“Far from Par”- How does Contract Design Impact the Return ...

"Far from Par"- How does Contract Design Impact the Return on Treasury Floating Rate Notes?

by

Karan Bhanot and Liang Guo1

Abstract

Floating Rate Notes were introduced by the US Treasury in 2014 to lower overall borrowing costs and diversify the funding base for US Government Bonds. Using daily data on all Floating Rate Notes issued from 1/2014 ? 10/2016 we find that they often trade at a premium or discount and earn an excess return relative to the underlying T-bill index. The excess return is driven by contractual features of these notes- the imperfect synchronization between the indexing dates used for coupon accruals, a spread over the index rate, and a deflation put option due to a lower zero bound on the accrued coupon. Since FRNs are derivative instruments, their returns also reflect in part the liquidity and arbitrage constraints implicit in the derivative markets.

JEL classification: G01, G12, G28 This Draft: January 15, 2017 Preliminary version. Do not circulate or cite

1Professor of Finance, Department of Finance, College of Business Administration, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX 78249 and Assistant Professor of Finance, College of Business, California State University at San Bernadino, CA 92407, Emails: karan.bhanot@utsa. edu and lguo@csusb.edu, Phone numbers: 210-458-7429 and 909-537-3257. Fax for both authors: 210-458-6320. Corresponding author- karan. bhanot@utsa.edu.

1

"Far from Par"- How does Contract Design Impact the Return on Treasury Floating Rate Notes?

1. Introduction

Treasury Floating Rate Notes (FRNs) are government bonds whose coupon payments are linked to the yield on thirteen-week Treasury Bills. The US Treasury introduced FRNs in the year 2014 as an avenue to diversify their funding base, lower borrowing costs, and manage the maturity profile of its outstanding debt. Prior to the first time sale of these bonds the Treasury provided market participants an opportunity to comment on the design, term, and other features of these notes keeping in mind the desired objectives of the Treasury.2 In addition to these traditional factors, the financial crisis has added concerns about the stability of the funding base for Treasury Debt. The adoption of structural and operational reforms to address run risks in Money Market Funds further accelerated the impetus to offer FRNs.3 As noted by Cochrane (2013)4:

"a benefit of floating rate debt is that it is a little bit safer. It is possible that markets refuse to roll over debt, leaving the Treasury technically insolvent, unable to pay the principal on its existing debt. This is basically what happened in Greece. .... The deeper benefit, I think, is that the floating-rate Treasury debt opens the way to a run-free financial system"

In this paper we ask: how do contractual features of FRNs impact their returns? We draw implications of the FRN security design for Treasury funding costs.

Contractual features (e.g., covenants, convertibility provisions) of a financial asset determine the cash flow profile of the asset. Investors analyze the nature and uncertainty of future cash flows that stem from these contractual features, and this in turn determines traded prices and consequently the returns on these instruments. Two key decisions are required with respect to the design of FRNs - First, what index must be used to accrue the coupon of these notes? Second, what is the frequency with which the relevant index should be reset? The thirteen-week T-bill yield was chosen as the index to determine the coupon on the FRN based on feedback from market participants keeping in mind that the index should reference a liquid, traded rate with a transparent

2 See . treasurydirect. gov/instit/statreg/auctreg/ANPR2012.pdf 3 See 4 See

1

pricing mechanism. Thus, the index rate setting mechanism was seen as robust and with sufficient pricing history.5 In particular, the reference yield for the FRN is based on the auction rate on a thirteen-week T-bill in the auction from the previous week. In addition to the lag period of one week between the index valuation date and coupon accrual, another feature of FRNs is that coupon accrual cannot be negative even when a T-bill auction is at a negative rate.

An investor considering the choice between a FRN and rolling over T-bills will find the FRN a good choice given that the investor does not incur any transaction costs involved in rolling over T-bills. This would suggest that FRN should deliver returns that are comparable or lower than a strategy of rolling over thirteen-week T-bills. However, equilibrium based arguments suggest that contractual features impose a risk relative to rolling over T-bills and this risk should be compensated in the returns on FRNs. First, the lag period between the index date and accrual date introduces uncertainty about the coupon at the next reset date, and the consequent accrued coupon when the contract matures. Second, when T-bill rates are negative, the lower zero bound on coupon rates caps the coupon accrual.

Concurrent with equilibrium considerations related to index lags and the deflation put option, arbitrage based arguments suggest that FRN prices should be linked with a strategy of rolling over T-bills. Since FRNs are derivative instruments on the underlying index, were the returns to differ to a large extent hedge funds and other institutional investors would have an incentive to exploit such deviations. An arbitrageur who seeks to exploit the return differential between FRN and T-bill returns will buy an FRN and short sell the corresponding T-bills or vice versa. The ability to consummate these transactions depends on the ability to short sell T-bills and finance the FRN purchase. Shleifer and Vishny (1997) argue that while the textbook version of arbitrage requires no capital and entails no risk, but in reality all arbitrage is constrained and is risky. In terms of constraints, none is more important than the availability of capital. As noted by Brunnermeier and Pedersen (2009): "trading requires capital and capital availability is subject to market conditions". For example, an arbitrageur seeking to exploit an apparent mispricing would finance the FRN purchase via a repurchase agreement using the asset as collateral. Thus, the ability to get collateralized loans to finance a purchase, collectively termed "funding liquidity", is important in exploiting the arbitrage and ensuring that relative prices converge. Such funding liquidity is subject to economic conditions and varies with time. In addition to the funding liquidity

5 See Goldman Sachs comment dated 01/22/2013 on CFR Part 356, Docket no. BPD-2012-0002.

2

constraints, an arbitrageur may not be able to fully finance a bond purchase using the asset as collateral because lenders require a margin (or haircut) to protect them against adverse movements in the collateral's price and their corresponding ability to sell the asset were the borrower to default. The margin therefore depends on the type of asset under consideration (asset risk and asset specific liquidity). Asset specific liquidity or the ease with which a security can be traded can be driven by a number of factors, some of which are unique to the asset under consideration. In general, these factors include exogenous transaction costs, inventory risk, and private information, amongst other considerations.

The funding and asset specific liquidity constraints discussed above are especially important in the case of derivative assets that require leveraged purchases of the underlying asset for replication of the derivative payoffs. Garleanu and Pedersen (2011) provide a theoretical model wherein deviations in relative prices of derivatives and the underlying asset are a result of changes in both funding liquidity (shadow cost of capital to arbitrageurs) as well as asset specific liquidity (determinants of margin requirements). In particular, they show that when risk-tolerant investors are margin constrained and risk-averse investors take on optimal allocations, the basis between a derivative and its underlying asset is non-zero in equilibrium. The basis depends on relative margins of the asset and the derivatives, and the leveraged investors shadow cost of capital.

Using daily data on all FRNs issued by the US Treasury from 1/2014 to 10/2016 we examine whether and the extent to which FRN returns are driven by contractual and arbitrage considerations. We find that Treasury FRNs deliver a daily return of 37.6 bps (reported on an annualized basis) and an excess return of 27.41 bps relative to the strategy of rolling over short term T-bills. FRNs also deliver an excess return relative to other benchmarks for short term funding costs.

Our empirical tests reveal that the excess returns on FRNs are positively linked to proxies that capture the impact of contractual features. First, the realized return is negatively related to a proxy for indexing lag. The excess return is negatively related to the fixed spread on FRNs and positively associated with the changes in the maturity matched yield to maturity on comparable fixed rate treasuries. Finally, excess returns are higher when a measure of the distance to the zero lower bound on coupon accruals are higher.

In addition to equilibrium consideration, our results also show that limited arbitrage results in returns that are not perfectly synchronized with the index. We find that excess FRN returns are

3

positively related to the corresponding maturity swap spread. The swap spread is in turn driven by liquidity in the interbank market and the risk premia associated with these liquidity constraints. Thus, collectively these empirical results suggest that FRN returns are larger than those on T-bills, and that in turn imposes higher borrowing costs on the US Treasury.

Our paper contributes to several branches of the finance literature. First, to our knowledge this is the first paper to analyze returns in the Treasury FRN market. The market has grown in size to over $300 billion and constitutes an important avenue for Treasury funding of US debt. Our results therefore have important implications for the determinants of funding costs and considerations in the design of such contracts. We show that investors demand a premium for investing in this market that in turn increases immediate funding costs for the Treasury. However, these costs should be weighed against the corresponding benefits from the increase in funding base and other benefits noted by the US Treasury. In related work, Greenwood, Hanson and Stein (2015) provide a theoretical model that trades off the benefits of short-term debt against the refinancing risk implied by the need to roll it over more often.

Our study is related to the work on the optimal maturity of government debt and borrowing costs. While the US Treasury had previously borrowed using a combination of Treasury Bills with a maturity of less than a year, Treasury Notes with maturities of two to ten years, and long term government bonds, the recent introduction of inflation indexed bonds followed by FRNs has changed the maturity structure of debt. Economic theory asserts that whether the government borrows long term or short term is irrelevant (Barro (1974)). The assertion is based on the assumptions that taxation creates no deadweight costs, capital markets are frictionless, and that government debt does not provide any other benefits. These assumptions may not hold in the practice since, for example, investors value the liquidity of government debt. Greenwood, et al. (2010) showed that these benefits are the basis of a theoretical model that trades off the advantages of short-term debt against the refinancing risk implied by the need to roll it over more often in. Thus, FRNs reduce interest rate risk and provide a liquidity benefit that may attract investors willing to accept this lower risk than relative to comparable maturity fixed coupon treasury bonds.

Second, our paper is related to the literature on contract design. The finance literature has extensively studied the impact of contract design and bond covenants on asset prices in the market with credit risk (Leland (1994), Leland and Toft (1996)). Such bond contractual features are intended to reduce agency costs between bond holders and equity holders (Smith and Warner

4

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download