to low yields and low price leads to high yields (refer example above if this is confusing).

In this scenario, longer duration bonds have lower demand because you are exposed to this 2% risk for a long time. While other investments may be generating 5% returns you will proud owner of a bond that will earn only 2% for 10 years.

Do note that in both type of bonds, you can sell them in the market. However, most investors think alike. So a bunch of investors may start selling these bonds at the same time to reinvest in higher return assets. In this case, if you sell your bonds you may end up selling at a lower price that you bought them. Also, if everyone is selling their 1 year and 10 year bonds, prices for both will fall – but since interest rate risk is still relatively higher on 10 year bonds- those will fall more.

When good times are expected, yields for short term bonds are lowest (high demand and price) while yields for long term bonds are highest (lowest demand and price). This is why the slope of the curve is upwards from left to right. In the August 2017 chart above, you can see that 10 year bonds had a yield of 2.1% while the 1 year bond had a yield of only 1.2%.

Now let’s look at how it can get inverted – i.e. slope downwards from left to right.

Imagine that the US government issues a 10 year bond with an interest of 2%. Now this time let’s assume that investors think that US economy is going to enter a recession. Given the US bonds are considered risk free (US, rarely, has never defaulted on a single bond-ever!), a recession will mean that investors may have not have safe other alternate avenues to invest. As a result, investors may rush to buy US bonds leading to high demand and bonds becoming expensive. Price of 10 year bonds could increase from $100 to $110. The yield of these bonds then will do down from 2% ($2/$100) to 1.81% ($2/$110). But remember, investors expect a recession – so even a guaranteed 1.81% return is better than losing your money.

At the same time, since outlook for future is not too bright and investors want to safely lock in their money for long term, demand for and price of 1 year and other short term bonds will go down as well. From the example above, let’s assume that the 1 year bond now trades only at $93. The yield from this bond goes from 2% ($2/$100) to 2.15% ($2/$93) – so the yield goes up. So, with expectation of a recession, the right hand side of the curve starts to slope downward, and the left side of the curve starts of upwards. And that’s the inversion of the curve.

Below is a comparison of bond yield curves from Aug 2017 and Aug 2019. You can how the shape has 'inversed' in these two years. Yield for a 1 year bond went up from 1.2% in Aug 2017 to 1.77% in Aug 2019, an increase of over 45%. But the yield of 10 year bond fell from 2.1% to 1.49% - a fall of almost 30%.

Now let’s understand the implications. The inversion as a predictor of recession came out of a paper by Campbell Harvey, a Duke University professor, in 1986. In this paper, he proposed that an continuous inversion for a calendar quarter predicts a recession (this has already happened). This prediction model has actually held true for last 60 years and 7 recessions! Not just that but Campbell Harvey also accurately forecasted a lack of recession after the infamous Black Monday stock market crash in 1987. After the crash of stock markets across the world, most economists and pundits were predicting a recession. Campbell however saw a lack of inversion in the yield curve and forecasted that there will be no recession – and there wasn’t!!

While the accuracy is scary- there are a few things to be cautious about. The US Federal Reserve knows that a potentially accurate forecasting model of recession exists– so they will pay attention to it (unless they think it’s a complete coincidence). If they do pay attention to it – they could take proactive steps to head off the recession. Also, the model doesn’t say anything about timing, severity and length of the recession – it could be possibly be a short recession that lasts a couple of quarters or it could be a deep prolonged recession. It could also start in the next month or may not happen for another 2 years. It would be also be a mistake to use the predictor as a standalone or only indicator. After all, market structure, behavior, scale, quantitative easing – a lot of things have changed in the bond market over the past decade, so it is possible that the model just isn't valid anymore. And finally, there is a tiny possibility that it was all a coincidence.

Only time can tell if it is still valid – I truly hope it is not!