Data and Methodology: Data

The literature review provided a qualitative basis for exploring the relationship between the energy sector and the development of Russian sanctions in the sample period 2012–2018. Domestic sanctions, as discussed in the Literature Review, are generally unilateral in nature and are most impactful in their duration when implemented by Acts of Congress. As such, the sample set for the present study are the five Acts of Congress passed between 2012–2018 that targeted the Russian Federation: H.R.6156; S.2124, H.R.4152; S.2828, H.R.5859; H.R.3364; and H.R.5040/H.R.5515 (full references can be found here). This set offered an exploration of person-vote analysis on sanctions that do not explicitly target the Federation’s energy sector (two: H.R.6156 and H.R.5040/H.R.5515) and those that do (three: S.2124, H.R.4152; S.2828, H.R. 5859; and H.R.3364).

The use of person-vote analysis permitted understanding of each individual piece of legislation as occurring in isolation. This was critical, as the limited number of samples in the sample period did not permit inclusion of existing voting trends for each congressional representative over their career. The purpose of the present study was to merely examine what relationship existed between the energy sector and the development of the five congressional Russian sanctions. The period of study permitted this to occur against the backdrop of changing political control of the executive and legislative branches, offering the ability to test for amplification of support based on party affiliation in congressional chambers and with the president, the chief executor of U.S. foreign policy.

Data

The data for the present study was hand-collected from several different repositories to explore the personal financial holdings, campaign contributions, and constituency stake in the energy sector as defined as financial interests associated with upstream activities and their industry associations. Upstream activities included the exploration and extraction of petroleum and natural gas resources onshore and offshore, supporting these activities, financing these activities, or manufacturing or providing machinery for these activities. The LNG industry was also included in this definition due to its geopolitical weight as discussed in the Energy Policy section of the literature review. This delimitation was added in order to limit the scope of analysis to those firms and industry associations that would most likely have financial interests in maintaining the status quo with the Russian Federation’s energy sector given the historical track record of U.S. energy sanctions.

Voting records for the full House of Representatives and Senate for each of the five sanctions were obtained through Congress’s online legislation database. Person-vote observations for the sample set were weighted on a spectrum, with a Nay vote yielding a value of -1, a Not Voting observation yielding a value of 0, and a Yay vote yielding a value of +1. To bias against ad populum voting habits, all person-vote records for each sanction in the sample set were additionally recorded relative to the full voting record of the individual chamber and the aggregate voting record of both chambers. In these calculations, Not Voting observations were ignored as missing data points. The specific formula for weighting the person-vote records relative to the individual chamber and the combined voting record of both chambers is provided in the Data Statistics section.

Of the five sanctions, only one, the Ukraine Freedom Support Act (S.2828, H.R.5859), was not passed with a roll call vote, which provides specific voting decisions for each voting representative. Instead, the Ukraine Freedom Support Act was passed through unanimous consent in both houses, which is a voting procedure used to expedite proceedings. All person-vote records for the Ukraine Freedom Support Act are thus listed as Yay (1) values. The inclusion of these records does cause concern for data bias because specific person-vote behavior cannot be observed. However, the use of unanimous consent for this piece of legislation could have been challenged and halted by any member of either chamber should the text have been of concern. The Ukraine Freedom Support Act of 2014 is the first sanction placed on Moscow that specifically targeted the Federation’s energy sector since the Reagan administration’s Yamal pipeline embargo, including any foreign parties engaged in upstream activities related to offshore development and the development of the Federation’s shale industry (2014).

No member of either chamber chose to challenge the use of unanimous consent for this legislation. In fact, in the House of Representatives, congressional representatives actively argued to unanimous consent passage with Representative Engel (D, NY-16) stating that the expedited passage of the legislation was necessary in order to signal a decisive U.S. response to President Putin’s “blatant disregard of international law,” with Representative Kaptur (D, OH-9) concurring that expedited action was necessary to respond to the Federation’s actions against Ukraine, NATO allies, and other European states because “liberty cannot bow to Putin’s aggression” (“Ukraine Freedom Support Act of 2014,” p. 10305). Because no action was taken to halt the unanimous consent approval procedure, the person-vote records for the Ukraine Freedom Support Act were included in the present study as a complete and substantive data observation.

In total, 328 person-vote observations were observed during the sample period from 158 different congressional representatives: 209 observations from 104 representatives were observed in the House of Representatives, and 119 observations from 55 senators were observed in the Senate. These congressional representatives were chosen as they were either the authors or cosponsors of the sanctions legislation that passed either chamber en route to the final votes as analyzed in the present study, or they were the (co-/vice)chairs or ranking members of congressional (sub)committees most associated with the energy sector, U.S. foreign policy toward the Russian Federation, or U.S. trade relations.

The chosen congressional representatives would likely have had the most concrete connections to the energy sector as they (1) led oversight activities relating to the sector or U.S. trade policy, (2) engaged in oversight activities related to U.S. foreign policy, or (3) authored or cosponsored the legislation that directly targeted a geopolitical energy market competitor. Additionally, many of the leaders of the congressional (sub)committees likely had input on the advancement of the pieces of legislation, given their respective oversight responsibilities. The only exception to this delimitation is Senator Walsh (D-MT), who was included in the sample set as he was the representative who petitioned the Presiding Officer to move forward with the unanimous consent procedure in the Senate for the Ukraine Freedom Support Act (“Imposing sanctions with respect to the Russian Federation and providing additional assistance to Ukraine,” 2014, p. 6819). The (sub)committee leadership rosters from both chambers were determined from the Clerk of the House of Representatives and the Secretary of the Senate.

Hypothesis 1

Data gathered for Hypothesis 1 (personal finances) came from the Center for Responsive Politics, the Office of the Clerk of the House of Representatives (Office of the Clerk), the Tarrant County Clerk, and the Senate Office of Public Records. This data was compiled primarily from congressional personal financial disclosure (PFD) forms covering the years in which the sanction legislation was passed (2012, 2014, 2017, and 2018), which are readily available through the House of Representatives and the Senate. These forms provide mandated information pertaining to each congressional representative’s personal finances and vary between the congressional chambers on how and what is reported. According the Office of the Clerk (n.d.), representatives must disclose information pertaining to:

● Assets that are valued at greater than US$1,000 by the end of the reporting period and any unearned income exceeding US$200. Unearned income is defined as including dividends, capital gains, and royalties.

● Transactions in the reporting period with values greater than US$1,000. Individual transactions with values less than US$1,000 are not mandated to be reported even if the total value of multiple transactions meets or exceeds the US$1,000 single-transaction threshold.

● Earned income beyond official duties that is in excess of US$200 in aggregate from a single source as well as any honoraria paid to spouses or principal staff.

● Liabilities for the representative, their spouse, or their dependent children in excess of US$10,000.

● Positions held outside of official duties that are unpaid. This reporting is limited to exclusively the representative and does not include household information.

● Agreements regarding future employment after departing from the House of Representatives, from a leave of absence, or with a previous employer concerning benefits and ongoing compensation plans.

● Individual gifts received during the reporting period valued in excess of US$390 for the representative, their spouse, or their dependent children.

● Individual travel payments and reimbursements during the reporting period for the representative valued in excess of US$390 that was not for personal or official duties.

● In lieu of honoraria payments valued at US$200 in aggregate from a single source in the reporting period. Only the source and value are required to be reported. The destination charity is not provided on the PFD.

● Compensation in excess of US$5,000 from each of the two immediately preceding calendar years prior to taking office — in the event that the representative held an ownership position.

● Whether or not any shares were purchased in an initial public offering.

● Whether or not a blind trust for the representative’s household was not disclosed on the PFD.

● Whether or not assets or unearned income, liabilities, or transactions for the representative’s spouse or dependent children were not disclosed on the PFD.

The U.S. Senate Select Committee on Ethics (2020) instructs senators to provide the same information as their counterparts in the House of Representatives with the notable exception of the final three questions on the Office of the Clerk’s PFD regarding initial public offerings, the exclusion of a blind trust from the report, or the exclusion of assets, unearned income, liabilities, and transactions from a spouse of a dependent child from the report.

These PFDs are mostly available electronically from the respective websites for the Office of the Clerk and the Senate Office of Public Records. However, with regard to the Senate Office, only PFDs from sitting senators and those departing the Senate no longer than six years prior to the data collection were available for the sample period 2012–2018. The Senate Office cited Section 8(c) of the STOCK Act, which sets out a timeline for maintaining PFDs for all congressional representatives to be six years after the representative is no longer a Member of Congress, as the rationale for this missing data in their collection (2012). However, this did not prove to be the case with the records of the Office of the Clerk, since all PFDs for the representatives in the sample set were accessible regardless if they had exited the House of Representatives longer than six years ago. The missing PFD records for the following senators were obtained through the Center for Responsive Politics:

● Bingaman, Jeff (D-NM)

● Kerry, John (D-MA)

● Lautenberg, Frank (D-NJ)

● Lugar, Richard (R-IN)

The PFDs for Senator Daniel Inouye (D-HI) and Representative Alan Nunnelee (R-MS) were not available for the period of study due to their deaths prior to their offices filing PFDs for the reporting years of 2012 (one data point, Inouye) and 2014 (two data points, Nunnelee) respectively. These data points are recorded as missing data.

In total, 332 PFD records were collected from the Center for Responsive Politics, the Office of the Clerk, and the Senate Office of Public Records. Of these records, 207 were related to members of the House of Representatives, and 105 were related to members of the Senate. A number of these records were collected as amendments to previously filed PFDs gathered from the sample period so as to ensure due-diligence data collection in recording which congressional representatives self-reported personal financial holdings in the energy sector. The weight or proportion of energy interests was not recorded in this data set, only a binary variable of whether the representative had self-reported energy interests (1) or not (0). Self-reported energy interests were examined under the assets and transactions sections of the PFDs and also in addendums provided by some representatives for their mutual funds, trusts, and other portfolios up to the dates of the final votes for the sanctions under analysis.

Due to the binary variable for personal energy interests, only the first example was recorded. Asset holdings were cross-referenced with transaction records to identify if they were sold leading up to the final vote in the given year. Any energy interest activity noted after the date of the final vote for the sanctions under analysis was ignored. If these assets were liquidated prior to the vote, they were ignored. If they were not liquidated prior to the vote, they were included as energy interests. The rationale used for including energy interests in portfolios that were not energy-specific as rendering the binary (1) variable was that if representatives were self-reporting more detailed information than was necessary, they were aware that all of their vested interests in the business community were subject to public scrutiny and were thereby implicitly submitting their finances for analysis.

The most common energy interests identified through the data collection for Hypothesis 1 were two of the supermajors, the ExxonMobil Corporation (Exxon) and the Chevron Corporation (Chevron), and energy-specific mutual funds, such as Energy Select Sector SPDR. Additional interests were identified through financial stakes in Wells Fargo — largely considered to be one of the most energy-exposed publicly traded financing firms in the North American market, with exposure totaling US$13.56 billion at the end of 2019 (2019) — and the Alaska Permanent Fund, the annual universal dividend paid to Alaskan residents from proceeds of the energy sector operating in Alaska. Other public-traded financing firms that have large energy exposure, such as Citi and JPMorgan Chase, were not included due to the delimitation restricting research to Wells Fargo due to its larger risk exposure to North American upstream activities vis-à-vis Citi and JPMorgan Chase.

A unique energy interest identified through the examination of the PFDs were two mineral leases and two warranty deeds related to the personal financial interests of a Texas representative and their spouse in Tarrant County. On the respective representative’s PFD, mineral rights for two properties were disclosed as assets without clarification regarding the type of mineral ownership that the representative held. Research in the Tarrant County Clerk’s database identified that these mineral rights were for hydrocarbon leases, one of which was jointly held with their spouse and the other wholly held by the spouse. These mineral leases corresponded to two warranty deeds ceding conveyances of the mineral rights to both properties over a decade prior to the sample period. The warranty deeds ceded one conveyance jointly to the representative and their spouse and ceded one conveyance solely to the spouse. This exploration into the mineral rights records was necessitated by the ambiguity of the reporting guidelines for mineral rights as assets on PFD records. These mineral rights constituted the only self-reported personal financial energy interests of the respective representative during the sample period, and thus, the exploration was justified with the records included in the data repository for the present study.

Hypothesis 2

Data gathered for Hypothesis 2 (campaign contributions) came from the Federal Elections Commission (FEC) with support from the Center for Responsive Politics. In contrast to the data collection and analysis for Hypothesis 1, the process for Hypothesis 2 was much more straightforward because all data was publicly available from the FEC’s database. The FEC requires that all campaign contributions above a US$50 aggregate per reporting period and election cycle be recorded (n.d.c). Additionally, the FEC limits the amount of funds that can be contributed to a candidate by individuals and PACs per reporting period and election cycle.

Energy interests are not permitted to donate to any candidate as a C-class corporation or an LLC that files as a corporation under the Internal Revenue Service or has publicly tradable shares (FEC, n.d.d); however, they are permitted to do so if they choose to file taxes as a partnership or have not made a determination to file as either a corporation or a partnership (FEC, n.d.d). Therefore, energy interests establish PACs to engage in the civic process. Under 2019 rules, PACs are permitted to donate up to US$5,000 per election if they are classified as a multicandidate PAC and up to US$2,800 per election if they are a non-multicandidate PAC (FEC, n.d.a). Multicandidate or non-multicandidate status is determined by the number of donors and recipients of PAC funds and the length of time that the PAC has been filed with the FEC (FEC, n.d.b).

The vast majority of FEC data came from Schedule A records, the form that details contribution receipt information. In total, 242 Schedule A records were collected from the FEC, 196 related to members of the House of Representatives and 46 related to members of the Senate. The weight or proportion of energy contributions relative to the aggregate was not recorded in this data set, only a binary variable of whether the representative had received campaign contributions from energy interests (1) or not (0). Contribution records were examined from the immediately preceding election, regardless of type (general, primary, or special), up to the final vote for each sanction. The present study focused only on the immediately preceding election and not the immediately preceding FEC filing period because, for most of the sample set, these periods were one and the same.

The outliers to this observation were senators since their terms span a six-year period. The examination of only the immediately preceding election rested on the assumption from the qualitative evidence that contributions to representatives are more impactful during election years than non-election years because representatives need financial support to win their respective elections. This creates a power dynamic in which any interest can increase its financial leverage over a representative following an election, as contributions are a form of a political investment in the representative and their voting habits (Fulmer, et al., 2013).

Schedule A records are available as large Comma Separated Values (CSV) files, and their CSV status permits ease of filtering for specific firms or any specific phrase or morpheme. Because the variable for energy contributions for the present study is binary, only the first example of energy contributions was recorded. The most common energy contributions came from Exxon and from several independent energy firms engaged in shale exploration and production, namely Anadarko Petroleum Corporation, Chesapeake Energy Corporation, and Occidental Petroleum Corporation. The traditional supermajors, aside from Exxon, did not appear as significant campaign contributors. The Chevron Corporation was a contributor to several representatives in the sample set but not to the extent of Exxon. BP plc and Royal Dutch Shell PLC were also rarely identified as among the donors.

Additionally, industry interests such as the American Petroleum Institute and Koch Industries, Inc. (Koch Industries) were common contributors. While Koch Industries is nearly exclusively involved in the midstream and downstream portions of the energy value chain, it was included in the present study as it is one of the largest leaseholders in the upstream Athabasca oil sands, if not the largest, and is well known for expending campaign contributions in the interests of the full energy sector, including upstream sectors (Mufson, S. & J. Eilperin, 2014; Mufson, S., 2014). Furthermore, contributions from Wells Fargo were also included in the analysis due to its large risk exposure to North American upstream activities (2019). While other publicly traded financial institutions also have large energy exposures, Wells Fargo is generally regarded as mostly focused on North American energy interests, while other institutions have a more global portfolio.

Due to the size of Schedule A reports, the Center for Responsible Politics’ Open Secrets database was consulted to streamline the identification of energy campaign contributions in the sample set. This proved to be critical in the identification of all energy contributions since the following four records of campaign contributions were not denoted on Schedule A records:

● Holden, Tim (D, PA-17): October 17, 2012, donation from Chesapeake Energy Corporation FED-PAC;

● Lowenthal, Alan (D, CA-47): June 7, 2018, donation from the Offshore Marine Service Association PAC;

● Rohrabacher, Dana (R, CA-48): October 29, 2018, donation from Chevron Employees PAC; and

● Ros-Lehtinen, Ileana (R, FL-27): May 16, 2012, donation from Chevron Employees PAC.

These contributions were identified as missing through consultation of the Open Secrets database, and subsequent outreach to the Center for Responsive Politics identified these as appearing on Schedule B records for the respective PACs. Schedule B records are itemized disbursements for a specific period for a PAC, and these include contributions made to a candidate. It is unclear why these records were not reported on the Schedule A reports for these representatives. The support of the Center for Responsive Politics staff was very much appreciated in aiding the completion of the data collection for Hypothesis 2.

Hypothesis 3

Data gathered for Hypothesis 3 (constituency stake) came exclusively from the United States Census Bureau (USCB), which provides detailed reports on the U.S. economy. The data collected for the present study to determine constituency stakes in the energy sector came from UCSB data reports for NAICS codes at the state and federal level. While UCSB data is accessible for NAICS codes at the House of Representatives-district level, UCSB data for the following delimited NAICS codes was only accessible at the state level:

● 211: Oil and gas extraction;

● 213111: Drilling oil and gas wells;

● 213112: Support activities for oil and gas operations; and

● 333132: Oil and gas machinery and equipment manufacturing.

Additionally, the most recent data available was through 2017, allowing exploration only for the 2012, 2014, and 2017 person-vote observations. Hypothesis 3 data for the 2018 person-vote observations was thus recorded as missing data. The broader definition of “energy” as used in the data collection for Hypotheses 1 and 2 was not used under this hypothesis as data was not available to determine constituency stakes across the full upstream, financing, and LNG sectors. Notably, exploration and offshore activities are not included in these NAICS code delimitations because these activities are classified under different NAICS codes that include other industries unrelated to energy.

Constituency stake in the energy sector was defined as meeting at least one of the following three criteria: (a) employment levels relative to the state aggregate, (b) total payroll relative to the state aggregate, and (c) the number of firms operating in the state relative to the state aggregate. Ratios for each of these criteria were required in order to establish equilibrium in the analysis of constituency stake in the energy sector as examining absolute size for each criterion biases against diversification and states with smaller economies. Creating ratios that measured the individual criterion against its respective state aggregate thus ensured equality of analysis. Statistical significance was measured as having ratio values greater than that of the national average. For the 2012 and 2014 observation periods, the following 14 states each ranked higher than the national average for all three criteria:

● Alaska;

● Arkansas;

● Colorado;

● Kansas;

● Louisiana;

● Mississippi;

● Montana;

● New Mexico;

● North Dakota;

● Oklahoma;

● Texas;

● Utah;

● West Virginia; and

● Wyoming.

For the 2017 observation period, though, only West Virginia ranked above the national average for the first criterion — employment levels relative to the state aggregate. Additionally, Florida ranked above the national average for the second criterion — total payroll relative to the state aggregate. Neither Florida nor West Virginia ranked above the national average for the third criterion — the number of firms operating in the state relative to the state aggregate. Due to its outlier status of appearing in only one of the criterion lists in only one of the observation periods, Florida was not included in the sample set as a state with constituency stake in the energy sector. West Virginia was included in the 2017 observation period as a state with constituency stake in the energy sector, despite ranking in only one of the criteria (a), due to the rationale that employment level is a more critical method of determining the importance of a sector to the overall state economy and the constituents than payroll levels and the number of firms in operation. It was not expected that this inclusion would bias the results of the present study since West Virginia was included in the two previous observation periods.

The results of this data were included in the sample set as a binary variable of whether the representative comes from a state with constituency stake in the energy sector (1) or not (0). Due to the inherent limitations of the available USCB data, the binary variables were associated with the entirety of the state delegation in the sample set within the observation period. While this was a logical step for senators because they represent the entirety of a state, this was extended to individual representatives because state delegations typically have a vested interest in representing the interests of their entire state in terms of international trade and international relations. This vested interest would likely be higher in delegations from states that have a stronger industry interest relative to a particular piece of legislation as compared to delegations from states that do not.

Concerning the energy sector, the delegations of the 14 states that were included in the sample set as demonstrating constituency stake in the energy sector were likely to have had a similar economic pressure from their constituents to support legislation that benefits their home state and constituents as related to the energy sector. Thus, the full delegation as represented in the sample set was included for analysis. Forty-nine observations were recorded for the 2012, 2014, and 2017 observation periods, with 16 occurring in the Senate and 33 occurring in the House of Representatives. It is worth noting that all representatives from the 14 states identified in the 2012, 2014, and 2017 observation periods were also included among the 200 person-vote observations from the same periods that received campaign contributions from the energy sector, thus mitigating concerns of data bias toward greater energy influence on voting behaviors for a full delegation.

References:

“Ukraine Freedom Support Act of 2014.” Congressional Record 160 (2014) pp. 10305–10310. (Text from: Congressional Record Permanent Digital Collection). https://www.govinfo.gov/content/pkg/CREC-2014-12-11/pdf/CREC-2014-12-11-bk3.pdf.

Engel, E. (NY-17). “Ukraine Freedom Support Act of 2014.” Congressional Record 160 (2014) pp. 10305. (Text from: Congressional Record Permanent Digital Collection). https://www.govinfo.gov/content/pkg/CREC-2014-12-11/pdf/CREC-2014-12-11-bk3.pdf.

Federal Elections Commission. (n.d.a). Contribution limits. https://www.fec.gov/help-candidates-and-committees/candidate-taking-receipts/contribution-limits/.

Federal Elections Commission. (n.d.b). Multicandidate status. https://www.fec.gov/help-candidates-and-committees/filing-pac-reports/multicandidate-status/.

Federal Elections Commission. (n.d.c). Recording receipts. https://www.fec.gov/help-candidates-and-committees/keeping-records/records-receipts/.

Federal Elections Commission. (n.d.d). Who can and can’t contribute. https://www.fec.gov/help-candidates-and-committees/candidate-taking-receipts/who-can-and-cant-contribute/.

Fulmer, S., Knill, A., & Yu, X. (2012). Political contributions and the severity of government enforcement (Working Paper). https://editorialexpress.com/cgi-bin/conference/download.cgi?db_name=CICF2014&paper_id=532.

Kaptur, M. (OH-9). “Ukraine Freedom Support Act of 2014.” Congressional Record 160 (2014) pp. 10305. (Text from: Congressional Record Permanent Digital Collection). https://www.govinfo.gov/content/pkg/CREC-2014-12-11/pdf/CREC-2014-12-11-bk3.pdf.

Mufson, S. (2014, April 7). Does Koch Industries hold most Canadian oil sands leases? It’s complicated. Washington Post. https://www.washingtonpost.com/business/economy/does-koch-industries-hold-most-canadian-oil-sands-leases-its-complicated/2014/04/07/2470e5e4-be70-11e3-b574-f8748871856a_story.html?itid=lk_inline_manual_2.

Mufson, S. & Eilperin, J. (2014, March 20). The biggest foreign lease holder in Canada’s oil sands isn’t Exxon Mobil or Chevron. It’s the Koch brothers. Washington Post. https://www.washingtonpost.com/news/wonk/wp/2014/03/20/the-biggest-land-owner-in-canadas-oil-sands-isnt-exxon-mobil-or-conoco-phillips-its-the-koch-brothers/.

Office of the Clerk. (n.d.). Instruction guide: Financial disclosure statements and periodic transaction reports. https://ethics.house.gov/sites/ethics.house.gov/files/documents/CY%202019%20Instruction%20Guide%20for%20Financial%20Disclosure%20Statements%20and%20PTRs.pdf.

Stop Trading on Congressional Knowledge (STOCK) Act, Pub. L. 112–105, § 8, 126 Stat. 291 (2012).

Ukraine Freedom Support Act of 2014, Pub. L. №113–272, § 2, 128 Stat. 2953 (2014).

U.S. Senate Select Committee on Ethics. (2020, February). Financial disclosure instructions for Calendar Year 2019. https://www.ethics.senate.gov/public/index.cfm/files/serve?File_id=EDADCA8C-D651-4244-BFDE-CC9C69C28687.

Wells Fargo & Company. (2019). Exhibit 13 (Financials) of the 2019 Annual Report. https://www08.wellsfargomedia.com/assets/pdf/about/investor-relations/sec-filings/2019/exhibit-13.pdf.