British Petroleum Linear and Multiple Linear Forecast for Q1 2015
Aubrey Martin
West Virginia State University

British Petroleum (BP) is one of the top oil producing companies in the world. They are responsible for all aspects of the business, from exploring to find new underground oil supplies, to marketing and selling the finished product and everything between. Because of BP’s wide array of operations, there are many different drivers of their business. BP’s main operation is sales of refined oil, where it earns about 70.1% of its revenues, while another 24.7% of its revenues come from the sale of crude oil. The other small percentage of their revenues come mainly from buying/selling other companies, joint ventures, and interest earned on investments.
The goal of this study is to run linear and multiple linear analyses on the data collected to predict first quarter 2015 results for net revenues, operating costs, and profits of BP. To establish the main drivers of revenue, operating expenses, and net profit, I used simple percentages and correlations. Most of the items in the research were extracted from BP’s quarterly financial statements from 2011-2015 (2015), with the exception of quarterly average Brent Crude oil prices during the same time period, which were pulled from Yahoo Finance (2015). By using simple percentages, I was able to see which line items from the income statement were responsible for the majority (in USD) of each category during the timeframe of the study. Brent Crude Oil prices were, of course, not on the financial statements of BP, but I assumed that they would play a major role in determining revenues for an oil-based company, so I elected to use the prices as a variable. Next, I ran correlation studies to determine how strongly correlated my selected variables were to their respective categories of study (net revenues, net operating costs, and net profit.) I determined that financing costs would be the best driver of operational costs because it had the strongest negative correlation with expenses of any item in the expenses category, along with the second highest percentage of total dollar costs. For net profit, I again chose Brent Crude Oil price per barrel as the variable that predicted profit the best.
After running my simple linear analysis, I used the linear function (y=a+bx) to forecast forward to Q1 2015 using the graphs and formulas created in Microsoft Excel. My results from simple linear forecasting were as follows: * Net Revenues: * Forecast – $52,459 million USD * Actual – $54,920 million USD * Forecast Error – $2461 million USD; 4% * Operating Expenses: * Forecast – $78,134 million USD * Actual - $52,318 million USD * Forecast Error - $25,816 million USD; -49% * Net Profit/Loss * Forecast – $12,195 million USD * Actual - $2602 million USD * Forecast Error – $9,593 million USD; -369%
The results were surprising to say the least. I determined that oil prices did seem to predict revenues well, but financing expenses did not predict operating costs, and oil prices did not predict net profit. To determine whether another variable may help these variables to predict revenues, operating costs, and profits, I ran a multiple linear analysis. In this analysis, I added downstream sales revenue to further predict total sales revenue, trade to predict additional operating costs, and taxation to predict net profit. Taxation, while always a factor in any company’s net profit, has a large effect on BP’s profit because it paid less than 40% taxes in 14/16 quarters studied, which allow its profitability to remain high. After running the same tests, this time with the additional variables as well as the original, the results were as follows: * Net Revenues: * Forecast – $91,029 million USD * Actual – $54,920 million USD * Forecast Error – $36,109 million USD; -66% * Operating Expenses: * Forecast – $90,896 million USD * Actual - $52,318 million USD * Forecast Error - $38,578 million USD; -74% * Net Profit/Loss * Forecast – $609 million USD * Actual - $2602 million USD * Forecast Error – $1,993 million USD; 77%
Clearly, my variables were not great predictors of BP’s performance. The estimates in net revenues and operating costs were much farther from the actual values than they were with simple linear forecasting. However, net profit’s estimation was much closer as expected, due to including taxation, which changes wildly from quarter to quarter.
My results were most likely skewed by the rapid swings in oil price per barrel, which I determined is the best predictor of BP’s revenue stream. During the last quarter of collected data, along with the predicted quarter, oil prices plummeted, which lowered revenues by almost 40%. The decline in oil prices, along with excess supply of oil, caused BP to slow operations, which is the explanation for both revenue and operating expenses declining. Since the estimates were based on linear trends from previous quarters, I believe the data was skewed and had no chance to be correct using linear methods. Had a forecaster been present to use judgment on economic conditions and oil prices in the upcoming quarter, the estimates would have been closer to actual results. Because the variables that could further explain BP’s revenues, expenses, and profit are all qualitative, a dummy variable would not help in the forecast.
Overall, I believe that linear forecasting is not a great way to forecast. There are simply too many variables outside of the company’s control that can affect its performance. Because linear forecasting does not take these variables into account, the potential for error is high, as demonstrated in my results. Oil prices are decided by simple supply and demand. BP cannot control demand for oil, thus they cannot control oil prices. This econometric factor is just one of many potential errors in this forecasting process.

References
BP Misses Earnings Estimates in Q4, Revenues Decline Y/Y. (2015, February 3). Retrieved
November 11, 2015, from http://www.zacks.com/stock/news/162839/bp-misses-earnings- estimates-in-q4-revenues-decline-yy
BZX15.NYM Interactive Stock Chart | Yahoo! Inc. Stock - Yahoo! Finance. (2015). Retrieved November 6, 2015, from http://finance.yahoo.com/echarts?s=BZX15.NYM Interactive#{"allowChartStacking":true}
Will SandRidge's Gulf of Mexico Asset Sale Pay Off? (2013). Retrieved November 7, 2015, from http://www.fool.com/investing/general/2014/01/26/will-sandridges-gulf-of-mexico- asset-sale-pay-off.aspx (2015, April 28). Retrieved November 7, 2015, from http://www.bp.com/content/dam/bp/pdf/investors/bp-first-quarter-2015-results.pdf