One of the main reasons policymakers cite for establishing a standalone, independent anticorruption agency is the need to strengthen the enforcement of their nation’s laws against bribery, conflict of interest, and other corruption crimes. In the past 25 year some 150 countries have created a specialized, independent agency to fight corruption (De Jaegere 2011), and virtually all have been given the lead responsibility for investigating criminal violations of the anticorruption laws. But while a broad international consensus exists on the value of creating a new agency with investigative powers, opinion remains sharply divided on whether these agencies should also have the power to prosecute the crimes it uncovers. As this is written, Indonesian lawmakers are considering legislation to strip its Corruption Eradication Commission (KPK) of the power to prosecute while a bill before the Kenyan parliament would grant its Ethics and Anticorruption Commission (EACC) the power to prosecute the cases it investigates.
No matter the country, debate about whether a single agency should have the power to both investigate and prosecute corruption cases inevitably comes down to a small set of conflicting claims. Those who oppose giving a single agency both powers raise an argument at the center of the older debate about the relative responsibilities of police and prosecutors — investigator bias. In the words of a British Royal Commission that studied the relationship between English police and prosecutors, an investigator “without any improper motive . . . may be inclined to shut his mind to other evidence telling against the guilt of the suspect or to overestimate the strength of the evidence he has assembled.” That is, once an investigator hones in on a suspect, confirmation bias sets in, and he or she will interpret all evidence as supporting the suspect’s guilt. Putting the decision about whether to prosecute a case in an agency wholly separate from the one that investigates provides a strong check against such bias, reducing the chances that the innocent will be put to a trial or weak cases brought to court.
The investigator bias argument has a long and distinguished pedigree, and a 2011 survey of the powers of 50 anticorruption agencies by World Bank economist Francesca Recanatini found that it often carries the day. Only half of the 50 agencies she surveyed have both investigative and prosecutions powers. But as the contemporary debates in Indonesia and Kenya suggest, proponents of combing investigation and prosecution in a single agency have a very powerful counter argument in their corner.
A quarter century of experience with both anticorruption agencies shows that those with the responsibility for both investigating and prosecuting corruption are likely to be more effective in enforcing the anticorruption laws. One obvious giveaway is to look at who backs and who opposes the proposed Indonesian and Kenyan legislation. In Indonesia the drive to take away the KPK’s power to prosecute is led by those who want to weaken the agency; it is opposed by civil society groups, the media, and others who support vigorous enforcement of the nation’s anticorruption laws. In Kenya the legislation to give the EACC prosecution power is backed by the Kenyan chapter of Transparency International and others wanting the nation’s corrupt elite prosecuted and convicted; it is bottled up in a parliament where many suspect the integrity of the majority has been compromised.
To be sure, some anticorruption agencies have succeeded without having prosecutorial powers, the Hong Kong and Singapore agencies being notable examples; at the same time there are many examples where agencies possessing both investigative and prosecutorial power have been ineffective. But by and large dividing the powers between two agencies makes the effective enforcement of the anticorruption laws all the more difficult to achieve.
The reasons are several. First and foremost is that unlike robbery, homicide, and other “street crime,” where the facts and the law are relatively straightforward, corruption offenses are complex, involving often convoluted and obscure transactions that raise complicated factual and legal issues. As then Director of South African’s National Prosecution Service Bulelani Ngcuka told a conference on modern criminal justice in 2001, in such cases it is critical “that the prosecutor who will ultimately conduct the prosecution [be] intimately involved in the investigation of the case [so that he or she] can exercise greater control over how the evidence is collected and can ensure that the evidence will be found to be admissible in court.” Housing prosecutors in a separate agency makes it very hard, if not impossible, to realize this involvement. (As an aside, the Scorpions, the nickname of the team of prosecutors and investigators Ngcuka oversaw to investigate and prosecute high level corruption, were so successful that the targets of their efforts succeeded in having them disbanded.)
Effective enforcement of all laws demands, as the Britain Attorney General told the House of Commons during its review of police-prosecutor relations, “a close, professional, and robust relationship” between investigators and prosecutors. But again, if the two are in separate agencies, this goal is all the harder to achieve. When prosecutors decline to prosecute a case an investigator has worked long and hard developing, it is easy for the investigator to treat it as a personal slight. Furthermore, prosecutors often need a good deal of time to review a lengthy file, but delays are commonly seen by civil society as a lack of commitment to prosecuting corruption case or, when after a delay the file is rejected, that corruption has have infected the review process.
Identifying who is accountable for weak enforcement of the anticorruption laws is virtually impossible when responsibility is split between two agencies. Investigators say weak enforcement is the prosecutors’ fault; they are refusing to prosecute solid cases. Prosecutors point the finger at the investigators, saying the cases they present are weak. As happened so often in the past in Kenya, the head of the anticorruption agency and the head prosecutor can get into a public spat over who is to blame, leaving the public unsure who is at fault while poisoning relations between working level staff in the two agencies.
Investigator bias is a real concern, but there are ways to address it short of putting the responsibility for investigation and prosecution in separate agencies. Close monitoring by the legislature or an independent review commission is one solution. Procedural changes that allow courts to weed out weak cases early in the process are another. And of course better, and public, data on the outcome of investigations will allow all to see if investigator bias is present. How many investigations matured into court cases? What percentage of cases resulted in acquittal? Dismissal at an early stage in the proceeding? Conviction?
After studying police-prosecutor relations in a number of countries, another of Britain’s many Royal Commissions tasked with recommending reforms to the criminal justice system concluded that no organizational arrangement “so clearly succeeds in resolving the problems which arise in any system of criminal justice that it furnishes an obvious model which all others should adopt” (chapter 1, para 13). That may be the case with the much broader question of how to structure police-prosecutor relations, but on the narrower question of whether anticorruption agencies should have both the power to investigate and prosecute, 25 years experience across a wide range of countries with different legal traditions points clearly to the need to consolidate these powers in a single agency.
GAB | The Global Anticorruption Blog 
To my surprise, I actually think I agree with you Rick. Maybe not for every country and in every instance, but I do think consolidating corruption and prosecution in one entity would be a good idea as a general rule.
I’d like to add to your argument a general point about what I think distinguishes anticorruption prosecutions from regular prosecutions, and therefore justifies consolidating functions. Anticorruption prosecutions are often intensely political. Powerful actors in public life have strong incentives to stall, obfuscate, and call in political favors to influence their outcome. Having separate offices might be a way for investigators and prosecutors to hold each other accountable, but more likely than not, they will simply serve as another veto point along the way to successful conviction. Further, anticorruption investigations often capture public attention and popular press, increasing the likelihood that improprieties will be noticed and be reported. Because the risks of anticorruption efforts stalling seem to be higher than ordinary crimes, and the risks of investigator bias may be lower (though this admittedly won’t always be true), consolidating functions makes solid, practical sense.
I support Rick and Courtney arguments and like to share here experience from Nepal that goes beyond investigation and prosecution debate. The first constitutional anti-graft agency in Nepal “Commission for the Prevention of Abuse of Authority (CPAA)” established in 1977 was bestowed with the third power, that is, to adjudicate corruption crimes. In spite of such a sweeping power, the agency virtually remained defunct. In 1990, with the change in regime, the country established Commission for the Investigation of Abuse of Authority (CIAA) giving with investigative and prosecution power. The adjudication power rested with the court.
In 2005, the then royal government taking advantage of the political situation, established another body called Royal Commission on Control of Corruption (RCCC) with a sweeping power similar to CPAA. But the body was declared extra-constitutional by the Supreme Court and got dissolved.
The regime change in 2006, particularly, during the drafting of the New Constitution in 2011, there was a brief debate over taking away prosecution power from CIAA. The then elected Constitution Assembly (CA) failed to deliver new constitution and new elections had to be held in 2013. In September 2015, we now have a new constitution where CIAA as an agency is retained but, this time, the law makers curtailed its power to investigate “improper conducts”. Earlier CIAA had a power to investigate and prosecute corruption crime and improper conducts by public officials. I will say the debate is still not over: there is now an overt conflict between CIAA and the government. Quiet an interesting drama to watch.
Thanks for the interesting observations. The Chinese procuratorate both investigates and prosecutes corruption cases and that particular institutional design has not been regarded as an advantage. I guess the argument for merging investigative and prosecutorial functions is valid in a mature legal system where criminal trial actually matters. This may not be the case in many of the high corruption countries.
This is a very nice summary of the conflicting considerations in this difficult institutional design question, and I don’t disagree with anything you say. I did, however, want to bring to your attention recent work by Gabe Kuris, in particular his paper “Watchdogs or Guard Dogs?” in Policy & Society (vol. 34(2), 2015) (http://www.sciencedirect.com/science/article/pii/S144940351500017X). Interestingly, he finds (based on a series of case studies), that anticorruption agencies “with teeth” are not always more effective than those that can only investigate and publicize — often because the latter are given more freedom of action, precisely because they are seen as less threatening. Indeed, some of your examples of “effective” ACAs (with prosecutorial power) getting shut down because they appeared too threatening might be further evidence in support of Kuris’s observations.
More food for thought, I suppose.
The news report below circulated by the Asia-Pacific Integrity in Action Network further illustrates the problems that can arise when the power to investigate and to prosecute are housed in separate agencies —
“At a ceremony held to mark the International Anti-Corruption Day, Maldives Anticorruption Commission member Aminath Minna said prosecution and convictions in corruption cases are rare world over. “If the few number of cases forwarded for prosecution are not filed at the court, it’s an issue we’ve to think about,” she said.
“Minna defended her commission’s track record saying that it had always sent cases, if backed by evidence, to the prosecutor general’s office. Prosecutors, however, reject many cases, she said.
The commission member also expressed hope that they would be able to have a better working relationship with the newly appointed prosecutor general Aishath Bisham.”
Source: http://www.haveeru.com.mv/news/64677#
My instinct is to agree that, in most cases, it is best to house the investigative and prosecutorial powers in a single anticorruption agency. I have some concern, however, that the potential responses to investigator bias leave something to be desired. Specifically, I remain doubtful that putting oversight in the hands of other agencies of government (potentially one or more of which includes members who are the subject of investigations and/or prosecutions by the anticorruption agency) will be effective. This actually gets to Courtney’s point about politicization, in that there is a potential that an entity that investigates and prosecutes could brush aside oversight efforts as attempts by the corrupt to tamp down on anticorruption efforts.
Of those listed in the post, I think the most promising idea to combat investigator bias is through dissemination of data to the public. Data can empower, but it is limited by how much power the public has in a given country to influence institutions, including anticorruption institutions. So long as the anticorruption agency is responsive to public pressure, there is reason to be less concerned about investigator bias. Of course, there is the chance that an anticorruption agency with the power to investigate and the power to prosecute could overstep its bounds and become more powerful than the entities and people it is supposed to be investigating and prosecuting. At that point, I am not sure how the investigator bias problem could be solved, or how to decrease that entity’s power.
There are certainly advantages to housing prosecution and investigation within the same entity, but I wonder about the definition of effectiveness we are using in this context. As Rick mentions, convictions are much more likely if the same actor investigates and prosecutes because of investigator/confirmation bias. If convictions are the standard for anticorruption effectiveness, then, such arrangements will be more “effective” systematically. Do single unit systems remain more effective if we measure anticorruption progress with a metric beyond convictions, though?
For those following the issue, the view of a senior Azeri anticorruption official —
KUALA LUMPUR: Many of the problems with anti-corruption bodies around the world stem from their lack of power to prosecute, said a senior anti-corruption figure today. Elnur Musayev, a senior prosecutor in Azerbaijan’s Anti-Corruption Department said . . . “A lot of problems with anti-corruption agencies all over the world arise when they submit investigation papers to the prosecution who can decide whether to prosecute or not,” he said adding, Azerbaijan’s system allowed for the anti-corruption department to both investigate and prosecute corruption cases.
Free Malaysia Today, Jan. 20, http://www.freemalaysiatoday.com/category/nation/2016/01/20/issues-arise-when-anti-graft-bodies-lack-power-to-prosecute/
Pingback: New Case Studies on Specialized Anticorruption Courts in Indonesia, the Philippines, Slovakia, and Uganda | Anti Corruption Digest
Pingback: When Should We Put Anticorruption Agencies in the Constitution? | Anti Corruption Digest
Pingback: “Jan Lokpal not Joke Pal” – A Review of India’s Anti-Corruption Efforts – Trot Along