Professor’s Study Questions Narrative in 2006 U.S. Attorney Dismissals

The dismissals of nine U.S. attorneys in 2006 generated headlines, congressional hearings and accusations that the decisions were politically motivated.

Although a federal investigation called the decision-making process in the firings unsystematic and arbitrary, new UT Dallas research argues that it was not as haphazard as many thought.

The study, published in the Journal of Law and Courts, found evidence that four of the attorneys were dismissed because they were in the path of the least political resistance.

“In effect, the U.S. attorneys who were removed were the easiest to replace,” said Dr. Banks Miller, program head of political science and public policy and political economy in the School of Economic, Political and Policy Sciences. Miller wrote the study with Dr. Brett Curry of Georgia Southern University, who is also a co-author of their forthcoming book, United States Attorneys, Political Control, and Career Ambition.

The administration of former President George W. Bush ordered the dismissals during the middle of Bush’s second term. U.S. attorneys, who are appointed by the president and must be confirmed by the Senate, typically serve until the end of the president’s term, although they can be removed at any time. As many as one-third of the nation’s 93 U.S. attorneys were on various Justice Department lists for potential removal during this time.

While the administration cited poor performance for the dismissals, many raised concerns that the decisions were based on improper political factors, such as how the attorneys handled certain voter fraud or public corruption investigations.

The Justice Department’s investigation found “significant evidence that political partisan considerations were an important factor in the removal of several of the U.S. attorneys” and called the process for determining which ones to dismiss “arbitrary.”

But, according to Miller’s study, four of the dismissed attorneys had something in common: Their districts were represented by two Democratic senators. At the time, Democrats were the minority party in the Senate.

“Because Democratic senators represented the districts of the U.S. attorneys who were removed, the Bush administration could make a replacement without any constraints on whom they put in that position,” Miller said. “This part of the story hasn’t been told. Underneath the narrative that these firings were haphazard, there is something systematic about what they did.”

The authors analyzed the number of cases the attorneys filed and completed, convictions, prison sentences and total prison terms for those convicted between 2002 and 2006. U.S. attorneys in districts represented by only Democratic senators were six times more likely to be asked to resign than those in districts represented by at least one Republican senator, the study found.

A variety of reasons explain the other dismissals, according to the research. The study also found that attorneys who filed more cases had a higher chance of keeping their positions. For example, a U.S. attorney who filed an average of 276 cases per year had a 1 in 3 chance of being asked to resign, while a U.S. attorney with an average of 2,252 cases per year had about a 1 in 12 chance of being asked to leave.

While the dismissals happened 12 years ago, Miller said he decided to examine the issue after obtaining the necessary data for his book. He said he wanted to focus on the dismissals when he was working on his dissertation for his PhD in political science at Ohio State University, but the data was not available at the time.

The findings demonstrate that politics were not the only factor in determining which U.S. attorneys to fire, Miller said. He said the findings also illustrate a trend over the past few decades in which the federal government has increased the oversight of U.S. attorneys.

“Despite a popular narrative of the dismissals being solely about retribution, we find evidence that it’s about control, something both parties seek,” Miller said.