Convenience Yields in Electricity Prices: Evidence from the Natural Gas Market

• DOI: http://doi.org/10.1002/fut.21807
• Author: Nikolaos Paratsiokas,Nikolaos Milonas
• Date: 01 May 2017
• Published date: 01 May 2017

Convenience Yields in Electricity Prices:

Evidence from the Natural Gas Market

Nikolaos Milonas* and Nikolaos Paratsiokas

This study develops a model in which changes in the electricity basis are directly related to

changes in the fuel’s storage level and the fuel’s convenience yield. The model is tested with

data from the US electricity and natural gas markets during the period 2004–2014. Empirical

results show a negative relationship between the electricity basis and the fuel’s convenience

yield giving support to the argument that non-storable commodities embody convenience

yields. © 2016 Wiley Periodicals, Inc. Jrl Fut Mark 37:522–538, 2017

1. INTRODUCTION

This paper examines the existence of convenience yields in electricity prices. Because

electricity is non-storable, it is largely assumed that convenience yields are absent from

electricity prices. The approach followed here focuses on the convenience yield embodied in

the prices of fuels used to produce electricity. Upon a sudden shortage of the fuel, its spot

price will increase relatively to the futures price giving rise to convenience yields. The idea

rests upon the fact that electricity is the output of burning a given fuel through a production

transformation process. The transformation to electricity will not occur unless the value of

fuel (including the convenience yield) is also transferred to the electricity output. Therefore,

in a free and competitive market, the increase in the fuel cost will give rise to an increase in

the electricity price. The latter will indirectly encompass the convenience yield of the

respective fuel.

Studying convenience yield is important to researchers and traders because it is a basic

determinant of the theory of storage and the pricing of futures contracts. Developed in the

earlier writings of Kaldor (1939) and Working (1949), this theory predicts that futures prices

reflect the cost of purchasing and storing the underlying commodity until the contract’s

maturity. Therefore, the cost of storage is perceived as the compensation to holders of

inventories to carry the commodity into the future and as such it affects futures pricing.

Following the work of Kaldor and Working, Brennan (1958) and Telser (1958)

formulated convenience yield as a negative function of the level of inventories. Recently,

following the same reasoning, Gorton, Hayashi, and Rouwenhorst (2013) treat convenience

Nikolaos Milonas is at National and Kapodistrian University of Athens, Athens, Greece. Nikolaos Paratsiokas

is at Institute of Economic and Industrial Research, Athens, Greece. The authors thank the participants of the

53rd Meeting of EWGCFM and the 2nd International Conference of RCEM at Chania, Greece, May 22–24,

2014 and the participants of the World Finance Conference in Venice, July 2–4, 2014 for useful comments

on an earlier draft. The authors alone have the responsibility for any remaining errors.

JEL Classification: G12, G13

*Correspondence author, National and Kapodistrian University of Athens, 1 Sofokleous Street, 105 59 Athens,

Greece. Tel: þ30 210 3689442, Fax: þ30 210 3689468, e-mail: nmilonas@econ.uoa.gr

Received October 2015; Accepted July 2016

The Journal of Futures Markets, Vol. 37, No. 5, 522–538 (2017)

© 2016 Wiley Periodicals, Inc.

Published online 25 August 2016 in Wiley Online Library (wileyonlinelibrary.com).

DOI: 10.1002/fut.21807

yield as a negative function of the inventory level in equilibrium. Convenience yield arises in

the case of a “stockout”of a commodity and represents the additional benefit received by

inventory holders after accommodating for the cost of storage. In another model, Routlege,

Seppi, and Spatt (2000) treat convenience yield as a function of inventories and consider an

exogenous shock arising from an imbalance between demand and supply. This imbalance

leads to an option in the hands of the commodity holder to sell the commodity when supplies

are scarce.

The option-like feature of convenience yield has been analyzed in a number of papers

including Heaney (2002), Henkel, Howe, and Hughes (1990), and Milonas and Thomadakis

(1997a). These models introduce the contingent nature of convenience yield attributed to

the level of inventories which, under various exogenous circumstances, may fall at low levels

that cannot meet current demand. To restore the demand/supply equilibrium, spot prices

increase above futures prices giving rise to convenience yields. In this case, the convenience

yield becomes “in-the-money.”At other times with ample inventories to meet current

demand, convenience yields are nil and become “out-of-the money.”

Convenience yields have been documented empirically in a number of studies. Fama

and French (1987) examined 21 commodities in the period 1966–1984 focusing on the

futures basis and futures risk premium. They found large variation of the basis due to changes

in interest rates and seasonality from which they inferred the existence of convenience yields.

Direct evidence that convenience yields have the characteristics of call options was provided

by Milonas and Thomadakis (1997b). The empirical evidence on corn, wheat, soybeans, and

copper over the period 1966–1995 shows the existence of convenience yields in direct

positive relationship with the volatility of spot prices and negative relationship with the level

of inventories.

Milonas and Henker (2001) examined the prices of Brent and WTI crude oils in the

period 1991–1995 and verified the positive relationship between convenience yield and spot

price volatility and the negative relationship with oil stocks. The authors found that

convenience yield estimates are significant portions of cash prices exceeding 2%, on average.

While the vast majority on commodity literature agrees on the importance and

applicability of the theory of storage and the existence of convenience yield in storable

commodities, concrete evidence in the case of non-storable commodities, like electricity, has

not been established yet.

Regarding electricity prices, Bessembinder and Lemmon (2002), Eydeland and Geman

(1998), Lucia and Schwartz (2002), and Pirrong and Jermakyan (2008) argued that due to

the non-storable nature of electricity, the cost-of-carry theory and the implied arbitrage

arguments are not applicable. Instead, they decompose futures electricity prices into the

expected spot price at delivery and the involved risk premium. Empirical studies by Cartea

and Villaplana (2008), Furio and Meneu (2010), and Longstaff and Wang (2004) employing

data of spot and futures prices from different power exchange markets, confirm the existence

of statistically significant risk premia that exhibit high volatility and regular changes in sign.

Other studies by Clewlow and Strickland (2000), Geman and Roncoroni (2006), and

Huisman and de Jong (2003) consider the term structure of electricity futures prices as

depending on the properties of the spot price such as its stochastic volatility, price jumps and

mean reversion. Yet, in doing so, they do not account for the potential impact of fuel inventory

levels on the electricity futures prices.

Two studies on electricity prices depart from previous research by taking the approach

of incorporating the fuel’s supply into the behavior of electricity prices. In this context,

Douglas and Popova (2008) accounted for the amount of natural gas in storage as a

determinant of the risk premium in futures electricity markets. In a similar approach,

Botterud et al. (2010) studied the impact of the level of water reserves on spot and futures

Convenience Yields in Electricity Prices 523