T-Bond Futures: Back to the 1990s? (Summary)

Investors take long or short positions using US Treasury bond futures to hedge long-term interest rate risk. This study assesses the risk that the T-bond futures market may suffer distortions in coming years, to the detriment of users of T-bond futures and the Treasuries markets.

The potential consequences of this risk are significant: Treasuries are used as benchmarks to price other securities, as collateral, as underlying assets for fixed-income derivatives, and to implement monetary policy. In addition, a shortage of supply and market manipulation of Treasuries can severely affect pricing and liquidity in the cash, repo, and futures markets, as well as in the T-bond basis (simultaneously trading bonds and the related bond futures).

The looming problem relates to the underlying assets used by investors with short positions to settle their T-bond contracts. Because no single bond has all the characteristics needed, repayment is made through a basket of bonds with varying coupons and maturities. Some bonds in this basket are cheaper to deliver than others—depending on their coupon, term to maturity, and price. Because bonds are constantly maturing, the cheapest bond changes regularly.

Unusually, a single bond will be the cheapest available for this basket during the next five years. In combination with other market factors, this scenario could lead to potential liquidity issues, price manipulation, and a squeeze on short sellers.

The authors suggest that current conditions could trigger these negative consequences but that the potentially damaging effects could be mitigated.

The authors start with the premise that a single T-bond (the 4½% of 2036) currently has the shortest maturity in the deliverable basket for the next five years.

They find parallels with the 1994–99 period, when the 11¼% issuance of 2015 was the cheapest in the basket for a full five years. This period is known as the “long dry spell of the 11¼%.” It occurred because of the elimination of the callable feature of T-bonds after 1985, which opened up a maturity gap in the deliverable basket, the effects of which were experienced only a decade later.

The authors make the case that this “long dry spell” would not have led to such negative consequences were it not for three additional (and concurrent) conditions:

1. Interest rates below the reference coupon rate. The yield curve was substantially lower than the coupon rate from 1994 through 1999, so bonds with the lowest durations tended to be the cheapest to deliver.

2. Fifteen years of falling long-term yields. All the T-bonds issued from August 1985 through August 1999 had coupon rates lower than the 11¼% T-bond of February 1985. So, the security with the shortest maturity also had the highest coupon.

3. A rising discrepancy between the long-term yield and the reference coupon rate. In the second half of the 1990s, interest rates fell, with yields on long-term T-bonds falling faster than those on short-term T-bonds.

As well as testing for these same conditions today, the authors assess what steps might be taken to mitigate the adverse consequences of history repeating itself.

The authors believe that two decades after the “long dry spell of the 11¼%,” similar conditions have arisen again. A five-year maturity gap has opened up in T-bond delivery baskets, between the 5 3/8% issue of February 2031 and the 4½% issue of February 2036. So, the latter issue will be entrenched as the cheapest bond. Similar to during the 1980s, this gap was caused by no issuance of 30-year T-bonds for a five-year period between 2001 and 2006.

In addition, the same three accompanying conditions that existed in 1994 exist today:

1. Interest rates below the reference coupon rate. Interest rates have stayed well below the 6% rate of the T-bond futures contract for almost seven years, and markets expect interest rates to rise only gradually.

2. Falling long-term yields. In the wake of the financial crisis, the Fed’s short-term interest rate fell close to zero, while long-term T-bond yields also hit historic lows.

3. A rising discrepancy between the long-term yield and the reference coupon rate. This discrepancy hit an all-time high of 4.14% in June 2016, well above the 2.94% of December 1998.

Given that the upcoming episode is predictable, the authors say it can be mitigated by two possible fixes. The first is to extend the terms to maturity of deliverable bonds so the five-year-gap issue would be postponed to a later date. The second is to lower the reference coupon rate to 3%, which better reflects prevailing long-term rates, reducing the potential for a market squeeze.

Investors should be aware of the potential negative consequences of another “long dry spell.”

One consequence is reduced effectiveness of T-bond futures hedging. From 1994 through 1999, the effective maturity of the contract diminished in tandem with the remaining term of the 11¼% of February 2015 issue.

A second consequence is that there could be a constricted market for the T-bond basis, a market central to many hedging and trading strategies, given that trading the basis relies on changes in the cheapest bond. No such changes took place during the 1994–99 period because of the entrenchment of the 11¼% issue.

Third, there could be short squeezes. These typically involve buying both the T-bond future and the cheapest bond in its delivery basket, leading to a reduction in the supply of the cheapest bond and higher costs for short traders seeking to unwind their futures positions. Short squeezes usually create pricing distortions, affecting all T-bond sectors, including the cash, repo, and futures markets.