Do Seasonal Tropical Storm Forecasts Affect Crack Spread Prices?

• Author: Kristin Fink,Jason Fink
• Published date: 01 May 2014
• DOI: http://doi.org/10.1002/fut.21607
• Date: 01 May 2014

DOSEASONAL TROPICAL STORM

FORECASTS AFFECT CRACK

SPREAD PRICES?

JASON FINK* and KRISTIN FINK

Individual storms in the Gulf of Mexico have been shown to affect crack spread futures prices.

As hurricanes strike refiner-dense areas, prices of refined petroleum products rise. We find that

crack spread prices are even affected by seasonal hurricane forecasts. We find this despite the

difficulties of long horizon forecasting, and that refiners are only exposed in a small portion of the

Atlantic basin. These effects are economically important. For example, a one standard deviation

increase in the June forecast of the net tropical cyclone activity in the upcoming season

increases 3-2-1 crack spread prices by over 9%. © 2013 Wiley Periodicals, Inc. Jrl Fut Mark

34:420–433, 2014

1. INTRODUCTION

The densest collection of petroleum refineries in the United States is a clustering of refiners in

Louisiana and Texas in the northwest Gulf of Mexico. By geographic happenstance, this

region is also one of the busiest locations for tropical storms on the planet. Therefore, the

market for refined petroleum in the United States is subject to significant risk from these

storms.

1

Crack spread futures, which provide a synthetic proxy for the profitability of

(functioning) refiners, have been shown to be sensitive to short term hurricane forecasts in the

Gulf region (Fink, Fink, & Russell, 2012). However, the reactions of these important futures

markets to seasonal hurricane forecasts, if any, have yet to be established. We test the effect of

the release of the leading seasonal hurricane forecast in the Atlantic basin (the Gray–

Klotzbach forecast from Colorado State University) on the crack spread market, and find

important economic effects.

2

Forecasts from the Gray–Klotzbach research team (hereafter, GK) are made at the

beginning of each December, June and August for the subsequent hurricane season (the peak

Jason Fink is the Chandler/Universal Professor of Banking at James Madison University, Harrisonburg,

Virginia. Kristin Fink is a Professor of Finance at James Madison University, Harrisonburg, Virginia.

The authors would like to thank Phil Klotzbach for providing verification data for the seasonal forecasts,

and an anonymous referee for valuable suggestions. We especially thank Justin Fink for excellent research

assistance.

*Correspondence author, James Madison University, MSC 0203, Harrisonburg, VA 22807. Tel: 540-568-8107,

Fax: 540-568-3017, e-mail: finkjd@jmu.edu

Received October 2012; Accepted January 2013

1

See, for example, Chen, Finney, and Lai (2005), Radchenko (2005), and Blair and Rezek (2008).

2

These effects may be increasing in importance. Bender et al. (2010) project a doubling of category 4 and 5 storms by

the end of this century.

The Journal of Futures Markets, Vol. 34, No. 5, 420–433 (2014)

© 2013 Wiley Periodicals, Inc.

Published online 16 February 2013 in Wiley Online Library (wileyonlinelibrary.com).

DOI: 10.1002/fut.21607