There are differences in the information content of the constant maturity Treasury and LIBOR term structures that the authors evaluate. The differences became more pronounced during the recent recession.
The nature of both the constant maturity Treasury (CMT) rate and LIBOR has changed along with the relationship between them. The differences became more pronounced during the recent recession and are consistent with major shocks experienced by credit default swap (CDS) and tenor swap rates.
How Is This Research Useful to Practitioners?
Several features differentiate LIBOR and the CMT rate. The former is an aggregate interbank market rate based on the responses of contributing AA rated banks. Unlike the CMT rate, LIBOR is not prone to a “flight to quality” effect when markets are in turmoil because it contains an element of default risk. In contrast, the CMT rate reflects liquid, direct US government obligations whose rates, unlike LIBOR, are influenced by supply fluctuations as well as some special features desired by investors: partial tax exemption, added compensation for those engaged in securities lending, and eligibility to satisfy banks’ numerous regulatory requirements.
The information content in the CMT rate and LIBOR has become quite different. Moreover, this content change seems to have persisted, which forecasters need to take into account when using these term structures to forecast the future direction of interest rates.
The authors use daily data from the British Bankers Association and the Board of Governors of the Federal Reserve System to construct term structures from January 1986 to December 2013 to observe information content differences. In addition, they apply and extend existing research to track the development of these differences before and during the recent recession.
LIBOR tells two stories. Prior to the recent recession, the LIBOR term structure was a harbinger of future spot rate changes but revealed little about return premiums. During the recession, the roles reversed. Changes in default risk appear to be behind this finding. CMT rates, in contrast, were more stable and changes tended to be more gradual. Changes in credit exposure seem to be the driver of changes in the difference between LIBOR and CMT rates, not LIBOR rate manipulation and more accommodative monetary policy in response to the crisis.
How Did the Authors Conduct This Research?
The authors use four time-series regressions to evaluate separately forward rates’ predictability and whether implied forward rates contain information on return premiums or future changes in spot rates for both the CMT and LIBOR markets. The analysis also looks for periods when behaviors in these markets diverge.
The authors use three distinct approaches to determine whether values between the markets are time varying and whether changes in the LIBOR and CMT markets exhibit statistical significance. The first approach is a rolling window analysis to capture any subperiods that would reveal a significant difference between these markets. For the second approach, the authors follow Bai and Perron (Econometrica 1998) to estimate how many structural breaks take place between the markets, their magnitude, and the time of their occurrence. Finally, they calculate the time-varying coefficient across the LIBOR and CMT markets for each maturity and regression to look for disparities in the correlation (the generalized autoregressive conditional heteroskedasticity methodology).
The authors evaluate both term structures using the three methodologies. The information content of the LIBOR term structure changed markedly. Prior to the onset of the financial crisis, the LIBOR term structure was rich in information on future spot rate changes for three-month LIBOR but contained little information regarding return premiums. After the start of the crisis, however, this information content reversed. In contrast, changes in the information content of the CMT rate term structure were more measured and gradual. Breaks tended to be less frequent, particularly those associated with the recent recession, reflecting what seemed to be more stable CMT markets.
Changes in default risk are correlated with meaningful information differences between LIBOR and the CMT rate term structures. Default premiums are often higher during a recession and decline with maturity, whereas in stronger economic environments, the reverse tends to hold. In fact, changes in CDS rates, credit spreads, and tenor swaps echo and were consistent with those that occurred in both the CMT and LIBOR markets. Looser monetary policy and LIBOR manipulation do not appear to have had any meaningful impact on the results.
The behavior of interest rates and credit markets tells an interesting story and apparently more than one, given the significant differences in information content that the authors observe and substantiate through their investigation of the CMT and LIBOR markets. US Treasury and LIBOR term structures differ in their message, particularly at the beginning of the recent financial crisis. The authors’ work appears to have global implications given the widespread use of LIBOR to price a myriad of financial products. An ongoing effort to determine the impermanence of changes that the financial crisis brought about in the information content of the two rates could be the subject of future research.