Gas Prices Understanding the Link

Gas Prices

UNDERSTANDING the LINK BETWEEN AMERICAN TROOP DEATHS and U.S. REFINED OIL PRICES

Experts have tried to connect skyrocketing gasoline prices in America with the current war in Iraq. The war, some argue, causes concerns about regional stability and future oil prices, and this impacts pricing. By this logic, the price of refined petroleum should increase or decrease, depending on the intensity of the war. However, a side-by-side analysis of U.S. refined oil prices and American troop deaths, over an 18-month period from the base month of March 2006 to August 2007, shows almost no correlation. No relationship is established even when a time lag is introduced to allow oil markets time to react to troop deaths.

Tony Bautista

Econ 2, Macro

TTh 7-10pm

Nov. 27, 2007

Introduction

With gasoline prices climbing to near-record levels in some parts of America, there has been a rush to connect the current war in Iraq with the pain consumers are feeling at the pump. Spikes in gas prices were predicted before the onset of the war, and academic research has speculated that oil prices are approximately 20% higher because of the Iraq conflict, largely because of concerns about regional stability and worries about future supplies (Hall, 2006 and Kopytoff, 2002). But if such concerns can spark higher oil prices, one would expect oil -- and, subsequently, gas prices -- to increase and decrease as the war intensifies or as violence lessens. This link is easy enough to examine if we look at American refined oil prices for an 18-month period (from March 2006 to August 2007, which is the last month for which the U.S. Department of Energy has data) against American troop casualties in the same months. Refined petroleum prices are a good benchmark, as they allow us to factor out various state and federal taxes that can cause significant differences regionally in gasoline costs. American troop deaths are only one yardstick, as they don't count Iraqi deaths -- but they are still useful, as they are a typical gauge for determining how the war effort is proceeding. For this paper, I will analyze percentage increases or decreases in American troop deaths against movement in petroleum oil prices during the same months. Then, I will introduce a one- and two-month lag to allow petroleum markets to react to troop deaths. This analysis shows that increases or decreases in troop deaths typically do not have the same effect on oil prices, and may even have a reverse effect.

The immediate effect

Excluding the base month of March 2006, there were 17 months analyzed. Refined petroleum prices and coalition deaths only moved in the same direction in eight of the 17 months studied (See Table 1 and Graph 1). In other words, if coalition deaths increased during a particular month, there was a slightly less than 50% chance that refined petroleum prices also increased. It is difficult to argue that there is a definite and immediate relationship between coalition deaths and petroleum prices when, more often than not, the figures do not trend in the same direction.

Interesting enough, there is a stronger argument for a link when we isolate only the months during which the greatest percentage change in coalition deaths occurred. If we isolate the three months with the largest increases (145.2% in April 2006; 60% in December 2006; and 51.2% in August 2006), we find that in April and December, refined prices also increased, and their decrease was only slight in August 2006. So, perhaps, once we cross a certain threshold of deaths, gas prices can be impacted. To test this, we can isolate only the months when there was at least a 20% change -- increase or decrease -- in coalition deaths, compared to the previous month. Out of the 10 months that experienced at least a 20% change, in six cases petroleum prices trended in the same direction. If we isolate months during which at least a 30% change occurred, in only two of the five months did petroleum prices trend in the same direction -- our link actually gets worse.

In short, it is difficult to argue that there is any immediate, monthly correlation between coalition deaths and refined petroleum prices, even when significant increases or decreases of coalition deaths occur.

One- and two-month lags

When we consider time lags between American deaths and refined petroleum prices, the case for a link does not become any stronger. A one-month lag leaves us with 16 months to analyze, as we must exclude the base month (March 2006) and the final month (August 2007). In only six of the 16 months studied did petroleum prices move in the same direction as troop casualties, a month after those casualties occurred (See Table 1 and Graph 1). When we consider the three months with the largest increases (145.2% in April 2006; 60% in December 2006; and 51.2% in August 2006), only once (May 2006) did the following month post increased refined petroleum prices. If we consider all months with at least 20% movement in troop casualties, in only four of the 10 months did refined petroleum prices move in the same direction the following month. Isolating months with at least 30% movement in troop casualties shows that in only one case of five did refined petroleum prices increase the following month. In short, the link gets worse as the percentage increase or decrease in casualties moves significantly.

By analyzing a two-month lag, we are left with 15 data points. In only six of the 15 months did refined petroleum prices move, two months later, in the same direction as troop casualty figures (See Table 1 and Graph 1). The three months with the largest increases in troop casualties (145.2% in April 2006; 60% in December 2006; and 51.2% in August 2006), were only followed two months later by an increase in refined petroleum prices once -- in February 2007. Of the nine months analyzed where at least a 20% change in troop deaths occurred, only four times was there similar movement in refined petroleum prices two months later. Interesting enough, in three of the five months where at least a 30% change in troop deaths occurred, refined petroleum prices moved in the same direction two months later. While this allows for an argument that refined petroleum prices, with a two-month lag, can be influenced by movements in troop deaths that exceed a certain trigger point, the argument is far from convincing. After all, the overall correlation between troop deaths and refined petroleum prices after a two-month lag is weak, and at the 20% trigger mark, refined petroleum prices moved in sync less than half the time.

Conclusion

The link between American troop deaths in Iraq and refined petroleum prices is weak at best, and arguably non-existent, even when we consider time lags. There is some evidence that, without time lags, there can be some relationship between troop deaths and refined petroleum prices during the months in which those deaths occurred. The majority of the time when there was a 20% movement in troop deaths, compared to the previous month, refined petroleum prices moved in the same direction. This relationship does not hold up once we cross the 30% barrier, but the sample size is smaller and arguably less relevant.