Do Volatility Smiles Matter for Pricing Asian Basket Options? The Case of Livestock Gross Margin Insurance for Dairy Cattle

Dec. 2014 - Dec. 2015

This project is a follow-up on a working paper of the same title my professor undertook back in 2012. The main research question is whether or not volatility smiles in commodity markets affect the premium of Livestock Gross Margin (LGM). The official LGM rating method assumes flat volatility curves for all futures contracts. Both the original working paper and this follow-up project show that the perceived volatility curves have negligible effect on LGM premiums in the official deductible range ($0.00 - $2.00). However, the curves make gradual and meaningful deviations from the official method as the deductibles wage higher and higher. 

My job was to expand the existing analysis to cover more years of data and bigger range of deductibles. The original working paper only used data from 2009 when the headwind of the financial crisis hit the dairy industry at its worst. Given 2009's historical notoriety, the original paper simply studied a special case of the markets. I extended the study to include ten years of data ranging from 2005 to 2014. Because the project uses 10 years worth of CME high-frequency data that the computer code from the original paper simply was not able to handle, I revamped the existing computer program and made numerous additions. 

Several highlights of this projects include:

• Managed 10 years of CME high-frequency trading data in Microsoft SQL Server 2012.
• Employed a moment-matching calibration method to fit generalized lambda distributions that model the 3-day arithmetic average prices of futures contracts. 
• Estimated a slew of copulas to test the relevancy of dependence structures.

Due to the complexity of the project, it is further explained in the following pages:

• Data Management: explains how data are transported from raw flat files to an in-house SQL Server. SQL Server techniques such as indices and stored procedures are utilized to improve data operation efficiency. 
• Distribution Calibration: details the statistical model and calibration procedures
• Simulation: describes how actuarially fair premiums are calculated through Monte Carlo simulation from the statistical model.