Investigative decision-making in public corruption cases: Factors influencing case outcomes

Abstract Public corruption undermines the rule of law that is crucial to the maintenance of governmental legitimacy and peaceful co-existence. As political divisiveness has grown, politicians have called public corruption investigations “witch hunts” and part of partisan politics, rather than a search for justice. Contemporary attacks on law enforcement charge that they are acting with political motivation which further undermines the legitimacy of these investigations. It is important for research to delve into how the investigative processes maintain their fact-finding missions in support of the rule of law. This study addresses how public corruption investigations occur and how the investigations themselves are conducted. First, federal prosecution data are used to understand the frequency of investigations over the last thirty years. Second, interviews with 40 former investigators and prosecutors examines the process, protocols, tools and resources needed for successful corruption investigations. Investigations require interagency cooperation (especially at different levels of government) and particular attention to the use of investigative tools that ensure the legitimacy of the law in the eyes of the accused, the courts and the general public. In addition, a decline in public corruption cases is seen in recent years with the US prosecution effort diverted to immigration cases.


PUBLIC INTEREST STATEMENT
No one is above the law. Criminal investigations are one way to hold public officials accountable. This article explores the history of public corruption investigations using federal prosecution data to see whether investigations have declined in a way corresponding to observed declines in convictions. We found that the decline in corruption cases corresponded to changes in FBI directors, suggesting a change in investigative priorities. At the same time, immigration cases increased from about 30,000 to 90,000 cases, a dramatic increase. We interviewed 40 former investigators and prosecutors to understand the process and protocols for such investigations. They told us that investigations require approvals at every stage, including permission to open a case. Successful investigations depended on multiple agencies working together. Further, public corruption investigations are complicated and require not only time and resources but also support from leadership for such these sensitive investigations.

Introduction
In a democracy, the rule of law is crucial to the maintenance of governmental legitimacy and peaceful co-existence. Leaders cannot be above the law if the letter and spirit of law is to be respected. Research has shown a decrease in public corruption convictions in the US since the high at 410 convictions since 2001 to the lowest point in 2015 at 330 (Albanese, Artello, & Nguyen, 2019). The public perception about U.S. government's ability to fight corruption declined steadily culminating in 2017, when twice as many people found that government is very bad at fighting corruption (Transparency International [TI], 2017). In 2018, the fewest public corruption cases were prosecuted ever in U.S. (Rampell, 2018).
As political divisiveness has grown, politicians have called public corruption investigations "witch hunts" and as part of partisan politics (Quinn, 2017). US Presidents Nixon, Clinton and G.W. Bush had their political proxies refer to allegations and investigations as "witch hunts" (Quinn, 2017). Unlike his predecessors, President Trump directly attacked the motives of U.S. law enforcement agencies concerning the investigation into Russian meddling into the 2016 election (Berenson, 2018). Recent attacks by government leaders on law enforcement such as FBI agents and directors, claiming the investigators are acting with political motivation, undermine the legitimacy of law and the legal process (Baker & Benner, 2018). As investigations continue, however, it is important to understand how the investigative process, especially in corruption investigations occurring outside of public view, are conducted to understand their legitimacy as guardians of the independent rule of law (e.g. those investigations conducted by law enforcement agencies; not Congressional oversight hearings). The process and methods used in corruption investigations have also come under fire (Bump, 2017). Individuals have questioned whether an alleged informant placed in the campaign entrapped two of Trump's advisors (Bump, 2017;Rucker, Costa, Leonning, & Dawsey, 2018). Surveillance and search practices, and even text messages by law enforcement officials, have also come under scrutiny (Klaas, 2018;Smith & Sheth, 2018).
In a democracy, the rule of law is central to the governing system. As part of the system, the fair and impartial investigation into wrongdoing of those with whom the public has entrusted public monies and authority is essential for a democracy to work well. If public officials are not held accountable to the law, the basic tenets of democracy are eroded. When public officials politicize the investigation process, they attack the legitimacy of the rule of law itself. By understanding the investigative process, the attacks on its legitimacy may be reduced, and consequently, protect the rule of law. This study fills a gap in the literature by examining potential explanations for the decrease in public corruption cases and investigative process in recent years regarding the process, tools used and resources needed for successful investigations.

Literature review
Public corruption is the misuse of one's public office and/or authority, either an elected office or appointed position, for personal gain (Rose-Ackerman & Palikfa, 2016;Transparency International 2017). Yet, the US Department of Justice [DOJ] has broadened this definition in its designation of public corruption beyond merely individuals acting in their public capacity as described above but also to include the misuse of public property and/or funds; and/or defrauding a government program even without any involvement of an elected or appointed official in said schemes or actions (U.S. Department of Justice, 2018). Consequently, DOJ data reflects more than the actions of those who are entrusted by the public to act on its behalf beyond the accepted definition of public corruption making a true measurement of the extent of prosecution corruption difficult to ascertain.
Research on public corruption has focused primarily on convictions as a means to measure objectively the extent of the problem. Public corruption cases are not tracked at the state or local level in the US. In an effort to learn the extent of the issue, Cordis and Milyo (2016) searched all newspaper and newswire coverage from 1986 to 2014, to identify public corruption cases by using terms related to political corruption, scandal, bribery, conspiracy, embezzlement, fraud, kickbacks and misappropriation and then cross-referencing to ensure that public employees were involved in some way. Only 910 public employees' convictions for public corruption were discovered outside federal courts. The remainder of the cases (n = 15,542) were found to be pursued by and prosecuted in federal system, which means that 94% of public corruption cases are addressed exclusively by the US Department of Justice (Cordis & Milyo, 2016). The Department of Justice records a referral as "public corruption" if the assigned Assistant U.S. Attorney (herein AUSA) designates the lead charge as a public corruption case (when the precise statutory charge might be bribery, embezzlement, extortion, conspiracy, or other charge). Additionally, DOJ also classifies all cases where individuals defraud a federal program (involving taxpayer funds), such as a Medicaid, although public officials were not involved (U.S. Department of Justice, 2018). Therefore, the database used in this study defines all cases designated as public corruption by DOJ to include individuals acting in their public capacity, the misuse of public property and/or funds; and/or defrauding a government program. Some research has examined prosecutor's decisions related to public corruption cases, such as the impact of resources, political considerations and organizational culture. One study found that by adding one more attorney to an office led to many more corruption prosecutions occurring (Alt & Lassen, 2012). Another study compared public corruption convictions between Clinton and Bush II administrations. The study concluded that both administrations prosecuted individuals in both parties, although the Bush Administration appeared to focus on Democratic defendants more than Republican, and defendants prosecuted in the Clinton era faced 6 months longer prison sentences (Gordon, 2009). Pavlik (2017) found slightly different results when examining key states as defined as states determining the outcomes of presidential elections. He found that key states had more public corruption convictions than those states not determinative in presidential elections (Pavlik, 2017). The difference between the two studies may be the result of which party is in political power, rather than targeting partisan enemies. As Republicans have had more control of state and local governments in recent years, more convictions of Republicans will be seen, if this explanation is accurate.
There has been a paucity of research on investigation process and tools used in public corruption cases. One study examined the rate of convictions for public corruption over the last thirty years . They found that public corruption convictions have dipped in the last decade from a high of 400 in 2008 to less than 330 in 2015 . The question becomes why the numbers of convictions continued to decline. One possible explanation explored was whether prosecutors were less willing to take these types of cases to court (Artello & Albanese, 2019). In this study, they compared public corruption cases to white collar crimes and organized crime because these types of cases share complexities in litigation (e.g. questions about causation and intent, statutory application and factual intricacies), more resources for legal representation, and longer investigations. Organized crime and public corruption also share undercover operations as well, unlike other types of white-collar crime. Based on data on federal prosecutions, gathered by Transactional Records Access Clearinghouse (herein TRAC), it was found that two-thirds of public corruption investigations were declined by prosecutors from 2004-2015, in contrast to organized crime cases where prosecutors prosecuted two-thirds of the investigative referrals (Artello & Albanese, 2019).
The current study investigates potential explanations for the reduction in convictions over time: what may be influencing the decrease in public corruption cases being prosecuted by the US Department of Justice? To address this question, this article explores the number of investigations over time, potential influences on the investigation process, the actual investigation process, and the tools and resources necessary to complete successful investigations into public corruption.

Methods
This study used a mixed data approach by examining federal data obtained via Transactional Records Access Clearinghouse, a data gathering, data research and data distribution organization at Syracuse University. TRAC gathers, organizes and makes available data on the operations of the government for public access and review. TRAC systematically collects and verifies government agency data using the Freedom of Information Act (FOIA). Using this source, records regarding referrals from investigative agencies to DOJ were reviewed covering more than 30 years 1986-2017 (n = 45,983).
A sample of indictments and information filings (n = 46) taken from this universe of cases were analyzed using court records on public corruption cases covering a period of 3 years. Some cases include multiple defendants that arise out corrupt schemes and behaviors (Albanese & Artello, 2019). Additionally, interviews (n = 40) were conducted to elucidate and expand on the investigation process and patterns discovered in the federal data available from TRAC and analysis of court documents. Corruption cases were identified from an analysis of US Attorneys' press releases when either indictments or convictions were announced.
A recent analysis found that the federal data organized by TRAC was more valid and reliable than the government agency reports themselves (Cordis & Milyo, 2016). Once TRAC obtains data through FOIA, a variety of statistical techniques are used to check and verify the data. The TRAC data are stripped of all identifiers, so they cannot be linked to specific cases or individuals. TRAC data on federal referrals designates a program category (e.g. public corruption, organized crime, white collar, terrorism, immigration) based on the classification assigned by the Assistant U.S. Attorney. We examined the program category of public corruption from 1986-2017 (n = 45,983) to understand the trends in these referrals. TRAC organizes these as cases and not by defendants. A referral may result in a prosecution or in declination. A declination is when a prosecutor decides not to move forward on a referral to create a prosecution case (Transactional Records Access Clearinghouse [TRAC], 1999).
In an effort to study the process of public corruption investigations, an examination was undertaken of the prosecution and conviction announcements in every case of public corruption. We gathered press releases (n = 2,419) from all of the US Attorneys' offices (n = 94) using a keyword searches for corruption, public corruption and corrupt acts over a period of 3 years (2013)(2014)(2015). We focused on a three-year period because most U.S. Attorneys' offices archived these announcements up to 3 years. Because public corruption cases engender a great deal of public interest, they are invariably announced through press releases, and sometimes the indictments themselves are also released to highlight and disseminate the prosecution effort against public corruption. If the office did not announce the action in a press release for some reason, there was no other way to locate potential prosecution information and link it to a specific case. Once the press releases were collected, they were reviewed using a pragmatic inductive approach to be grouped by type of corrupt conduct, defined by involving public officials using their position or power for personal gain (Albanese & Artello, 2019;Savin-Baden & Major, 2013). Examples include bribery, embezzlement, extortion, fraud, official misconduct and obstruction of justice.
A sample of 313 cases was selected for secondary coding, because they represented distinct kinds of corrupt behavior identified by the authors. The distinct kinds of behaviors include the following: receiving and soliciting bribes, extorting favor or money, defrauding a contract, embezzling from government, official misconduct, obstructing justice and violation of regulatory laws (Albanese & Artello, 2019). The secondary coding of corruption cases was based on fact patterns, behaviors, agencies, parties involved and extent of corruption. To summarize, an initial descriptive coding (n = 313 cases) was completed, and a second, more detailed, analytical coding occurred to designate cases representative (n = 33) of the different types of factual patterns discovered to make a deeper analysis of the cases to include pleadings. Each of these samples were derived from the original sample of cases (n = 2,419) listed in press releases (Albanese & Artello, 2019). In the federal system, cases can be started with an information pleading and then followed up with a grand jury indictment after a grand jury is empaneled. Some cases will have information pleadings as well as an indictment. Other cases also have more than one indictment because new information was uncovered and indictments were amended. Consequently, using the federal case management system (PACER), the cases' informations and indictments (n = 46) were identified and gathered.
Using purposive and snowball sampling, interviewees were discovered through a general recruitment email to former US Attorney and FBI associations with information and contact information about the study. Some interviewees self-identified from the initial emails and then these interviewees suggested future interviewees for former US Attorneys and FBI agents having significant experience in corruption cases. The refereed potential interviewees were contacted via email, and if interested, they would call or email the researchers. Interviews (n = 40) were semi-structured with open-ended questions and lasted from 40 min to 2 and a half hours. The former U.S. attorneys and assistant U.S. Attorneys (n = 22) in the sample had served in offices across the country and had experience prosecuting and/or public corruption cases. Former FBI agents (n = 18) served in different offices across the US and had experience investigating multiple public corruption cases at all levels of government: local, state and federal.
An inductive coding, where the coding comes from the actual data rather than imposed by a hypothesis, was used to create an initial descriptive coding focusing on the methods used in investigations. Investigation types were identified, e.g. undercover, documents, informants. The second analytical coding examined the process which investigations followed, including how evidence is developed and how discretion and decision points are employed in making investigative decisions including stages of investigations and resources used (Savin-Baden & Major, 2013).

Results
The results fall into three different substantive categories: number of investigation referrals for prosecution, investigation process and/or tools, and resources/impediments to completing such investigations.

Investigation referrals
A conviction cannot occur without an investigation. Public corruption investigations can be as damaging to an individual's reputation as a conviction, as reported by Hillary Clinton about the 2016 presidential election (Clinton, 2017). Reputational damage to public officials is a major concern, so corruption allegations are investigated carefully. Consequently, investigations that do not proceed to prosecution are not transparent to the public (Interviews A-02, I-02, A-05, A-09, A-20 and A-22). The federal data provided by TRAC provides information regarding completed investigations because they are then referred to Department of Justice. If an investigation is incomplete or discontinued for any reason, there is no mechanism to capture the reasoning behind these decisions, whether it was discontinued for lack of resources, change in leadership, or insufficient evidence. Therefore, the most complete information available on the number of public corruption investigations comes from the federal data on formal charges, indictments and convictions gathered by TRAC.
As noted earlier, TRAC compiles data without any identifiers from Department of Justice through FOIA requests to examine how prosecutors are actually conducting the business for the American (i.e. taxpaying) people (TRAC, 2016). As seen in Figure 1 Next, the rate of referrals by US President from Reagan (starting in 1986) through the first year of Trump's presidency was undertaken to ascertain whether political party of the President made a significant difference in public corruption referrals in order to explain the recent decline. We divided their annual case referral totals by President's years in office to create an annual rate for each President as seen in Table 1. Findings reveal the pattern does not appear to be based on party, as Bush I and Clinton are less than 150 referrals apart. Additionally, Trump and Obama are less than 200 investigations apart per year through 2017. Bush I had the most annual investigations at a rate of 2,179 followed closely by Clinton at 2,070. Although Bush II shifted resources towards terrorism after the attacks of 9/11, the rate of referrals during his presidency was nearly twice that seen during Obama's eight years. However, the lowest case referral rate since 1986 occurred in 2017 during Trump's presidency at only 454 referrals, a drop of nearly 80% from the Bush I era. Source: TRAC Referrals for public corruption program prosecutions to US DOJ Annual rate of investigation referrals by President Regan 1986-1988Bush I 1989-1992Clinton 1993-2000Bush II 2001-2008Obama 2009-2016Trump 2017 Source: TRAC data from 1986-2017.
Since presidential administration does not appear to explain the pattern seen in referrals in Table  1, we then analyzed the referral rate by the identity of the Director of the Federal Bureau of Investigation. While other agencies may also be involved in public corruption cases (e.g. interviewee A-12 relied on IRS agents for information on corruption), the Federal Bureau of Investigation [FBI] was the lead investigative agency in the vast majority of public corruption cases. Additionally, we considered that some FBI directors spanned more than one presidency, as they have maximum 10year terms. When the case referral rate is separated by each FBI director's term from 1986-2017, as seen in Figure 2, the pattern appears similar to the one seen in Table 1. Some directors overlapped for a year, for example, FBI director Webster resigned on 25 May 1987 and his successor was named on 2 November 1987. In these cases, we counted 1987 to Webster instead of Sessions (Federal Bureau of Investigation [FBI], 2017). Since the investigations were probably started under the predecessor, we counted that year for the prior Director during a transition time, because the prior director had to authorize resources for those investigations. The average for Webster came from only two years 1986 and 1987, as those were the only years available with corruption case referral numbers. Figure 2 follows a similar pattern found (in Figure 1) in the referral rate reported by DOJ. The public corruption referral numbers correspond with the terms of FBI Directors more closely than with presidential administrations. Figure 3 overlays the public corruption referrals by year with the average annual rate per FBI Director. It clearly shows the close relationship between corruption case referrals and FBI Directors. Many of the directors span multiple presidents including administrations of differing parties (e.g. Director Freeh served under both Clinton and Bush II, Director Mueller served under Bush II and Obama, and Director Comey served under both Obama and Trump). This analysis suggests that the investigative priorities of each FBI directors play a significant role in the types of cases investigated.

Investigation process and tools
Public corruption investigations are unlike any other type of investigation completed by law enforcement agencies (Interviews A-02, A-09, A-21, A-22). Unlike bank robberies where a report from the bank about the robbery will trigger an investigation, public corruption cases require agents to predicate a case before proceeding to a preliminary inquiry that may eventually lead to a full investigation. Throughout this process, agents must obtain authorizations to proceed and then continually renew the investigation, which differs from any other type of investigation. One investigator summed it well as "To do anything [on a public corruption case] you need Attorney General approval and subpoena, you need approval very early in the process. It's not like bank robbery" (Interview I-17).

Predication of the case
Before a full investigation or even a preliminary one can begin, an agent must predicate the case. As one investigator explained, "First, you would have a control file. You have to correlate the rumor. You have to meet the predication standard which is higher than reasonable suspicion" (Interview I-02). To predicate a case, agents first identify the subject of the investigation is a public official. They must demonstrate that the public official either took an oath of office, serves in a public appointment, and/or is being paid from public money. Furthermore, "the agent has to complete extensive paperwork in the field. You're going to have to do interviews very carefully" (Interview I-02). This is because suspects in public corruption cases are public officials and therefore often educated, comfortable answering questions and careful to hide unauthorized activities.
Once those facts are established, agents need to develop a theory of investigation based on a fact pattern that is related to a potential statutory violation. Unlike other types of crimes, federal agents contact prosecutors at this stage to help proceed with corruption cases; having a prosecutor support the investigation is key to having an investigation proceed into the courtroom for prosecution. Once agents have an identified prosecutor with whom to work, the agent opens a preliminary inquiry. "In some states, young prosecutors and the FBI work together to investigate, and then you have the prosecutor will bring the case when it's ready. FBI has been charged to include the prosecutor at the early stages" (Interview S-14).

Preliminary inquiry
A preliminary inquiry uses non-invasive investigative tools. Some of the tools include a review of local contracts, publicly available financial information including property records, loan applications, mortgage documents and other financial disclosures. This type of inquiry is designed to explore potential areas where money may be used and/or hidden that exceed publicly disclosed salaries, while avoiding possible damage to the official's reputation. I-04 summed it well in an interview: We had certain types of investigations to use at the beginning. They would be records checks, open-source checks, and talk to a few people … You want to make sure that it's not a political rival or formal lover or someone with different motives. Once the case is mentioned as part of an investigation it's disparaging on that individual. It is important to get it right, because once you disparage a person it is bad. Consequently, an investigator may debrief a confidential informant to obtain as much detail as possible about a suspected corruption scheme, but very few witnesses are interviewed at the initial stages, especially if the inquiry is likely to damage the reputation of public officials. The purpose of this process is to provide collaboration to the original information. Once an agent has collaboration, a full investigation may be authorized based on the probable cause gathered through the preliminary inquiry.

Full investigations
Once an investigation has been approved, agents and prosecutors reach a decision point about the process to be used. One path involves the use of the grand jury investigation process at the federal level. A-03 describes the advantages to using the grand jury: Sometimes we would issue grand jury subpoenas and do all the investigations that way. … The power of the grand jury is to gather documents, and through the documents we can get people's names to be interviewed. If they refused to be interviewed, we can then subpoena them under the grand jury.
Grand juries may issue subpoenas for documents and call witnesses to testify under oath. This pathway usually results in potential public exposure because individuals may talk about being subpoenaed or preparing documents. Additionally, the person who is the focus of the investigation will likely learn of the subpoenas, and may attempt to evade or interfere with the investigation. In addition to subpoenas, prosecutors may ask for Title III wiretap authorization on an official, which is the federal law that permits electronic surveillance with a warrant stating probable cause of a serious law violation. Wiretaps are helpful, not only to gather information about the scheme and potentially other conspirators, but many investigators and prosecutors prefer them because they capture the actual words of targeted individuals. On the other hand, electronic surveillance requires a warrant, so other independent evidence must first be gathered to link the individual to a suspected felony.
A second path involves a continuation of sensitive investigation techniques such as undercover operations although, like electronic surveillance, undercover operations are expensive and can pose danger to undercover officers (Albanese, 2015). Undercover operations can be an effective tool for investigations because through such investigations, agents are more likely to capture the planning of illegal acts on video and/or audiotape, a type of evidence that is convincing to juries. It becomes more difficult for the public official to deny the activity when presented with a video or audio recording of the actions. I-04 explained, "Sensitive investigations may require an undercover, witnesses, intercepted communications." Undercover operations also provide another advantage because sophisticated and smart public officials avoid prosecution by placing an intermediary between themselves and the corruptible act (Interview A-03). For example, "we had to have an undercover operation because it was a very close tight knit society for these kickbacks" (Interview I-17).
Approval for undercover operations can be a long process as described by I-17: "It took us working a year to three years to get the wiretap after I went to the class in Los Angeles and through the DC approval predication process to lay it out the need for an undercover operation." Undercover operations usually involve a group task force consisting of multiple law enforcement agencies and require Department of Justice and FBI Headquarters approval to begin the operation. These investigations are reviewed every 90 days, especially when Title III wiretaps and undercover agents are being used. These approvals include a review of the investigative plan, the scenario and/or schemes used, and potential targets. Approvals from the top officials in the Department of Justice are necessary for any cases involving public officials in order to avoid cases that result in unfounded allegations that damage reputations. If the case is targeting nonpublic officials, the Special Agent in Charge of the local FBI offices may approve and supervise an undercover operation.
Sometimes a public corruption case may bypass the preliminary inquiry or have it waived by Department of Justice and FBI headquarters, if the original information provides sufficient evidence of corruption because "it is very important to develop an intelligent space to do public corruption cases" (Interview I-02). For example, In the investigation, there were [casino] licenses on the table-14 of the 15 had been issued. [Governor had] steered all the licenses for the riverboats. The last license for the riverboat became an undercover operation to establish public corruption. We applied for Title 3 wiretaps. We went after the Commissioner and Governor's son. We had 13 operations. Group 1 [operations had to be] approved by DC headquarters and supervisors. We ran a criminal undercover operation. Group 2 were at the field offices. (Interview I-02).
The case above reflects government licensing situations where decisions are made and illegal payments exchanged. Through these types of investigations, undercover agents work to get close to public officials in order to learn the actual operation of the corrupt acts, usually recording illicit payments with video or audio, to reduce alternative explanations for such transactions that might be offered by the targets (Interviews I-

Resources and challenges
Public corruption investigations depend on resources. Without proper resources, investigators cannot gather the evidence necessary to proceed. For example, law enforcement needs to be trained on how to detect such cases, to develop community resources and to formulate the facts to obtain prosecutorial buy-in. In many cases, prosecutors and investigators stumble on to the case as described by A-05.
It is important to teach prosecutors and investigators how to do these types of investigations and prosecutions. Many times, the prosecutor finds these cases by happenstance; [they are] not determined or focused well on these types of prosecutions. The focus of the FBI on these types of cases goes in and out. It is rare that the prosecutors are actually involved in the initial investigation. It is usually dependent on what bubbles up. We are a reactive force. Many of these cases are dependent on how others develop them.
Investigations require a task force of multiple agents and US Attorneys to be successful (Interviews I-06, A-20). One former US Attorney described one investigation requiring "6 to 8 DOJ attorneys and 68 FBI agents because of all of the logistics" (Interview A-04). In Operation Greylord (involving judicial corruption in Chicago), the investigation required FBI agents to be arrested as defendants and then make a deal to reduce or dismiss the case. This investigation proceeded for three and half years resulting in over 100 indictments (Hake & Klatt, 2015). Many of the former FBI agents interviewed for this study described investigations that took more than 2 years to gather sufficient information for a referral to prosecutors (I-01, I-02, I-04, I-07, I-10). I-11 explained it well "Public corruption cases last three, four, or maybe five years. It is harder to get agents on these cases" because working on "one case for several years cannot look good for your evaluation" (Interview I-17). These interviewees suggested that the multi-year nature of corruption cases makes them undesirable because of their time commitment and resource-intensive nature, resulting in convictions only after months or years of investigative work.
Challenges also push against public corruption cases from different directions. A noted earlier, federal authorities, rather than state or local authorities, pursue the vast majority of public corruption cases. Most state and local agencies lack the resources and training necessary to pursue them (Interview I-04). Many state and local investigators do not want to work on public corruption because it adversely affects them in the community. One investigator explains: Many detectives do not want to work public corruption. Look at internal affairs. It is very tough to staff them. I had a case in Hartford where we arrested a prosecutor for sex. We had two guys from the Hartford Police Department on the task force but they were stigmatized. (Interview I-06).
Former US Attorneys and FBI investigators concurred that local and state investigators do not pursue public corruption because they (state and locals) must continue to work within the community and many times work with individuals who would be under investigation (Interviews I-02, I-12, O-02, A-20, A-06, A-07, & A-15). Federal investigators and attorneys have distance from the local political milieu which allows them not only to see the community with greater clarity, but also, they are not beholding to state or local officials for their budgets or approvals for other investigations (Interviews A-02, A-15, A-20, & A-22). I-09 explains that the FBI is best to investigate public corruption because "we are not entrenched in the local fabric of the community." "Local district attorneys and police chiefs are elected too much and are dependent on the political party machine too much. They are also discouraged about going after their own party and discredited to go after the other party," A-20 explains the lack of local and state prosecutions.
Federal investigators expressed some reluctance about these cases as well because not only do they require significant resources but also they are time consuming. Undercover investigations are preferred although they require many resources and long-time commitment as described by a former US Attorney: "The question becomes how to break into this close[d] system. You need to have intelligence. When you don't have intelligence, it's hard to break into it. They would have to have time to be able to do the investigation" (Interview A-20).

Resource allocation
Public corruption investigations in the US have dropped dramatically since 2001 as seen in referrals for prosecutions. Interviews clearly identified that public corruption cases were a resource intensive endeavors which led the authors to examine the entire federal prosecution effort during this period to identify where the effort against pubic corruption cases was being diverted.
One possible explanation is that law enforcement has focused on other crimes, such as terrorism after the 9/11 attacks. Figure 4 presents data on federal referrals for prosecutions for both public corruption and terrorism cases. While the data shows a spike (n = 3,628) in 2003 for terrorism (soon after the attacks of 11 September 2001), the terrorism investigations dropped by two-thirds the next year to 1,221. By 2006, terrorism and public corruption referrals show comparable numbers of referrals. Consequently, the shift in focus to terrorism following 9/11/2001 does not explain the drop-off in terrorism cases seen following 2006. This finding led to further examination of where the investigative resources went that were formerly assigned to public corruption cases. If not terrorism cases, then where?
We examined organized crime cases but found no significant increases in prosecution referrals over time. We followed that by examining immigration cases. From 1986From -1996 cases held steady at about 8% of the total referrals for prosecution to DOJ. However, in 1997-2004, these cases more than doubled to over 23% and then doubled again in 5 years to 41% in 2008. It ranged from 41% to 46% through Obama's administration (2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016). Cases dipped to 38% in 2017. The dramatic rise in criminal cases involving immigration at the same time that public corruption cases were declining suggests that the federal investigative and prosecution resources focused on public corruption may have shifted to immigration cases. Figure 5 illustrates trends in referrals for prosecution in immigration, terrorism and public corruption cases over 30 years.

Discussion
Public corruption investigations referred to DOJ in the last twenty years have declined. If investigations are not completed, then prosecutions do not have any cases to bring to court; and therefore, the public corruption conviction rate will decrease as well. In the last twenty years, the decrease has been steady. Based on DOJ data, immigration may have become the proxy for terrorism or public corruption in law enforcement priorities. The question remains about whether the resources are justified to be directed towards immigration rather than public corruption.
Public corruption differs from any other type of investigation because of the resources, expertise and the length and breadth of such investigations. and their potential ramifications in the community. Federal investigators have a certain space between themselves and the community because most were transferred into the community and lack close local connections. Federal laws offer well-developed tools such as Title III wiretaps, undercover operations and the possibility of grand jury investigations.
Although federal law enforcement has had successful prosecutions, they face many obstacles to investigating public corruption. The obstacles expressed in multiple interviews noted above illustrate why corruption cases are not more common:

Immigration
Public Corruption &Terrorism Public Corruption Referrals Terrorism Immigration Referrals Figure 5. Public corruption, terrorism and immigration referrals 1986-2017.
(a) The extended time period (months or years) needed to make a public corruption case (given the need for multiple forms of corroboration to establish criminal intention).
(b) Significant investigative resources are required to carry out these investigations involving tracking down tips, document reviews, following monetary exchanges, and surveillance involving public officials, who often take precautions to insulate themselves from exposure.
(c) The federal government has the burden of prosecuting nearly all corruption cases in the US, given the problems of inadequate resources, training and insufficient distance from targets at the state and local levels, where very few corruption cases are made.
(d) The need to demonstrate active corruption, rather than historical cases based on documentary evidence from the past. The documentary record is easy for a suspect to explain away, so contemporary evidence obtained from others is usually needed. "You need to have witnesses at the table to show a quid pro quo. Judges want to see the quid pro quo" (Interview I-13). Showing active corruption and criminal intent often requires working insiders, informants, and/or electronic surveillance beyond mere tips, suspicious activity, and past observations. "We had these text messages from the parties. In our case, the text messages collaborated the other witnesses' testimony. It supported their testimony. [An insider] was a very useful narrator and explained the context of the messages." (Interview A-21). These investigative sources are difficult and time-consuming to cultivate and monitor.
(e) The potential consequences of public corruption investigations can be wide-ranging for the investigators, prosecutors and others involved in the case. The careers of public officials are sometimes ended and, in some cases, the career aspirations of investigators and prosecutors thwarted in retaliation by the powerful political targets or powerful friends of those targeted in corruption investigations. I-08 described adverse impacts well: "An AUSA lead prosecutor on many cases and mafia cases … took out lawyers, cops, fire inspectors, licensing inspectors.
[AUSA] went up for a federal judgeship. A Boston city councilman called her 'unladylike.' She never made a federal judgeship." Several recurring themes emerged in the 40 interviews of former investigators and prosecutors. These themes included the problem of a great deal of investigative time and resources needed, the fact that only the federal government (and not the states) are making corruption cases, the difficulty in proving the existence corrupt agreements, and the potential consequences to the investigators and prosecutors have been seen in multiple cases. According to one former investigator, "investigating political corruption is far more lethal than investigating other forms of crime. It is probably the most dangerous type of investigation an agency can undertake" due to the possible deleterious impacts on those investigating (Interview I-06). In addition, "Public corruption investigations leave many an innocent person in their wake and the collateral damage is quite extensive!" (Woodiwiss, 2015, p. 67). An example is provided by the dissolution of the Pennsylvania Crime Commission after it exposed high-level corruption in the state (Woodiwiss, 2015). In other cases, promotions were denied and ostracism occurred against those who initiated a public corruption investigation (Barer, Girardit, & Eurell, 2016;Carter, 2016;Martens, 2015;York, 2014).
In the words of a former investigator, "there are sufficient laws on the books to address public corruption … what is often lacking is the political will" (Martens, 2015, p. xi). The lack of political will is expressed by the failure of many agencies, agency leaders, public pressure and political leadership to insist on devoting the time required, investigative resources, making corruption cases at both the state and federal levels, and not to tolerate negative consequences against those who investigate and prosecute public corruption cases.
In a related way, organized crime cannot operate over long periods without the complicity or protection of public officials (Albanese, 2015;Martens, 2015). For example, corruption leads to public frustration in the failure of the government to deliver taxpayer-funded services in a fair and just manner. This abuse of public office leads to public cynicism, protest and sometimes to violent conflict. Therefore, ineffective investigation of public corruption has cascading effects to other crimes, as even the city riots of the 1960s and beyond were linked in part to a pervasive feeling of corruption among the public in the cities affected (Hartman & Lewis, 2002;National Advisory Commission on Civil Disorders, 1968). Despite the impediments to making public corruption cases, their importance is clear in their connection to other crimes (e.g. money laundering, nepotism, embezzlement, fraud), making sure taxpayer's fund are spent in accordance with the law, and that public servants act in the public interest and not in their own self-interest. The 40 interviews summarized here illustrate how public corruption cases are made and how they might be more successful in the future.