Corruption in the developing world is a major cause of poverty.
Corruption is certainly not limited to poor nations, and since the 1980s, western countries have been able to use Information and Communications Technology (ICT) to bring corrupt officials to justice.
Given this, how can developed countries use more sophisticated means to deter corruption by early intervention, and secondly how can ICT help poor countries attack corruption?
Donald Cressey, an American criminologist, recognised that corrupt activity requires three elements: opportunity, motivation and self justification, forming the Fraud Triangle and that removing one of these elements will reduce levels of corruption.
In developed countries, where ICT systems manage and record all day to day activities, a mountain of data is collected as a by-product of the running the business, including interactions between people.
This electronic data provides a source of raw material for data and text mining; advanced analysis methods that draw on statistics and artificial intelligence to find unseen patterns in data.
In these data rich environments, these techniques can help break any element of the Fraud Triangle.
But how to break the fraud triangle? The opportunity to be corrupt is reduced as the likelihood of being found out increases. Abnormal patterns of behaviour can provide warning indicators. In law enforcement these might include police officers who issue low numbers of on the spot fines, or are the subject of many citizen complaints.
Self justification can be revealed through an analysis of staff emails, using sentiment analysis based on text mining methods, recognising telling phrases like “I deserve it”, “Everyone is doing it”.
Indeed, some experts in the field claim that warning signs of wrong doing can always be found in emails. This is a technique used by a major accountancy company as part of its forensic practice.
Motivation is invariably driven by some financial need. Rule based methods perhaps using HR information (“if employee A is a heavy gambler then risk level is high”) or employee email content analysis can identify changes in personal circumstance that are increasing financial pressure.
More importantly than using advanced detection methods, all anti-corruption systems, wherever they are situated, must be able to take all cases of potential or actual corruption to an appropriate conclusion.
They need strong capabilities: for example, they must enforce business processes, exercise strong access control, guarantee security and guard against corrupt operators.
Anti-corruption systems using advanced technologies can identify corruption at an early stage and help to prevent it. For example, some police forces in USA have moved to prevent corruption through computer based early intervention systems.
These developments are attracting interest from Australian governments. These systems break the fraud triangle using data gathered in the organization and offer real hope that corruption can be overcome in many organisations by identifying people at risk of becoming corrupt.
What of the developing world? The Spider Centre, a Swedish program focussed on ICT in developing regions recommends using technology to empower people and communities against corruption.
This includes measures to remove monopolies of access to information by officials and also their ability to exercise their discretion over its use, recognise that technology is not a complete solution, and that education plays an essential role in reducing corruption.
An example of an anti-corruption system operating in a developing country is the system used by KACC, the Kenyan Anti-Corruption Commission. It supports both public complaints and case management. Whilst it does not use advanced analytical methods, it provides a robust system that offers strong confidentiality and enforces processes.
Looking forward, the availability of more electronic data will increase allowing more advanced techniques to be applied.
In this way, better methods of early corruption detection can be applied in developing countries.