Politicizing Agency Spending Authority: Lessons from a Bush-era Scandal

Abstract

W hen can presidents direct bureaucrats to allocate government expenditures for electoral purposes? To address this question, I exploit a scandal concerning the General Services Administration (GSA), an agency that contracts with private vendors to provide supplies and real estate to other agencies. Shortly after Republican losses in 2006, a White House deputy gave a presentation to GSA political appointees identifying potentially vulnerable congressional districts. I find that vendors in prioritized Republican districts experienced unusually large new contract actions from the GSA's Public Buildings Service following the presentation relative to unmentioned districts, a discrepancy that disappeared once the Washington Post broke the story. Contracts supervised by the agency's Federal Acquisition Service, by contrast, were largely unresponsive to the briefing and media scrutiny. My findings suggest that the extent to which executives succeed in politicizing discretionary allocation decisions depends upon key features of the implementing agency's tasks and its informational environment.

Electoral cycles in pork barrel politics: Evidence from South Korea 1989–2008

“…In other words, ignoring cyclical changes in allocation patterns may result in Type II errors for swing voter models, which could fail to detect the fact that marginal voters are targeted but only in a short period of time just prior to an election. This conjecture is consistent with the tendency found in empirical studies on distributive politics: studies analyzing the overall allocation of resources over multiple elections often present findings that core supporters receive more government resources (Ansolabehere and Snyder, 2006;Berry et al, 2010;Larcinese et al, 2008), while studies focusing on the preelectoral period tend to support the swing voter model (Dahlberg and Johansson, 2002;Denemark, 2000;Gordon, 2011;Herron and Theodos, 2004).…”

“…Studies that find support for the core voter model include Ansolabehere and Snyder (2006), Balla et al (2002), Bickers and Stein (2000), Larcinese et al (2008) and Levitt and Snyder (1995). Conversely, Berry et al (2010), Case (2001), Dahlberg and Johansson (2002), Denemark (2000), Herron and Theodos (2004), Gordon (2011), Kwon (2005), and Stokes (2005) all marshal empirical evidence to support the swing voter model of resource distribution.…”

The Statistics of Causal Inference: A View from Political Methodology

“…Take DID as a second example. Gordon (2011) is one example where a DID identification strategy is highly plausible. Alternatively, Keele and Minozzi (2012) outline an example where a DID identification strategy generally fails.…”

The Contemporary Presidency: Presidents Profiting from Disasters: Evidence of Presidential Distributive Politics

“…The above research indicates that executive agencies appear to use distributive politics for partisan advantage but does not put forth evidence that these decisions are being made in the White House. Gersen and Berry () as well as Gordon () indicate that political appointees likely play a role, but these appointees may be acting on explicit or implicit direction of the president or simply demonstrating their own partisan biases. This article is focused on presidential influence in this process.…”