The time‐to‐maturity pattern of futures price sensitivity to news

• DOI: http://doi.org/10.1002/fut.22046
• Date: 01 January 2020
• Author: Hoang‐Long Phan,Ralf Zurbruegg
• Published date: 01 January 2020

J Futures Markets. 2020;40:126–144.wileyonlinelibrary.com/journal/fut126

|

© 2019 Wiley Periodicals, Inc.

Received: 13 November 2018

|

Accepted: 19 July 2019

DOI: 10.1002/fut.22046

RESEARCH ARTICLE

The time‐to‐maturity pattern of futures price sensitivity

to news

Hoang‐Long Phan

1,2

|

Ralf Zurbruegg

2

1

School of Economics, University of

Danang, Vietnam

2

Business School, University of Adelaide,

Australia

Correspondence

Ralf Zurbruegg, Business School,

University of Adelaide, SA 5005,

Australia.

Email: ralf.zurbrugg@adelaide.edu.au

Abstract

This paper examines the effect time‐to‐maturity has on how sensitive futures

prices are to news flows. Unscheduled daily news flows that relate to the

underlying asset of a futures contract are related to the daily realized volatility

of futures to calculate a price‐news sensitivity ratio. The observed pattern

follows an inverted U‐shape relationship and has a bearing on whether the

maturity effect will be noticeable in a futures contract. This paper also shows

that by examining the peak‐to‐maturity of the price sensitivity to news pattern,

it is possible to better identify which contracts are more likely to yield higher

volatility.

KEYWORDS

commodity futures, price sensitivity to news, time‐to‐maturity

JEL CLASSIFICATION

G10; G12; G14

1

|

INTRODUCTION

With the increased financialization of the commodity markets and the corresponding rise of speculative trading activity

within commodity futures, the pricing behavior of these futures has undergone substantial change. These changes

include greater cross‐market linkages that commodity futures returns have with equity indices (Büyükşahin & Robe,

2014), the impact that open interest has in predicting commodity returns (Hong & Yogo, 2012), changes in the risk‐

bearing preferences of commodity producers (Acharya, Schnabl, & Suarez, 2013), as well as changes in the relation

between futures prices across contract maturities (Buyuksahin, Haigh, Harris, Overdahl, & Robe, 2008).

One other consequence of having a greater amount of speculative trading within commodity futures is the role that

information flows have on trading behavior, and its subsequent impact on futures price volatility. While Samuelson

(1965) initially proposed that futures return volatility should increase closer to the maturity date of the contract (i.e., the

maturity effect), it is based on the logic that the closer a futures contract is to maturity, the more sensitive the futures

price is to information regarding the value of the underlying asset (i.e., the fundamental). However, as speculative

trading is driven by information flows, and as commodity futures continue to experience a rise in speculative trading

activity, the sensitivity that futures prices have to information will also necessarily be related to this trading activity.

This motivates us to directly examine the futures price sensitivity to information flows over the life of a futures contract

as an explanation for why the maturity effect is not always observed.

1

While the Samuelson (1965) hypothesis suggests

1

Rutledge (1976) finds the maturity effect is present for silver and cocoa futures, but not for wheat and soybean oil. Milonas (1986) shows the presence of the maturity effect in several commodity

futures. Bessembinder, Coughenour, Seguin, and Smoller (1996) document the effect for agricultural futures, crude oil, and, to a certain extent, metals. Duong and Kalev (2008) find similar results for

agricultural futures but not for crude oil and gold. Daal, Farhat, and Wei (2005) find little evidence of it across a wide range of contracts while Jaeck and Lautier (2016) show the maturity effect is

present in electricity futures.

we should find futures price sensitivity to information on the fundamental increases as a contract nears maturity, we

speculate that this may not be the case and that it will be more closely related to the trading activity of the futures

contracts.

Previous studies have shown that price sensitivity to information is related to the level of trading activity (see Admati

& Pfleiderer, 1988; Bessembinder & Seguin, 1993; Kyle, 1985; Ripple & Moosa, 2009). Specifically for futures markets,

Bessembinder and Seguin (1993) document that futures return volatility is positively related to expected trading

volumes and negatively related to expected open interest. In addition, Ripple and Moosa (2009) show that trading

volume and open interest play dominant roles in explaining the time pattern of futures volatility. This leads us to

hypothesize that the sensitivity of futures prices to news will be a quadratic function of time‐to‐maturity. This is based

on the fact that the level of market activity of commodity futures (open interest and trading volume) generally follows

an inverted U‐shape pattern. Work dating back to the U.S. Department of Agriculture (1960) and Powers (1967) shows

that open interest, reflecting hedging demand, rises as commodities are produced and stored. Trading volume is

expected to follow a similar pattern since trading by speculators generally corresponds to hedging activity (Leuthold,

1983; Powers, 1967; Working, 1970). As a simple illustration, Figure 1 shows the average trading volume and open

interest over the life of September wheat futures contracts traded between 2003 and 2014. Both the levels of trading

volume and open interest display an inverted U‐shape pattern. It often tops after harvest (for agricultural futures) when

stocks are at their peak and starts to decline as hedgers close their positions when stocks are moved out of storage.

We also expect that the price‐news sensitivity pattern will subsequently impact the volatility pattern of futures

contracts. Specifically, our expectation is that there will be stronger (weaker) evidence of volatility increasing closer to

the maturity date if the inverted U‐shape of the price‐news sensitivity pattern tilts more (less) toward maturity. Our

findings, therefore, can help to explain the mixed results of previous studies that examine the maturity effect. In

particular, we conjecture that only when the price‐news sensitivity pattern tilts more toward maturity (and thereby

behaves more like something resembling a linear pattern), will the maturity effect be noticeable.

To directly test the price‐news sensitivity pattern, we utilize the Thomson Reuters News Analytics (TRNA) database,

which provides comprehensive data on daily commodity news coverage and has been used in recent studies examining

commodity news flows (see, e.g., Clements & Todorova, 2016; Smales, 2014, 2015). We measure news flows as the

number of daily news items relating to a specific commodity and calculate a measure of the sensitivity of futures return

volatility to news flows,

S

ENSITIVITY

, as the ratio of return volatility to the number of daily news items. We

empirically test our hypothesis on 12 commodity futures trading on four exchanges under the Chicago Mercantile

Exchange group of exchanges (CME Group). The futures include agricultural (grains, oilseeds, and livestock), metals,

and energy futures for all contracts traded between January 1, 2003 and June 30, 2014. We represent volatility as the

daily realized volatility calculated from the sum of the 5‐min squared realized returns (Andersen & Bollerslev, 1998)

using intraday data from Thomson Reuters Tick History (TRTH).

Our empirical analysis starts by establishing that both trading volume and open interest follow an inverted U‐shape

pattern over the life of a commodity futures contract, before proceeding to examine how

S

ENSITIVITY

changes over

the lifespan of these futures. Multivariate regression analyses provide evidence of an inverted U‐shape price‐news

sensitivity pattern for the commodity futures. To analyze the pattern in more detail, we also calculate the peak‐to‐

maturity (

P

T

M

) for each futures contract’s open interest, trading volume and

S

ENSITIVITY

.

P

T

M

is the number of days

between the peak of the price‐news sensitivity pattern and the maturity date, measured as a percentage of a contract’s

life. A smaller

P

T

M

indicates that the pattern peaks later in the life of the contract (i.e., the pattern tilts more toward

0

20,000

40,000

60,000

80,000

100,000

120,000

140,000

160,000

180,000

200,000

020406080100120140160180200220240260280300320340360

Number of contracts

Days to maturity

Volume Open Interest

FIGURE 1 Average trading volume and open interest over the life of September wheat futures contracts traded between 2003–2014

PHAN AND ZURBRUEGG

|

127