A Protocol to Assess the Impact of Crude Oil and Fuel Price Fluctuations on Future Asphalt Prices in the State of Alabama: A Stochastic Risk Assessment Approach
Type of DegreeMaster's Thesis
MetadataShow full item record
Since the introduction of asphalt in the year 1870, it has become one of the main components of horizontal construction. According to the National Asphalt Paving Association (NAPA), asphalt covers approximately 94 percent of the 2.7 million miles of paved roads in the United States (U.S.), and it contributes approximately $80 billion dollars to the horizontal construction industry annually. With bitumen, a byproduct of crude oil, as a critical ingredient of asphalt, the projected price of asphalt has been based on an assumed relationship with crude oil and fuel. Multiple publications, researchers, estimators, engineers, and other proponents of the horizontal construction industry have mentioned this correlation, yet there is a lack of research defining these items as dependent on each other. Using crude oil and fuel price indexes, this thesis researches the long held hypothesis of their relationship with asphalt price indexes. Based on the monthly-posted asphalt, crude oil, and fuel price indexes by the Alabama Department of Transportation (ALDOT) and U.S Energy Information Administration (EIA), a stochastic risk assessment tool using a modified cumulative sum (CUSUM) statistical analysis method was created to determine their relationship. The risk assessment tool enables the prediction of the time line of asphalt price changes along with the magnitude of the changes. Risks associated with those changes based on relationships with crude oil and fuel were answered. In terms of using crude oil price indexes to predict asphalt price index changes, the results indicated that the most likely time gap between them was 3 months and the percent change was 58 percent. This means that for a 1 percent change in the crude oil price index, 58 percent of that change would most likely be reflected in asphalt’s price index 3 months later. In terms of using fuel to predict asphalt’s price changes, the results indicated that the most likely time gap between them was 2 months and the percent change was 46 percent. This means that for any 1 percent change in the fuel’s price index, 46 percent of that change would most likely be reflected in asphalt’s price index two months later. In comparison with previous studies on this topic, revealed that the methods used to predict changes in asphalt price indexes in this thesis were significantly more effective (p = 0.05) than at least one of the previous studies. Likewise, a cross-validation process conducted on the deterministic results revealed a 2.2 and 3.9 mean average percentage error (MAPE) in the estimation of future asphalt prices using observed crude oil and fuel price fluctuations. Although the time gap and percent change ratios between crude oil and fuel price indexes showed to be significantly more accurate at estimating future asphalt price indexes (p = 0.05), both MAPE values could be considered to reflect high accuracy levels when compared against the expected cost estimating errors determined by the America Association of State Highway and Transportation Officials (AASHTO). According to AASHTO, expected accuracy in final design estimates is expected to be between -5 percent and +10 percent. This study meets these requirements with MAPE values under +4 percent, in a commodity (asphalt) that sometimes represents over 80 percent of the total project construction cost. Recognizing the unavoidable variability in the time gaps and percent change ratios mentioned above, this study conducted a stochastic analysis to create a risk assessment tool. This tool would allow estimators to make better-informed decisions by providing them with the probability of occurrence of different case scenarios in terms of potential time gaps and percent change ratios. For example, although the most likely results were 3 months and 58 percent between crude oil and asphalt price indexes, the stochastic analysis revealed that there was only a 14 percent probability of having a percent change ratio between 40 and 60 percent with a 3 month time gap. Through updated processes and refined methods of cost estimating, the findings within this thesis are expected to improve cost estimating effectiveness in asphalt paving projects. In any project-oriented organization, such as state transportation agencies (STAs), an improvement in cost estimating effectiveness is reflected in a better allocation of available resources, which in the case of STAs, are continuously shrinking to address all transportation infrastructure needs. Although the results and risk assessment tools presented in this study are applied to assess the future asphalt price indexes in the state of Alabama, the methodology presented throughout this thesis could be replicated for other STAs with local price indexes. In conclusion, the study presented in this thesis has the potential to improve STAs’ resource allocation practices through a better understanding of the factors triggering price index fluctuations in the most critical commodity used by STAs, asphalt.