THE STORY OF TIGRS, CATS, LIONS, AND STRIPS

Financial engineering can be defined as the creation of a security having a risk-return profile that is otherwise unavailable. The creation of U.S. Treasury STRIPS is a classic example. The story starts in the early 1980s, when interest rates were high due to double-digit inflation rates. When rates later dropped, the descent was steep and dramatic. For example, yields on 10-year Treasury notes averaged 14.30% during the month of June 1982 and fell to 10.85% by June 1983. Treasury yields then rose and averaged 13.56% for June 1984 before another descent to 10.16% in June 1985 and farther down to 7.80% for June 1986. Figure 2.1 displays the monthly averages of daily 10-year Treasury yields from April 1953 through May 2010.

Savvy investors during times of decreasing inflation and lower market interest rates naturally prefer long-term, low-coupon bonds. When yields go down, these securities appreciate in value much more than shorter-term, higher-coupon bonds. In Chapter 6, we return to this idea using the concept of duration. Even buy-and-hold institutional investors see value in low- coupon debt or, better yet, zero-coupon bonds, because cash flows received prior to maturity have to be reinvested at lower and lower rates, reducing the total return over the investment horizon.

The problem in the 1980s for such savvy investors was a limited supply of Treasury zero-coupon bonds. Until 1983, the Treasury issued “bearer” bonds – investors actually would have to clip each coupon from the bond

FIGURE 2.1 Monthly Averages of Daily 10-Year U.S. Treasury Note Yields from April 1953 to May 2010

Source: Federal Reserve Bank of St. Louis, FRED data series GS10

certificate and present it to the government for payment. (This usually was handled by the investor's broker and the Federal Reserve.) There was at the time a small market in zero-coupon Treasury debt created by physically clipping coupons corresponding to future payments and selling them as separate obligations. Some corporations issued zeros, but many investors seeking to benefit from lower market interest rates did not want to bear long-term corporate credit risk because the economy was just coming out of a deep recession.

This scenario provided fertile ground for financial engineering. Investment banks, notably Merrill Lynch in this story, found a way to supply the security that the market demanded. The bank would buy coupon-bearing Treasury securities – for instance, $100 million in par value of 30-year bonds having a coupon rate of 12.50% – and place them in a special-purpose entity (SPE). The SPE here is a single-purpose dedicated trust – it is empowered only to own the bonds and collect the payments; it cannot sell or lend the bonds, write options on them, or use them as collateral on loans in the repo market. The SPE then issues zero-coupon securities, which essentially are ownership rights corresponding to the coupon and principal payments. For example, the SPE could issue 0.5-year, 1.0-year, 1.5-year, out to 29.5-year zero-coupon debt having total face value of $6.25 million for each maturity and 30-year zeros having a face value of $106.25 million.

Merrill Lynch pioneered the market for “synthetic” zero-coupon Treasuries and cleverly named them Treasury Investment Growth Receipts, known by the acronym TIGRS. Selling the TIGRS for more than the purchase price of the coupon Treasuries that were placed in the SPEs became a significant source of profit for Merrill for several years in the early 1980s. Given that success, it is no surprise that other investment banks copied the design (and feline-inspired acronym) – Salomon Brothers created CATS (Certificates of Accrual on Treasury Securities), and Lehman Brothers created LIONS (Lehman Investment Opportunity Notes).

An important sales outlet for the financially engineered Treasury zeros was Individual Retirement Accounts (IRAs). Back then, all taxpayers could put up to $2,000 into an IRA and subtract that amount from pretax income. For example, Merrill priced the zero-coupon TIGRS, each of which had a face value of $1,000, to fill out the allotment. For instance, 30-year TIGRS could be priced at $50 to yield 10.239% (s.a.). The thundering herd of Merrill brokers would suggest putting 40 such TIGRS into your IRA for the year, or perhaps for older taxpayers, 8 TIGRS priced at $250 to yield 10.151% (s.a.) over 14 years. (How those yields are calculated is covered in the next section.)

At the time, this use of an SPE to create a new security was fairly new. The key legal aspect of the design was that the structure allowed the TIGRS, CATS, and LIONS to be deemed U.S. Treasury credit risk and not a liability of the investment bank behind the process. Moreover, the SPE was “bankruptcy remote” in that default by Merrill, Salomon, or Lehman, while unthinkable at the time, would not allow their creditors to go after the underlying Treasuries. Also, the creation of synthetic zeros involved a change in tax status that had to be approved by the Internal Revenue Service – more on that in Chapter 4.

In 1985, the U.S. Treasury responded to the success (and profitability) of TIGRS, CATS, and LIONS with some clever financial engineering of its own – the STRIPS program. After 1983, Treasury securities were no longer issued in bearer form and were registered by a CUSIP (the acronym for Committee on Uniform Security Identification Procedures). Each Treasury bill, note, and bond has its own CUSIP. The innovation was to assign a CUSIP to each coupon and principal cash flow in addition to the overall security. Lor example, an 8%, 10-year Treasury note effectively became a portfolio of 20 separately registered coupon interest securities each with a face value of 4 (per 100 of par value) and one principal security for a face value of 100. That each security had its own CUSIP facilitated trading – the coupons could be stripped off and sold as zero-coupon C-STRIPS, and the principal could be sold as zero-coupon P-STRIPS.

C-STRIPS have a special feature in that the supply for each CUSIP increases over time, thereby enhancing the liquidity of the security. That is, coupon interest to be paid on a given date – say, February 15,2018 – has the same CUSIP regardless of which Treasury note or bond it originally comes from. The original coupon rate is irrelevant. P-STRIPS, however, always correspond to the original security and have a unique CUSIP. Their supply is fixed at issuance.

The STRIPS program has become very successful, and nowadays government securities dealers quote bid and ask prices on a full term structure of C-STRIPS and P-STRIPS. That success eliminated the profitability of TIGRS, CATS, and LIONS to the investment banks because STRIPS did not need the cumbersome SPE structure. Also, the arbitrage strategy of bond reconstitution emerged. This strategy is to buy the various C-STRIPS and P-STRIPS in sufficient quantity to rebuild a specific Treasury note or bond. When the purchase price for the parts is less than the sale price of the assembled whole, a profit is made.

An interesting phenomenon is that the prices on long-dated P-STRIPS typically are a bit higher than C-STRIPS that mature on the same date. For example, for trading on August 19, 2010, the Wall Street Journal reported that the ask prices on C-STRIPS and P-STRIPS due February 15,2040, were 30.926 (percent of par value) and 31.422, respectively. The prices for trading on January 28, 2014, were 36.872 and 37.552. This pricing pattern has been quite persistent even though the credit risks and taxation on the zero- coupon bonds are the same.

One reason for the price difference is the greater supply, and hence liquidity, of the P-STRIPS. The coupon rate on the underlying Treasury bond is 4.625%, so for every 100 in par value of P-STRIPS, there initially are only 2.3125 of C-STRIPS. Another reason is that P-STRIPS allow the owner to carry out bond reconstitution arbitrage more easily. Suppose that STRIPS suddenly start to trade at low prices (and high yields) relative to the original Treasury bond. If a dealer or hedge fund already owns the P-STRIPS, only the sequence of C-STRIPS needs to be purchased to reconstitute and sell the T-bond. However, if the C-STRIPS are owned, the arbitrageur would have to buy a large quantity of P-STRIPS as well as the remaining C-STRIPS. So, the “option” to reconstitute when profitable is priced into the P-STRIPS. A buy-and-hold investor naturally prefers the higher yield on the C-STRIPS.