The Effects Of Light Crude Oil Costs And Stock Prices On Five Class I Railroads

The Effects of Light Crude Oil Costs and Stock Prices on Five Class I Railroads

I chose this study to determine if there was any significant effect on the relationship between the cost of (Brent) light, sweet crude oil, the largest type of oil in volume traded, and the stock prices of five Class I Railroad stocks operating in the United States. I am currently employed with the Brotherhood of Locomotive Engineers & Trainmen, the bargaining agent for the majority of engineers on the major Class I Railroads in the U. S. During the recent contract negotiations one of the important issues for labor's position in bargaining for wage increases, along with keeping the health & welfare contributions to a minimum, was the argument that the Class I Railroads have been making record profits and it would be fair and equitable to share some of that wealth with the men and women who actually do the work. I wanted to see if the numbers supported that conclusion.

In 1981, when the Staggers Act became law, the railroads were no longer regulated for commerce purposes. They no longer needed the Interstate Commerce Commissions approval for rate increases. This change in law allowed the railroads to take the position that they were now in a competitive market place. Based on recent financial reporting, three of the five railroads that I chose have broken their first quarter earnings records in 2007 by as much as 4% (Brotherhood of Locomotive Engineers & Trainmen News, June/July 2007, pg. 3).

This change in law has also resulted in the "cost neutral" (concessionary bargaining) that has taken place up until the latest agreement. Prior to this agreement the carriers demanded and were successful in getting work rule relief to counter all general wage increases. However, in light of the record profits that these railroads are enjoying, labor fought to get a larger piece of the pie and successfully negotiated a 17% wage increase over a 5-year period, along with setting a manageable cap on the health & welfare contributions. After analysis of the data that I have collected, I strongly believe that Labor's argument definitely had some weight.

Methodology

For this study, the closing stock prices of five Class I Railroads were gathered for the period of March 1, 2007 through May 31, 2007: Burlington Northern Sante Fe Railway Corp. (BNI), Canadian Pacific Railway (CP), CSX Transportation, Inc. (CSX), Norfolk Southern Railway Corp. (NSC), and Union Pacific Railroad Co. (UNP). For the same period of time, the daily closing price of a barrel of (Brent) light, sweet crude oil was collected. The stock price data was collected from Yahoo Finance (http://biz.yahoo.com/p/776conameu.html). The daily oil prices were collected from http://futures.tradingcharts.com/chart/CO/W.

Line Graphs and Linear Regression

Line graphs were created for this study using the daily closing stock price for each of the Class I Railroads selected and the daily closing cost of a price of light, sweet crude oil. All of the graphs contain a trend line, along with the estimated linear regression equation for each of the individual data groups. (see Charts 2-6). These equations were also used to calculate the P-value, which determines the probability that the independent variable, the daily cost of a barrel of oil, predicts the value of the dependent variable, the individual stock prices. Using a 95% confidence interval, there appears to be no significant relationship among any of the stocks. Using a 90% confidence level there appears to be a slight relationship with BNI and NSC (see Summary Output, pgs. 1-3).

Confidence Intervals

For this study, the data values gathered were from a select sample period and as such, the exact value of the population is not known. The range of values can be adjusted by choosing a confidence coefficient, which for this study I did at both the 90% and 95% range (see Tables 2a - 2b). This confidence interval does not give the exact value of the population mean (м), but it can determine with a 90% and 95% certainty, the range in which the population mean (м) lies. This range can be found using the confidence interval calculation using either the z-score or the t-score (see Equations 1 and 2). Both of these equations use the standard error calculation, which modifies either the population standard deviation (у) or the sample standard deviation (s), by the square root of the sample size (n) (Anderson et al., 2006). The difference between these two calculations is that the z-score is used when the population standard deviation (у) is known and the t-score is used when it is not. For purposes of this study, both were used to show the difference in the confidence intervals based on the known and unknown.

Correlation

One of the primary functions of the proxy is to provide a means of determining the direction of movement, in this case either the increase or the decrease in a stock's price, based upon the movement of that proxy, in this case the daily cost of a barrel of light, sweet crude oil. For purposes of this study, the Pearson Product Moment Correlation Coefficient for Sample Data (Anderson, et al., 2006)(see Equation No. 3), hereinafter referred to as the "correlation", was used due its ability to measure the linear relationship between two variables that is affected by the units of measurement contained within the x and y variables.

Another function of the correlation coefficient is that, when compared to the covariance, a value between -1 and +1 is given, allowing for comparison between each individual correlation. This value can also show either a positive or negative relationship. If the correlation is positive, when one variable increases, so does the other. If a correlation is negative, when one variable increases, the other variable decreases. As an example from this study, a positive correlation indicates that as the daily cost of a barrel of oil increases, the railroad's stock price is expected to increase. In contrast, if the correlation was negative, as the cost of a barrel of oil increased, the stock price would decrease.

One more important feature of the correlation is that it can show the strength of the relationship between the two variables. This is determined by the absolute value of the correlation coefficient with higher values indicating a stronger relationship and values nearer to zero indicating a lack of relationship. For this study, the paired correlations for each of the Class I Railroad stock prices and the daily cost of a barrel of oil were calculated using Equation No. 1 and done in Excel (see Table 1).

Results

The