Rack sellers can purchase bulk petroleum diesel and BBD and blend them, then sell the blended fuel to retailers at the rack. Retailers, or rack buyers, can also purchase bulk petroleum diesel and BBD above the rack and store and blend the fuels at the retail station.For both the tax and subsidy pass through analyses, I utilize daily spot prices for ULSD, jet fuel, and bio-diesel in the major U.S. spot markets purchased from OPIS. I obtained spot prices for ultra-low sulfur diesel and jet fuel from the following markets: Los Angeles, San Francisco, Pacific Northwest, Gulf Coast, Group 3, and New York Harbor Barge. Three of those spot markets – Chicago, NY Harbor, and Gulf Coast – also post spot prices for soy methyl ester B100. Spot markets only exist for SME B100 because, as mentioned in Section 1.1.1, soybean oil makes up the majority of bio-diesel used in the U.S. for RFS compliance. The sample includes observations from each spot market covering business days from 1/1/ 2015 to 7/31/ 2021. Table 1 summarizes the sample of spot prices. ULSD spot prices range from $1.63/gal to $1.75/gal on average over the entire sample. Jet fuel spot prices are slightly lower than ULSD,vertical grow system ranging between $1.57/gal and $1.68/gal on average. SME B100 spot prices are almost twice as large as ULSD on average. Section 1.4 shows that ULSD prices exceed the jet fuel price by the amount of the RIN tax on ULSD, on average.
Section 1.5 shows that B100 prices exceed ULSD prices by the amount of the RIN subsidy for B100.For our analysis we purchased several types of pricing data from OPIS. This data includes supplier-level, daily prices paid for blended diesel fuels at rack terminals in six cities: Los Angeles, CA; San Francisco, CA; Bakersfield, CA; Stockton, CA; Portland, OR; Eugene, OR; Wood River, IL; Trenton, NJ; Dallas, TX; St. Louis, MO. Rack prices are differentiated by the blend level and feedstock type. Biodiesel is assigned one of three feedstock types in the OPIS data: soy methyl ester , yellow grease methyl ester , and a mix of multiple feedstocks . To make the analysis internally consistent, I only utilize rack prices for SME bio-diesel blends, since B100 spot prices are only available for SME-based bio-diesel . Suppliers enter and exit the sample in all cities. For this reason, I aggregate to city-blend averages, instead of using the supplier-level data. I only include suppliers with consistent pricing over at least a one-year period. Therefore, the unit of observation when studying rack margins is a city-blend-day. This process is done separately for branded and unbranded products. For each rack market, the number of suppliers is listed in Table A- 2. As shown in Figure 6 the implicit taxes and subsidies have varied significantly over the sample. Rack spreads also vary over time, and across markets and blends. Additionally, different cities have different blend offerings . Table 2 summarizes the sample of rack spreads in the sub-samples outlined in Figure 6 and the full sample. A general trend is that rack margins are negative for many blends and inversely correlated with the percentage of bio-diesel in the blend.
They are also lower in sub-samples where the RIN subsidy is higher. This reflects the fact that RIN subsidies are being fully passed through at the rack, because suppliers can sell more RINs for a higher blend, and therefore, can accept a lower price. In California and Oregon, this trend does not always hold, and margins are generally much higher than the rest of the country. Higher margins in California and Oregon are consistent with incomplete pass through of LCFS and CFP credits. The pass through of RIN, LCFS credit, and CFP credit subsidies is discussed in detail in Section 1.5.Refiners and petroleum importers are the obligated parties are the obligated parties in each policy, meaning they must purchase RINs and LCFS credits proportional to their output which act as implicit taxes. Refiners sell ULSD in the major spot markets, meaning the implicit tax created from their RIN obligation will be embedded in spot market prices to the extent refiners pass through the cost. In this section, I test whether obligated parties fully pass through the implicit taxes created by the RFS and LCFS in spot market transactions. I extend the methodology developed in Knittel et al. analyzing spreads between ULSD and jet fuel spot prices. ULSD and jet fuel are nearly identical products, so their prices tend to move together. 19 Figure A- 2 shows daily ULSD and jet fuel spot prices since 2015 in the major spot markets. The only substantive difference between the two is their policy treatment: ULSD has a RIN obligation and generates LCFS/CFP deficits in California and Oregon, whereas jet fuel does not. Other federal and state taxes may differ between the two fuels but are assessed below the rack, so those differences won’t be captured in the spot prices. Additionally, spreads are calculated within spot markets, so transportation costs will not differ between ULSD and jet fuel.
The spread between ULSD and jet fuel may capture fluctuations other than changes in the RIN tax, but I assume that any differences are constant over time. This provides a clean setting where one of two identical products are treated. Figure 7 plots the daily spreads between ULSD and jet fuel prices along with the RIN tax, and where applicable, the LCFS/CFP tax. The RIN tax is calculated according to and the LCFS/CFP taxes are calculated according to . The Gulf Coast and NY Harbor have the most efficient spot markets for ULSD and jet fuel; in those markets it is easy to see that the spread tends to equal the RIN tax, suggesting the tax is fully passed through. Group 3 exhibits a similar pattern, but the spread is much noisier. Spreads in California and the Pacific Northwest are also very noisy, however Figure 7 clearly shows that the spreads generally follow the RIN tax while seemingly unaffected by LCFS and CFP taxes. If the LCFS and CFP taxes were fully passed through, we would see that the sum of the LCFS tax and the RIN tax would equal the ULSD-jet spread in California .The bottom panels of Figure 8 present estimates of the LCFS tax pass-through in Los Angeles and San Francisco, which shows that little to none of the LCFS tax is passed through to ULSD spot prices in California. Using the full sample, a $1/gal increase in the LCFS tax results in a 37 cents/gal increase in the spot price in Los Angeles on average, 16 cents/gal in San Francisco. When dropping the RIN Price Shock Period, which is the preferred specification because spreads are volatile in that period, the coefficients for Los Angeles and San Francisco fall to 0.28 and -0.2, respectively, and pass through is statistically indistinguishable from zero in San Francisco. The 95 percent confidence interval in Los Angeles is [0.03, 0.53] when dropping the RIN Price Shock Period. These results suggest that much of the LCFS tax is not passed through to spot prices but does not suggest that refiners are eating the cost. In the remainder of this section, I present evidence that LCFS deficit obligations are transferred downstream to blenders, and they completely pass through the LCFS tax to rack prices. As described in Section 1.1.2, obligated parties under the LCFS are allowed to transfer their status as the deficit generator to another entity if agreed upon by both parties. If deficit obligations are transferred to a different point in the supply chain,cannabis grow supplies the tax will be priced in at the new point. Additionally, in some cases, refineries also own blending operations which would lead to the same results – the LCFS tax being passed through at the rack rather than the spot market. Blenders may agree to fulfill the deficit obligation as they already trade credits. If refiners exercise this option, blenders will incur the LCFS deficit obligation and pass through the tax to the rack price. There is no rack price for pure ULSD in the OPIS rack pricing data. However, there are rack prices for B0-5 – ULSD that may contain up to 5 percent bio-diesel by volume but doesn’t earn credits because the amount of bio-diesel is unidentifiable. This is essentially ULSD sold at the rack. I calculate the margin at the rack for B0-5 in city ? on day ? using the previous day’s nearest spot price of ULSD.Figure 9 plots the daily rack margins, calculated according to , for both branded and unbranded ULSD in Los Angeles – the only city in California or Oregon with consistent reporting of B0-5 rack prices. Prior to April 2020, unbranded ULSD rack prices were higher than the spot price by the amount of the LCFS tax on average. The branded ULSD rack margin is less noisy than its unbranded counterpart but is slightly marked-up above the amount of the LCFS tax.
This is consistent with the fact that branded rack buyers are limited to purchasing diesel that is of their brand and may therefore exhibit a higher willingness to pay than unbranded buyers who have more options .The level of the LCFS tax shifts upward at the start of each year with the annual decline in the diesel CI standard, and the reference diesel CI score often changes annually. The standard has been laid out in advance and changes to reference CI scores are specified by CARB and are not determined in any way by rack margins; therefore, they act as exogenous shifters in this setting. The most striking rise in the tax was on January 1st, 2019. The CI reference score for ULSD fell only slightly from 100.48 to 100.45 in 2019, the diesel CI standard fell from 96.91 to 94.17. The standard decreasing by nearly three full points, coupled with a record-high market price for LCFS credits, led to an average LCFS tax that was, on average, 6.72 cents/gallon higher in January 2019 than December 2018. In the same two-month period, the rack margin for branded ULSD was 6.13 cents/gallon higher in 2019. Rack margins for unbranded ULSD also responded to the change in the tax but later, around mid-March of 2019. It is unclear why the response was delayed, and it only occurred in that year. Results from this model are presented in Table 3. I estimate for both branded and unbranded rack ULSD spreads, yielding pass-through coefficients of 0.24 and 0.4, respectively. However, this result is heavily driven by the downward shock to rack ULSD spreads beginning in midApril 2020, likely due to the COVID-19 pandemic. Therefore, columns 3 and 4 of Table 3 presents results from the preferred specification, which includes a dummy variable for observations after April 15th, 2020. This set of results present evidence that refiners transfer their LCFS deficit obligation downstream and is consistent with complete pass through of the LCFS tax at racks in California.Table 3 also shows the significant and persistent impact that the COVID-19 pandemic had on ULSD rack margins, which were around 20 cents lower on average following mid-March of 2020 than their long-run average since 2015. Evident from Figure 9, since the pandemic started in spring 2020, rack sellers were either unable to continue to pass through the full cost of the LCFS tax at the rack or they were able to pass their obligation even further downstream. The rack pricing data for Portland and Eugene do not include B0-5, so this exercise can’t be done for Oregon and the CFP. However, given the similarities between the Oregon and California programs, and the results from Figure 7 and Figure 8, it is plausible that CFP obligated parties act in similar fashion. First, the RIN tax appears to be fully passed through to ULSD spot prices, consistent with previous studies . ULSD-jet spreads are noisy in most of the major spot markets except for the Gulf Coast, where the long-run pass through coefficient ?# / equals 1.00, and its 95 percent confidence interval is [.84, 1.16]. This implies that refiners have still been able to recoup the cost of RFS compliance, even as RIN prices have reached all-time highs. Second, little to none of the implicit tax from LCFS deficit obligations is passed through to ULSD spot prices because refiners shift the obligation downstream to blenders. However, Figure 8 shows that, in Los Angeles especially, some of the LCFS tax is passed through to the spot price of ULSD. This could occur because some refiners keep the obligation and others don’t, and the coefficients represent an average of This result, qualitatively, raises interesting questions. How do these agreements work? Does this occur because refiners have more bargaining power than blenders? These questions are outside the scope of this paper but present further research opportunities. Third, blenders have passed through the full LCFS tax to rack buyers for both branded and unbranded ULSD. Together, these findings suggest that the tax side of these policies operate as intended. That is, raising the price of petroleum for blenders and, ultimately, consumers.