The information in global interest rate futures contracts

• Published date: 01 June 2022
• Author: Robert Brooks,Brandon N. Cline,Pavel Teterin,Yu You
• Date: 01 June 2022
• DOI: http://doi.org/10.1002/fut.22323

Received: 4 June 2020

|

Accepted: 6 March 2022

DOI: 10.1002/fut.22323

RESEARCH ARTICLE

The information in global interest rate futures contracts

Robert Brooks

1

|Brandon N. Cline

2

|Pavel Teterin

3

|Yu You

4

1

Department of Economics, Finance, &

Legal Studies, Culverhouse College of

Business, The University of Alabama,

Tuscaloosa, Alabama, USA

2

Department of Finance and Economics,

Mississippi State University, Mississippi

State, Mississippi, USA

3

John B. and Lillian E. Neff Department

of Finance, John B. and Lillian E. Neff

College of Business and Innovation, The

University of Toledo, Toledo, Ohio, USA

4

Li Anmin Institute of Economic

Research, Liaoning University, Shenyang,

China

Correspondence

Yu You, Li Anmin Institute of Economic

Research, Liaoning University,

66 Chongshan Middle Road, Huanggu

District, Shenyang 110036, China.

Email: f91yy@163.com

Funding information

National Office for Philosophy and Social

Sciences (China), Grant/Award Number:

21BJL028

Abstract

We investigate information contained in the term structure of interest rate

futures contracts in the United States, Eurozone, UK, and Switzerland. We

find that current forward‐spot differentials often predict return premiums and,

especially, future spot rates. This predictability follows time‐series patterns

common to all four markets, except around crises. Macroeconomic indicators

are important determinants of predictability within and between markets. One

common factor captures a significant portion of variation in predictability. No

single market has a dominant share of macroeconomic indicators linked with

the common predictability factor. Inflation and exchange rates arise as the

most important determinants of the common factor.

KEYWORDS

expectations hypothesis, rate‐based futures contracts, term structure

JEL CLASSIFICATION

E43, F36, F44, G13

1|INTRODUCTION

Studying cross‐country differences in the information content of the interest rate term structure is important

to policymakers and entities involved in international finance since differences in the yield curve are linked

to exchange rate movements and excess currency returns (Y. Chen & Tsang, 2013). Interest rate futures

contracts are also tied to the average of selected benchmark interest rates based on the loans funded between two

entities. Although these futures contracts reflect market expectations of future rates, they also contain other information,

such as anticipated default of the loans reflected in the benchmark rates and futures term premiums.

Our study uniquely combines the research techniques applied to the information content of the spot term structure

of interest rates with the interest rate futures market. Do the interest rate futures markets simply replicate the

information content of the respective spot market, or is the information content different? If different, as we document

here, other insights can be gained regarding the information conveyed in these futures markets. To our knowledge,

there are no other articles examining this connection.

Estrella and Hardouvelis (1991) show that the US term structure reflects monetary‐policy‐independent factors that

predict real economic activity. Consequently, the term structure provides useful information to both private investors and

policy makers. Kuttner (2018) explores the consequences of the unconventional monetary policies enacted after the Great

J Futures Markets. 2022;42:1135–1166. wileyonlinelibrary.com/journal/fut © 2022 Wiley Periodicals LLC

|

1135

Recession. Based in part on numerous recent studies, he concludes, “A preponderance of evidence nonetheless suggests that

forward guidance and quantitative easing (QE) succeeded in lowering long‐term interest rates”(p. 142). As monetary policy

normally only targets short‐term interest rates, we are particularly interested in differences in information content based on

short‐term (3 months) and long‐term (3 years). Kuttner (2018) further reports on the influence of QE and term premiums.

We provide a unique examination of this issue based on interest rate futures contracts across four developed markets.

Estrella and Mishkin (1997) likewise find that the inference of real economic activity persists for European economies.

Evidence also suggests that the domestic term structure of one country predicts future real growth of the economy of

another country (Plosser & Rouwenhorst, 1994). Brand and Tripier (2021) note that the Euro area and the United

States were highly synchronized during the Great Recession, but they diverged afterward. They conclude, “Our results

show that risk shocks have stimulated US growth in the aftermath of the Great Recession and have been the main driver of

the double‐dip recession in the Euro area”(p. 137). Thus, we explore linkages between macroeconomic and financial

indicators across four developed economies: the United States, Eurozone, United Kingdom, and Switzerland.

Just as the ability of the term structure to forecast real economic activity and growth in different countries is

important to investors and policy makers, so is the ability of current forward rates to predict future rate changes and

holding period returns in different countries. Most of the existing literature on the estimation of term premiums and

the information about future rate changes and holding period returns, however, uses data only from a single country

and only spot rates, most often the United States (Wright, 2011).

Studies that do examine the term structure across countries find that there exist global yield factors that explain

significant fractions of a country's yield curve dynamics (Diebold et al., 2008). The studies also find that the term

premium component of forward rates in the 10 major industrialized countries with independent monetary policies has

trended down for the last 20 years, a phenomenon best explained by lower inflation uncertainty resulting from changes

in monetary policy (Wright, 2011). Consistent with this evidence, Dahlquist and Hasseltoft (2013) find that

international bond returns are predicted not only by local factors, but also by a global factor closely linked to US bond

risk premiums and international business cycles.

1

These drive risk premiums and expected short‐term interest rates in

opposite directions. They also observe increased correlations between international bond risk premiums over time and

advocate an increase in the integration of markets as a possible explanation.

The goal of this study is to examine the association between the informational content of the interest rate futures term

structure across economies. Evidence suggesting convergence of the informational content across countries would be

consistent with the literature on global yield factors and a global financial cycle. Alternatively, evidence suggesting that the

term structures of these markets contain distinct information would suggest that market participants in different

economies behave in different ways, perhaps due to differences in macroeconomic conditions or monetary policy.

We focus our analysis of the term structure on Libor‐or its equivalent‐based interest rate futures contracts. There

are several reasons why the information content of global interest rate futures may vary from the information content

of domestic interest rates. First, the observed rate is based on the futures market rather than the spot market. Thus, we

expect differences between the spot rates and implied futures rates because the futures market allows market

participants without access to interbank transactions to trade on their information or sentiment regarding Libor. By

examining Libor‐based futures, we are potentially capturing a key component of all financial valuation, the freely

traded base discount rate. Financial valuation involves projecting future expected cash flows and discounting these

expectations at some risk‐adjusted discount rate. Key to formulating this risk‐adjusted discount rate is the base interest

rate, historically proxied for by the Libor rate. Unfortunately, Libor suffered from market manipulation for a period.

Abrantes‐Metz and Metz (2012), Abrantes‐Metz and Verstein (2013), and Abrantes‐Metz et al. (2011) document that

at the beginning of the 2008–2009 financial crisis, Libor displayed patterns that differed quite dramatically from the

other interest rates. By the end of 2012, several large‐scale investigations were underway examining the alleged

fraudulent manipulation of Libor settings. The results of these investigations resulted in several criminal convictions

and massive fines paid.

2

As the scandal unfolded, rate rigging was alleged in all four markets covered in this paper. This

Libor scandal has resulted in some regulators calling for the termination of Libor calculations altogether, which, if

successful, would result in the disappearance of the Libor‐based futures and deny the market participants the

opportunity to incorporate their beliefs about the discount rates. On the basis of the scandal, there could be differential

effects of the scandal in each market and at different maturities.

1

Existence of a global risk factor and its dependence on macroeconomic conditions in larger markets are the two features echoed by the emerging

literature on the global financial cycle (Miranda‐Agrippino & Rey, 2015; Rey, 2015).

2

See also Datoo (2020) for a historical review of Libor and the related Libor scandal.

1136

|

BROOKS ET AL.

Second, the information content could differ from market to market due to the influence of macroeconomic

conditions and government policies. This prediction is in line with the literature linking macroeconomic indicators and

monetary policy to the term structure. See, for instance, Rudebusch (1995) and Rudebusch and Wu (2008).

3

Rather than impose a factor structure on the yield curve, we investigate time‐series variation in the information

content of the term structure by employing simple forecasting regressions involving forward rates. In this regard, we

extend the work of Brooks, Cline, and Enders (BCE, 2015) by applying their methodology to global interest rate futures

contracts. BCE (2015) explore the informational differences between constant maturity treasury (CMT) yields and

London Interbank Offered Rate (Libor) in the United States. They extend the work of BCE (2012), which examines the

informational content in Libor spot rates. BCE (2012) document that during recent periods more information exists in

Libor spot rates concerning future interest rates than expected holding period returns. In their extension, BCE (2015)

document significant differences in the informational content between Libor and CMT. They also observe that the

information content changed significantly as the financial crisis began and that these changes were related to credit

default swap rates and tenor swap rates.

Overall, we examine 3‐month interbank rates in four markets—United States Dollar, Euro, Great Britain Pound,

and Swiss Franc—and empirically document five unique results. First, the information contained in implied USD

futures rates is significantly different from the information contained in USD spot rates previously explored by BCE

(2015). Intuitively, implied forward rates derived from Eurodollar futures contracts incorporate information that differs

from information in spot Libor because futures contracts are tied solely to 3‐month Libor, whereas USD spot interbank

rates address various maturities.

Our second major finding suggests that the four rate‐based futures contracts contain similar information. We

document based on univariate time‐series analysis of return‐premium predictability that there are significant

commonalities in the patterns. On the basis of differential beta analysis, we further document that unique patterns for

return‐premium predictability concentrated primarily around the Great Recession and SARS–CoV‐2 pandemic.

Similarly, we document based on univariate time‐series analysis of future spot rate predictability that there are

significant commonalities in these patterns. Further, the forward‐spot differentials were much better at predicting

future spot rates rather than term premiums. On the basis of differential beta analysis, we further document that spot

rate predictability was particularly strong around the Great Recession and SARS–CoV‐2 pandemic.

Third, we further examine macroeconomic and financial market indicators and find that these indicators are

important for both return‐premium predictability, as well as future spot rate predictability. Macroeconomic variables

are important for both predictability measures but explain a larger portion of the future spot rate predictability. Many

of the indicators have regression coefficients with opposite signs (i.e., if coefficient is positively linked with futures spot

rate predictability, then it is negatively linked with return‐premium predictability). As expected, monetary‐policy‐

related variables are better at explaining short maturity future spot rate predictability. Interestingly, M1 growth is

important for both shorter and longer maturities.

Fourth, using least absolute shrinkage and selection operator (LASSO) regressions, we document that inflation is

universally an important determinant of differences in predictability of both return premiums and future spot rate.

Surprisingly, macrovariables associated with the largest market (USD) did not drive the largest share of differences in

return‐premium predictability or future spot rate predictability.

Finally, applying principal component analysis (PCA), we conclude that a significant portion of variation in both

types of predictability at long and short maturities can be attributed to one common factor that is global in nature

because it accounts for various economic conditions in all markets and not only the largest market. Using LASSO

regressions, we show that inflation and exchanges rates appear to be most frequently linked to the common factor. The

fact that inflation can explain both the commonalities in predictability patterns as well as the departures from the

average pattern suggests that central banks follow a version of Taylor (1993) rule but use different tuning parameters.

The contribution of our findings to the existing literature is in providing distinct evidence that rate‐based futures

from four industrialized nations provide a useful window into future spot rate predictability as well as return‐premium

predictability. Forecasting both future spot rates and return premiums lies at the very foundation of critical investment

decision‐making and government policy decisions. The evidence presented here suggest that each market contains

similar information. This evidence is consistent with the findings in Diebold et al. (2008), Wright (2011), and Dahlquist

3

Despite its influence, monetary policy does not completely determine the term structure. In fact, the term structure contains useful information

beyond what can be explained by monetary policy and macroeconomic indicators (Estrella and Hardouvelis, 1991).

BROOKS ET AL.

|

1137