Analyzing the Risk of Transporting Crude Oil by Rail Motivation Incidents Rail traffic Empirics Summary Analyzing the Risk of Transporting Crude Oil by Rail Charles F. Mason H.A. True Chair in Petroleum and Natural Gas Economics Department of Economics & Finance University of Wyoming Laramie, Wyoming 7 January, 2017 ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

July 6, 2013: Lac-Me´ gantic, Quebec Motivation Incidents Rail traffic Empirics Summary July 6, 2013: Lac-Me´ gantic, Quebec ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

December 30, 2013: Casselton, North Dakota Motivation Incidents Rail traffic Empirics Summary December 30, 2013: Casselton, North Dakota ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

April 30, 2014: Lynchburg, Virginia Motivation Incidents Rail traffic Empirics Summary April 30, 2014: Lynchburg, Virginia ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

February 16, 2015: Mount Carbon, West Virginia Motivation Incidents Rail traffic Empirics Summary February 16, 2015: Mount Carbon, West Virginia ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

March 5, 2015: Galena, IL Motivation Incidents Rail traffic Empirics Summary March 5, 2015: Galena, IL ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Pushback Motivation Incidents Rail traffic Empirics Summary ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Pushback Motivation Incidents Rail traffic Empirics Summary ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Serious Incidents Originating State Frequency Percent Colorado 1 4.35 Motivation Incidents Rail traffic Empirics Summary Serious Incidents Originating State Frequency Percent Colorado 1 4.35 Delaware Kansas Montana 2 8.7 New Mexico North Dakota 15 65.22 Texas Wyoming Total 23 100 ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Crude Oil Rail Incidents Motivation Incidents Rail traffic Empirics Summary Crude Oil Rail Incidents A. Serious Incidents Fraction of weeks with an event 0.07 Number of weeks between events Mean Std. Dev. Median Skewness 13.23 20.34 6.50 3.18 B. Minor Incidents Fraction of weeks with an event 0.50 Number of events per week Mean Std. Dev. Median Skewness 2.27 1.72 2.00 2.08 ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017 10 / 23

Major Incidents: Quantity of Oil Spilled and Economic Motivation Incidents Rail traffic Empirics Summary Major Incidents: Quantity of Oil Spilled and Economic Damages 15000 150 Thousands of Barrels 10000 Millions of US Dollars 100 5000 50 2009m1 2010m1 2011m1 2012m1 2013m1 2014m1 2015m1 month amount of crude oil released in major incidents total economic damages in major incidents ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Minor Incidents and Time Between Major Incidents Motivation Incidents Rail traffic Empirics Summary Minor Incidents and Time Between Major Incidents 100 2009w26 2010w26 2011w26 2012w27 2013w26 2014w26 week 10 weeks between serious oil spills 80 minor oil spills per week 5 60 40 20 2010w26 2011w26 2012w27 week 2013w26 2014w26 ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Minor Rail Incidents vs. Rail Oil Traffic Motivation Incidents Rail traffic Empirics Summary Minor Rail Incidents vs. Rail Oil Traffic 2009m1 2010m1 2011m1 2012m1 2013m1 2014m1 2015m1 ym monthly number of minor incidents monthly oil car shipped by rail 15 20 monthly number of minor incidents monthly oil car shipped by rail 15 10 10 5 5 ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Rail Oil Traffic: Small Shipping States Motivation Incidents Rail traffic Empirics Summary Rail Oil Traffic: Small Shipping States 250 200 monthly rail cars 150 100 50 2009m1 2010m1 2011m1 2012m1 month 2013m1 2014m1 ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Rail Oil Traffic: Large Shipping States Motivation Incidents Rail traffic Empirics Summary Rail Oil Traffic: Large Shipping States 2000 1500 monthly rail cars 1000 500 2008m7 2010m1 2011m7 month 2013m1 2014m7 ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Rail Oil Traffic: North Dakota Motivation Incidents Rail traffic Empirics Summary Rail Oil Traffic: North Dakota 15000 10000 monthly rail cars 5000 2008m7 2010m1 2011m7 month 2013m1 2014m7 ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Annual Crude Oil Shipments Motivation Incidents Rail traffic Empirics Summary Annual Crude Oil Shipments State 2009 2010 2011 Year 2012 2013 2014 Total IL 1 3 19 22 91 136 WY 11 53 65 191 320 TX 10 14 118 264 286 163 855 CO 21 4 26 111 166 328 ND 86 205 417 1080 1447 1403 4638 Other 50 74 112 577 494 1627 Note: Trains delivering crude oil, listed by originating state. ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Average Number of Cars Carrying Crude Oil, by Year Motivation Incidents Rail traffic Empirics Summary Average Number of Cars Carrying Crude Oil, by Year State 2009 2010 2011 Year 2012 2013 2014 Total IL 1 2 81.37 34.27 90.18 77.30 WY 4.23 12.46 65.18 42.17 TX 22.17 22.83 22.74 13.94 20.40 CO 2.90 29.75 32.46 23.33 83.33 53.19 ND 4.02 43.19 29.31 59.70 92.68 100.39 77.80 Other 10.50 9.26 7.53 4.40 5.50 18.92 9.82 Note: Average number of cars carrying crude oil per train. Listed by originating state. ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Data � Incidents � Rail Traffic � merged these sets ✄ PHMSA reports Motivation Incidents Rail traffic Empirics Summary Data � Incidents ✄ PHMSA reports any (self-reported) “incident” (restrict to crude oil) can be minor (common) or serious (infrequent) observations collected for 1 Jan 2009 – 31 Dec 2014 info on amount oil released, total econ. damage, originating state � Rail Traffic ✄ DOT waybill sample most large carrier shipments detailed information on every shipment, 2009 - 2014 retained all shipments carrying oil, originating in US � merged these sets ✄ aggregated to monthly observations “obs.”: number of oil cars shipped in month t from a state k no. incidents (0 – 8) no. serious incidents (0/1) amt. oil spilled; total costs ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Time Between Serious Incidents Motivation Incidents Rail traffic Empirics Summary Time Between Serious Incidents Regression model regressor Cox Exponential Weibull Cumulative number of minor incidents -0.025∗∗∗ (0.009) -0.019∗ (0.011) -0.019∗ (0.010) constant -2.255∗∗∗ (0.504) -1.956∗∗∗ (0.272) p 0.905 (0.146) χ2 statistic 7.644∗∗∗ 3.101∗ 3.599∗ Standard errors in parentheses *: significant at 10%; **: significant at 5%; ***: significant at 1% ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017 20 / 23

χ2 Rail Car Shipments and Minor Incidents Poisson Negative Binomial Motivation Incidents Rail traffic Empirics Summary Rail Car Shipments and Minor Incidents Poisson Negative Binomial (1) (2) (3) (4) Thousand cars 0.205∗∗∗ (0.014) 0.136∗∗∗ (0.006) 0.236∗∗∗ (0.018) 0.154∗∗∗ (0.024) constant -1.333∗∗∗ -1.378∗∗∗ -0.387 (0.104) (0.257) State-level FE? no yes N 681 562 χ2 229.0 442.7 167.1 42.7 Standard errors in parentheses *: significant at 10%; **: significant at 5%; ***: significant at 1% ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

χ2 Rail Shipments and (a) Oil Spilled, (b) Total Damages Motivation Incidents Rail traffic Empirics Summary Rail Shipments and (a) Oil Spilled, (b) Total Damages Dep. Vbl.: (a) Quantity of Oil Spilled (b) Total Economic Damages Poisson (1) Negative Binomial (2) Poisson Negative Binomial (3) (4) Thousand cars 0.026∗∗∗ (0.007) 0.137∗∗∗ (0.021) 0.215∗∗∗ (0.004) 0.227∗∗∗ (0.022) constant -2.333∗∗∗ (0.126) yes -4.183∗∗∗ (0.122) yes State-level FE? yes yes N χ2 562 16.03 562 42.80 539 2587 539 102.3 Standard errors in parentheses *: significant at 10%; **: significant at 5%; ***: significant at 1% ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017

Motivation Incidents Rail traffic Empirics Summary Conclusion � statistically important, negative relation b/w accumulated minor incidents and time between serious events � statistically important, positive rel’n b/w rail traffic and pdf over minor incidents ✄ adding 10,000 rail cars shipping oil ⇒ .4 add’n’l incidents / week � fixed effects largest for states with significant tight oil production ✄ OK, ND, TX, NM, WY � statistically important positive rel’n b/w rail traffic and pdf over costs ✄ implies impact on expected costs: marginal impact of one-unit increase in rail shipments = $725 ✄ costs reported in database include lost product and damaged capital (private costs) costs from response, closure of main transportation arteries (social costs) ✄ costs do not include social costs associated with environmental damages from oil spills property damages resulting from serious events (e.g., spill-induced fires) value of lost life ASSA/IAEE (C. Mason) Oil by Rail 7 January, 2017