This post will concern the Chicago-Kansas wheat futures marketplace. The main goal of this article will be to introduce the reader to these markets, recent histories, trading strategies, and some off the shelve statistics such as correlation, covariance then apply a principle component analysis (PCA)in the next post of this series to show the end user how to construct a basic term structure model.

Wheat is a commodity that is utilized as a food or food derivative trading in many location such as Chicago (CME), Kansas (CBOT), Minneapolis (MGEX), Australia, etc. For this study we will focus only on the CME and Kansas contracts.

Please if you are going to trade these be familiar with the contract specifications:

CME wheat contract specifications: http://www.cmegroup.com/trading/commodities/grain-and-oilseed/wheat_contract_specifications.html

Kansas wheat contract specifications: http://www.kcbot.com/symbols_trading_hours.html

Some terminology, many grain traders refer to the spread between outrights of different exchanges such as front month Chicago to front month Kansas as the basis. I will keep this terminology and additionally add that Chicago is always the front basis contract. So if I say you are long the August basis, I mean long Chicago and short Kansas.

In terms of recent histories both of these contracts have only been trading electronically on CME Globex for a few years now and have picked up tremendous liquidity with such change. There is now listed calendar swaps, options, and other derivative products such as the SP GCSI index listed on the CME.

There are several trading strategies involving these contracts but the most prevalent is to trade the Chicago to Kansas basis in a 1 to -1 contract fashion in the same expiration. This by far has been the most popular but with the recent reduction in volatility, I have been showing traders how to take curve plays by doing the front Chicago expiration against the deferred Kansas expiration in a 1 to -1 trading ratio to get more fills. The risk is somewhat higher as these contracts tend to move more as you are trading more risk on the future expectation of the crop. There are term structure strategies which we will cover in the next posting covering principle component analysis (PCA) and independent component analysis (ICA).

Many wheat traders say that in the long term that the basis is mean-reverting. In the spreadsheet I posted on SourceForge recently (CBOT-KC_WheatAnalysis.xlsx), I showed this with a simple spread analysis and variance ratio. The variance ratio is shown to stabilize very fast and the price ratio is shown to be very stable (i.e. mean reverting). Is this the case intra-day? I will have to get data for another future analysis for us.

A clip from this analysis may be seen below:

Moving on to correlation but wait what is correlation? Correlation is a metric that describes the statistical dependence between two or more quantities. The most widely used correlation is Pearson correlation which is defined as:

Covariance(Chicago,Kansas)/(Stdev(Chicago)*Stdev(Kansas)) where stdev is the standard deviation and these are all based one the asset returns.

There are many types of correlation such as Rank correlation. See http://en.wikipedia.org/wiki/Correlation_and_dependence for more decent information.

A plot of this correlation may be found below which shows correlation by contract of Chicago against Kansas:

But wait correlation depends on the covariance? Yes, that is correct and covariance depends on the variance on the individual assets involved. The covariance may be seen below.

Gadzooks, that crossover of variance may explain the ackward correlations in the Kansas back months term structure.

This is a basic intro to the mechanics of Chicago-Kansas wheat market. Next time we will venture into constructing a wheat principle component analysis model to trade with!