The formation of forward freight agreement rates in dry bulk shipping: Spot rates, risk premia, and heterogeneous expectations

• DOI: http://doi.org/10.1002/fut.21980
• Date: 01 August 2019
• Author: Nikos K. Nomikos,Ioannis C. Moutzouris
• Published date: 01 August 2019

Received: 27 September 2018

|

Accepted: 7 October 2018

DOI: 10.1002/fut.21980

RESEARCH ARTICLE

The formation of forward freight agreement rates in dry

bulk shipping: Spot rates, risk premia, and heterogeneous

expectations

Ioannis C. Moutzouris

|

Nikos K. Nomikos

Cass Business School, Faculty of Finance,

City, University of London, London, UK

Correspondence

Nikos K. Nomikos, Cass Business School,

Faculty of Finance, City, University of

London, 106 Bunhill Row, London EC1Y

8TZ, UK.

Email: N.Nomikos@city.ac.uk

Abstract

We examine the formation of forward rates in the dry bulk shipping industry.

We illustrate that the bulk of basis volatility can be attributed to expectations

about future physical market conditions rather than expectations about future

risk premia. However, there exists significant predictability of risk premia by

both price‐based signals and economic variables. To explain this finding, we

develop a dynamic asset pricing framework where, apart from having different

objective functions, agents might also differ in the way they form expectations

about future market conditions. Accordingly, we argue that the average investor

should hold “near‐rational”but slightly contrarian beliefs.

KEYWORDS

asset pricing, behavioral finance, biased beliefs, contrarian strategy, heterogeneous agents, law of

small numbers, shipping industry

JEL CLASSIFICATION

C13, G12, G13, G40

1

|

INTRODUCTION

In this paper, we extend the application of heterogeneous beliefs models to commodity derivatives. In particular, we

focus on the forward market for shipping freight rates, that is the market for forward freight agreements (FFAs).

Shipping is a very important sector of the world economy since 90% of the world trade is transported by sea and it is

justifiably considered as a leading indicator of world economic activity (Killian, 2009). During the last decade, in

addition to traditional shipping investors, the market for freight derivatives has also attracted the interest and

participation of investors from other sectors of the economy. Due to the distinct features of the shipping industry and its

highly volatile character, trading volume in shipping derivatives markets is expected to increase significantly over the

next years. Hence, understanding the pricing and trading dynamics of this market is important.

To the best of our knowledge, this is the first time that a structural, heterogeneous‐beliefs asset pricing model is

applied to a futures or forward market.

1

Thus, we provide a framework that can be adapted and evaluated empirically

in other commodity derivative markets. In doing so, our contribution to the literature is threefold. First, we are the

first to apply the variance decomposition framework in a derivatives market where the underlying asset is a

nonstorable service. Second, we document for the first time several noticeable empirical regularities related to FFA

J Futures Markets. 2019;39:1008–1031.wileyonlinelibrary.com/journal/fut1008

|

© 2018 Wiley Periodicals, Inc.

1

For instance, Ellen and Zwinkels (2010) and Bredin, Potì, and Salvador (2018) also use behavioral models with heterogeneous speculators to explain commodity price dynamics. Despite some

similarities to our model, commodity demand in their frameworks is not derived explicitly through a structural economic model as in our case.

rates and risk premia. Third, we propose a theoretical behavioral asset pricing model that can account for these

stylized facts.

We begin by analyzing the formation of the basis in the freight forward market. Fama (1984a, 1984b) and Fama and

French (1987) show that the variance of the basis of any futures and forward contract can be decomposed into the sum

of the covariance between the basis and the expected change in the spot price and the covariance between the basis and

the expected premium over the spot price at maturity; this premium reflects the excess return for an investor who goes

short on the derivative contract. We illustrate formally that volatility in the FFA basis can be attributed primarily to

expectations about future physical market conditions rather than expectations about future risk premia, as is generally

the case in commodity markets (Fama & French, 1987). This is justified on the basis that freight rates are subject to

supply and demand shocks which cannot be smoothed through short‐term adjustments in supply; the reader can

parallelize this to lack of commodity storability. This results in predictable variation of spot rates which, consequently,

increases the forecasting ability of the FFA basis. This is consistent with previous empirical evidence that predictability

of future spot rates is a decreasing function of commodity storability (Fama & French, 1987; French, 1986;

Hazuka, 1984).

While volatility in the FFA basis is primarily attributed to changes in expected spot rates, we cannot rule out the

existence of (possibly, time‐varying) risk premia. Accordingly, we provide evidence of several stylized features that

might be of interest to academic and practitioners alike. Specifically, in contrast to most commodity futures markets, we

find strong statistical evidence of contango; that is, realized risk premia are, on average, positive. In addition, there

exists significant predictability of future risk premia, consistent with the existence of a momentum effect: Lagged risk

premia strongly and positively forecast future risk premia. Finally, FFA risk premia can be strongly negatively

forecasted by both spot market signals and economic indicators related to commodity trade and shipping demand. The

existence of statistically significant predictability of future risk premia contradicts the unbiased expectations hypothesis

and, in turn, the efficiency of the FFA market. We further examine the validity of the unbiasedness hypothesis by

performing three frequently incorporated econometric tests which unequivocally suggest the existence of a bias in the

dry bulk FFA market. From a market participant’s perspective, those stylized features can be used to develop potentially

profitable trading strategies.

We develop a theoretical model of FFA price determination to reproduce our main empirical findings. The

proposed framework draws its main features from the latest generation of structural economic models in the

commodity futures literature (e.g., Acharya, Lochstoer, & Ramadorai, 2013; Gorton, Hayashi, & Rouwenhorst, 2012)

and has been modified and extended in two, quantitatively simple but conceptually important, manners. First, our

framework departs from the “theory of storage”explanation of “time‐varying”risk premia (e.g., Ekeland, Lautier, &

Villeneuve, 2018; Gorton et al., 2012) since shipping freight is a nonstorable service. An immediate consequence of

this is the extension of the (widely used) two‐period economic environment to an infinite horizon model, which

simplifies the empirical evaluation of the generated framework; accordingly, we validate the theoretical predictions

of our model through numerical simulations. Second, we incorporate the existence of distorted beliefs on a fraction of

the investor population; heterogeneous beliefs models provide an alternative way for researchers to explain empirical

regularities in asset prices that cannot be explained by traditional rational expectations models. This way, we

contribute to the generic commodity finance literature by incorporating explicitly the behavioral dimension in the

formation of derivative contracts rates.

Our discrete‐time economy consists of three types of agents; ship owners, charterers, and speculators. Apart from

having different objective functions, agents also differ in the way they form expectations about future market

conditions. While ship owners and charterers are fully rational investors, speculators are characterized by bounded

rationality and suffer from a form of “representativeness heuristic”which means that they “exaggerate how likely it is

that a small sample resembles the parent population from which is drawn”(Shefrin, 2000; Tversky & Kahneman, 1971).

As a result, following a shock in freight rates, speculators believe that rates will revert more rapidly to their previous

level than is the case in reality which results in a contrarian investment behavior on their behalf.

The use of behavioral models in equity markets is often justified by survey data which confirms the theoretical

predictions (Greenwood & Shleifer, 2014). Since there are no comparable detailed surveys regarding shipping industry

participants’beliefs and investment strategies, we justify the use of the proposed model by contradiction using both

theoretical predictions and model simulations. Namely, we show that a rational expectations model with a hedging

pressure bias, cannot explain the documented empirical regularities. Similarly, simulation tests suggest that to

simultaneously match all observed regularities sufficiently well, the average investor should hold “near‐rational”but

slightly contrarian beliefs.

MOUTZOURIS AND NOMIKOS

|

1009