Abstract
We find that federal bureaucrats award more, larger, and less risky contracts to politically connected firms when they have greater discretion over contracting outcomes. Using a sample of 4.3 million federal government contract actions obligating $2.47 trillion between 2000 and 2015, we show that this result varies predictably across contract and agency characteristics, over time, and in placebo tests, and is robust to a comprehensive fixed effect structure and seven alternate measures of political connectedness. Our evidence illustrates the overlooked role of the bureaucrat in facilitating political bias in federal contracting outcomes.
| Original language | English (US) |
|---|---|
| Article number | 101173 |
| Journal | Accounting, Organizations and Society |
| Volume | 88 |
| DOIs | |
| State | Published - Jan 2021 |
| Externally published | Yes |
Keywords
- Corporate political activity
- Government spending
- Monitoring and control
- Procurement contracts
ASJC Scopus subject areas
- Accounting
- Sociology and Political Science
- Organizational Behavior and Human Resource Management
- Information Systems and Management
Fingerprint
Dive into the research topics of 'Bureaucratic discretion and contracting outcomes'. Together they form a unique fingerprint.Cite this
- APA
- Standard
- Harvard
- Vancouver
- Author
- BIBTEX
- RIS
In: Accounting, Organizations and Society, Vol. 88, 101173, 01.2021.
Research output: Contribution to journal › Article › peer-review
}
TY - JOUR
T1 - Bureaucratic discretion and contracting outcomes
AU - Boland, Matthew
AU - Godsell, David
N1 - Funding Information: Using the quarter-end increase in non-competed contracts as a source of variation in bureaucratic discretion over contracting outcomes, and an estimation sample of 4.3 million federal government contract actions obligating $2.47 trillion between 2000 and 2015, we test whether bureaucratic discretion over contracting outcomes is associated with political bias in contracting.4 Consistent with CDT, we find that bureaucrats award more, larger, and less risky non-competed contracts to PC firms at quarter-end when they have more discretion over contracting outcomes. We find no political bias in contracting prior to the quarter-end or for fully competed contracts during the quarter-end. We draw credible inferences from our setting because we use multivariate regressions employing a large suite of controls and a comprehensive fixed-effect structure capturing firm, time, industry, and agency characteristics. We draw generalizable inferences because we use contract data for the entire federal bureaucracy rather than an individual agency or branch of the U.S. government. Overall, our credible and generalizable results illustrate that, consistent with CDT, bureaucrats autonomously inject political bias into contracting outcomes.We next identify cross-contract, cross-agency, and temporal variation in our main results consistent with CDT. First, to corroborate bureaucratic discretion over contracting outcomes as the mechanism driving contracts to PC firms, we examine a subgroup of non-competed contract actions that are especially susceptible to bureaucratic discretion: Indefinite Delivery Vehicles (IDV) contracts. IDV contracts permit greater bureaucratic discretion over contracting outcomes because IDV contracts are open-ended contracts with new orders and spending occurring on an as-needed basis without the encumbrance of a new tendering process. We find that bureaucrats award more and larger non-competed IDV contracts to PC firms at quarter-end, relative to non-PC firms. Second, we expect that bureaucrats in agencies more influenced by politicians more often adopt politicians’ contracting preferences when they have discretion over contracting outcomes. This prediction follows from CDT because under CDT bureaucrats adopt politicians’ contracting preferences when politicians have greater influence over the agencies’ operating environment (e.g., bureaucrats’ resources and promotion prospects). We measure politician influence over an agency using a time-varying measure of agency discretion provided by Bolton and Thrower (2019) that captures the extent to which politicians layer agency budget appropriations with procedural requirements and spending instructions. Consistent with CDT, we find that bureaucrats award more and larger non-competed contracts to PC firms at quarter-end when they operate in an agency more susceptible to political influence. Third, we examine whether our results vary as the quarter-end window shrinks. Budget underspending penalties gain primacy over procurement rule compliance, and increase the likelihood that bureaucrats switch to quick-to-complete non-competed contracts, as the quarter-end nears. We re-estimate our main results using four quarter-end windows decreasing in size (one month, two weeks, one week, three days). We find that as the quarter-end window shrinks 1) bureaucrats award more non-competed contracts, and 2) bureaucrats increasingly award more, larger, and less risky non-competed contracts to PC firms.Our findings offer several insights to policymakers and the prior literature. First, we extend the CDT literature by inferring from bureaucrats’ revealed contracting actions that bureaucrats adopt politicians’ contracting preferences and facilitate political bias in contracting outcomes when they have the discretion to do so. While other papers describe empirical patterns consistent with CDT, these papers examine emergency or ad hoc settings such that inferences gleaned therefrom do not generalize to year-round federal operations. For example, Ryan (2014) finds that H1N1 vaccines were initially distributed to states with Congressional representatives that oversaw the Department of Health and Human Services, and Garrett and Sobel (2003) find that FEMA disaster relief funds are more quickly allocated to politically important states.5 Confounding inferences in the prior literature is the study of irregular events more likely to attract politician intervention. Such interventions by politicians curb rather than accentuate the bureaucratic discretion that is essential to tests of whether bureaucrats autonomously inject political bias into contracting outcomes. Relative to this prior literature, we draw more credible inferences regarding the validity of CDT because we exploit a systematically recurring source of variation in bureaucratic contracting discretion that 1) is not systematically susceptible to direct political intervention, 2) permits the identification of a credible counterfactual,6 3) not only varies over time but also across contract types and agency, and 4) allows inferences generalizable to the entire federal bureaucracy rather than an individual agency or branch of the U.S. government.Category A apportionments are legally binding under the Anti-Deficiency Act (ADA). Agencies report on budget spending quarterly to facilitate compliance with ADA 31 U.S.C. 1511–1514, which requires that the President review federal expenditures four times each year. The GAO enforces the ADA, viewing overspending a Category A quarterly apportionment as just as illegal as overspending the annual appropriation (GAO, 2006, chapter 12, p. 21). The ADA specifies that “[i]ncurring of obligations or the making of expenditures (outlays) in excess of amounts available in appropriations or funds” can result in criminal penalties, including fines and imprisonment (Title 31, Sections 1349–1350 of the United States Code).Politicians can generate financial resources that facilitate re-election through the disbursement of government contracts to companies that support their electoral efforts. Boddewyn and Brewer (1994) argue that political actors are for sale while others argue that there exists a market for political favors (Pfeffer & Salancik, 1978) and that these political transactions are enforced by implicit contracts (Williamson, 1985). Stratmann (2002) provides evidence suggesting that politicians reverse decisions based on the receipt of political contributions. Fisman (2001); Faccio, Masulis and McConnell (2006); and Claessens, Feijen and Laeven (2008) find strong links between political connections and economic benefits of firm sponsors while others find that political contributions and connections appear to be reciprocated via government procurement contracts (Tahoun, 2014), a greater likelihood of government bailout (Duchin & Sosyura, 2012), restrained SEC oversight (Correia, 2014) and more favorable lending from government-owned banks (Khwaja & Mian, 2005).A positive coefficient on B3 will support CDT because it will show that bureaucrats award more contracts to PC firms when their discretion over contracting outcomes increases. In falsification tests, we expect no variation in the likelihood that a bureaucrat awards competed contracts to a PC firm at quarter-end because bureaucrats do not have discretion over competed contract awardees.The test variable coefficient in Table 8 is IDV × Non-Competed Contract × Quarter-End. The test variable in Table 8 captures the incremental probability that a bureaucrat awards a non-competed IDV contract to a PC firm at quarter-end. The test variable coefficient in Table 8 is positive and statistically significant, consistent with bureaucrats diverting contracts to PC firms when they have greater discretion over contracting outcomes. We corroborate this inference in tests examining contract size and risk in Table 9. The test variable is PC × IDV × Non-Competed Contract × Quarter-End in Table 9. We regress Contract Size and Cost-Plus on firm and contract characteristics in Columns 1 and 2, respectively. Table 9 presents the results of the modified Equation (2). The test variable coefficient in Column 1 of Table 9 is positive and statistically significant, consistent with bureaucrats awarding larger non-competed IDV contracts to PC firms than to non-PC firms at quarter-end. The test variable coefficient in Column 2, where the dependent variable is Cost-Plus, is indistinguishable from zero. Overall, however, evidence presented in Tables 8 and 9 is consistent with bureaucrats awarding more and larger, if not less risky, non-competed IDV contracts to PC firms at quarter-end. This evidence supports CDT because it demonstrates that political bias in contracting outcomes increases with bureaucrats’ discretion over contracting outcomes. Publisher Copyright: © 2020 Elsevier Ltd
PY - 2021/1
Y1 - 2021/1
N2 - We find that federal bureaucrats award more, larger, and less risky contracts to politically connected firms when they have greater discretion over contracting outcomes. Using a sample of 4.3 million federal government contract actions obligating $2.47 trillion between 2000 and 2015, we show that this result varies predictably across contract and agency characteristics, over time, and in placebo tests, and is robust to a comprehensive fixed effect structure and seven alternate measures of political connectedness. Our evidence illustrates the overlooked role of the bureaucrat in facilitating political bias in federal contracting outcomes.
AB - We find that federal bureaucrats award more, larger, and less risky contracts to politically connected firms when they have greater discretion over contracting outcomes. Using a sample of 4.3 million federal government contract actions obligating $2.47 trillion between 2000 and 2015, we show that this result varies predictably across contract and agency characteristics, over time, and in placebo tests, and is robust to a comprehensive fixed effect structure and seven alternate measures of political connectedness. Our evidence illustrates the overlooked role of the bureaucrat in facilitating political bias in federal contracting outcomes.
KW - Corporate political activity
KW - Government spending
KW - Monitoring and control
KW - Procurement contracts
UR - http://www.scopus.com/inward/record.url?scp=85091821767&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85091821767&partnerID=8YFLogxK
U2 - 10.1016/j.aos.2020.101173
DO - 10.1016/j.aos.2020.101173
M3 - Article
AN - SCOPUS:85091821767
SN - 0361-3682
VL - 88
JO - Accounting, Organizations and Society
JF - Accounting, Organizations and Society
M1 - 101173
ER -