Our Story
How one data team is rooting out procurement corruption in Kazakhstan
Finding evidence of corruption in paper-based procurement can be like looking for a needle in a haystack. But with new data tools, Kazakhstan is turning the spotlight on its bad actors — and recovering millions of dollars in the process.
Public procurement is a risky business — quite literally.
Procurement officers are tasked with buying goods and services for millions and even billions of dollars, and therein lies the challenge: because of the sheer volume of contracts and the astronomical figures involved, there will always be criminals who are tempted to game the system to their own advantage.
In Kazakhstan, for instance, one such dubious contract saw a kindergarten pay US$500 for every single ream of paper purchased. When public procurement is all paper-based or happens behind closed doors, it becomes a Herculean task to find these needles in the haystack, and the risk of corruption grows exponentially. But what if there was a way to spot risks in public procurement and catch bad faith actors before they are able to drain public budgets and public trust?
That’s what Datanomix, a company specializing in business analysis tools for large companies in Kazakhstan, set out in 2010 to do. So far, by their own estimates, they have been able to save the Kazakh government millions and provide documentation for hundreds of criminal cases.
Procurement officers are tasked with buying goods and services for millions and even billions of dollars, and therein lies the challenge: because of the sheer volume of contracts and the astronomical figures involved, there will always be criminals who are tempted to game the system to their own advantage.
In Kazakhstan, for instance, one such dubious contract saw a kindergarten pay US$500 for every single ream of paper purchased. When public procurement is all paper-based or happens behind closed doors, it becomes a Herculean task to find these needles in the haystack, and the risk of corruption grows exponentially. But what if there was a way to spot risks in public procurement and catch bad faith actors before they are able to drain public budgets and public trust?
That’s what Datanomix, a company specializing in business analysis tools for large companies in Kazakhstan, set out in 2010 to do. So far, by their own estimates, they have been able to save the Kazakh government millions and provide documentation for hundreds of criminal cases.
A spotlight on risky procurements
For their first project, launched in 2010, Datanomix partnered with Kazakhstan’s Financial police. While all procurement was still paper-based, the data on payments from the Treasury had been made available online. Datanomix used this data to create a set of ‘risk indicators’ — algorithms that can automatically calculate and highlight potentially risky public procurements.
Within just six months, the financial police had used this data to open 120 criminal cases, netting the government $77 million in recovered costs.
“This caused a mental shift,” Vitaly Trenkenshu, a managing partner at Datanomix, explains. “We showed law enforcement officers that they could find risks and solve economic crimes without even having to leave their desks.”
Spurred on by this success, the financial police launched the Government Situation Center analytical system, which, like a vacuum, pulled data from about 50 government databases. From this data, the Datanomix team developed more than 100 risk indicators, covering both fraud in procurement as well as other financial crimes such as tax evasion.
These indicators helped law enforcement officers find the procurements with the highest risk of malfeasance, such as procurement with the participation of family members. To Vitaly’s surprise, this was a common practice, and there were many instances where government officials gave contracts to their brother or wife’s company without even batting an eyelid.
In just two years, from 2013 to 2014, the financial police officers were able to document damages of $647,000,000 to the government, resulting in over 630 criminal cases. Vitaly is convinced that working with supervisory agencies can yield a significant economic effect:
“If you find some kind of scheme, a whole team of qualified people will take it on. They understand what economic crimes are and can stop them.”
Since then, risk indicators have been developed for all public bodies investigating corruption and financial crime in Kazakhstan — the Prosecutor General’s Office, the Economic Investigation Service, the Anti-Corruption Service, the Internal State Audit Committee, and many others. To start collaborating with each of them, Datanomix launched pilot projects at the company’s own expense to show the real, tangible results of this approach. Vitaly notes that this was key in overcoming the hesitation that was initially prevalent in many of these organizations.
For their first project, launched in 2010, Datanomix partnered with Kazakhstan’s Financial police. While all procurement was still paper-based, the data on payments from the Treasury had been made available online. Datanomix used this data to create a set of ‘risk indicators’ — algorithms that can automatically calculate and highlight potentially risky public procurements.
Within just six months, the financial police had used this data to open 120 criminal cases, netting the government $77 million in recovered costs.
“This caused a mental shift,” Vitaly Trenkenshu, a managing partner at Datanomix, explains. “We showed law enforcement officers that they could find risks and solve economic crimes without even having to leave their desks.”
Spurred on by this success, the financial police launched the Government Situation Center analytical system, which, like a vacuum, pulled data from about 50 government databases. From this data, the Datanomix team developed more than 100 risk indicators, covering both fraud in procurement as well as other financial crimes such as tax evasion.
These indicators helped law enforcement officers find the procurements with the highest risk of malfeasance, such as procurement with the participation of family members. To Vitaly’s surprise, this was a common practice, and there were many instances where government officials gave contracts to their brother or wife’s company without even batting an eyelid.
In just two years, from 2013 to 2014, the financial police officers were able to document damages of $647,000,000 to the government, resulting in over 630 criminal cases. Vitaly is convinced that working with supervisory agencies can yield a significant economic effect:
“If you find some kind of scheme, a whole team of qualified people will take it on. They understand what economic crimes are and can stop them.”
Since then, risk indicators have been developed for all public bodies investigating corruption and financial crime in Kazakhstan — the Prosecutor General’s Office, the Economic Investigation Service, the Anti-Corruption Service, the Internal State Audit Committee, and many others. To start collaborating with each of them, Datanomix launched pilot projects at the company’s own expense to show the real, tangible results of this approach. Vitaly notes that this was key in overcoming the hesitation that was initially prevalent in many of these organizations.
Seeing risks in real-time
In 2015 Kazakhstan developed a new electronic public procurement portal, which for the first time made data about procurement open to the public. This is the fuel that Datanomix uses to identify suspicious contracts. Here’s how it works:
Every night, Datanomix automatically reloads all data from the procurement portal into the “Eye”, the company’s analytical system, which is based on Qlik Sense.
This data contains all public procurement plans, announcements, lots, contracts, procuring entities, and suppliers — roughly equivalent to 100 million individual records. Reloading the data is necessary to ensure the accuracy of the information because government purchases can be amended retroactively by court order.
In addition to data on public procurement, Datanomix collects payments from the treasury, the register of state-owned enterprises and institutions, the register of legal entities, and certificates of origin of goods and other registers.
The system then compares the data with a set of known indicators to identify potential risks in public procurement.
In 2015 Kazakhstan developed a new electronic public procurement portal, which for the first time made data about procurement open to the public. This is the fuel that Datanomix uses to identify suspicious contracts. Here’s how it works:
Every night, Datanomix automatically reloads all data from the procurement portal into the “Eye”, the company’s analytical system, which is based on Qlik Sense.
This data contains all public procurement plans, announcements, lots, contracts, procuring entities, and suppliers — roughly equivalent to 100 million individual records. Reloading the data is necessary to ensure the accuracy of the information because government purchases can be amended retroactively by court order.
In addition to data on public procurement, Datanomix collects payments from the treasury, the register of state-owned enterprises and institutions, the register of legal entities, and certificates of origin of goods and other registers.
The system then compares the data with a set of known indicators to identify potential risks in public procurement.
To develop these indicators, the Datanomix team analyzed current best practices from leading institutions such as the World Bank, OECD, EBRD, OCP, and adapted relevant risk indicators to the public procurement system of Kazakhstan. They then added indicators inspired by the methods of law enforcement agencies in Kazakhstan.
From this, Datanomix created a list of 43 individual risk indicators that could be calculated automatically, reaching across all stages from planning to contract execution — a smaller list than the 100 indicators that were developed for the financial police since this system more narrowly focuses solely on procurement.
One risk indicator is short delivery times. Vitaly explains that there were many cases where government procuring entities would announce a tender to supply, for instance, thousands of computers in just two weeks. In such cases, the affiliated supplier likely knew about the procurement in advance, giving them an unfair advantage over potential competitors. For the first time, oversight authorities could now quickly identify and respond to these potentially fraudulent contracts in real time.
While there’s nothing stopping other countries from emulating the Datanomix approach, many countries lack the necessary data to calculate, for instance, median prices and pricing anomalies. Vitaly Trenkenshu applauds the Center for Electronic Finance of Kazakhstan for making their data available in a machine-readable format, without which the Datanomix system wouldn’t be able to function.
“In addition, there is a Unified Directory of Goods, Works and Services in Kazakhstan. It is much more detailed than the Common Procurement Vocabulary of the European Union, and helps to build reasonably accurate price analysis in the system,” he says.
From this, Datanomix created a list of 43 individual risk indicators that could be calculated automatically, reaching across all stages from planning to contract execution — a smaller list than the 100 indicators that were developed for the financial police since this system more narrowly focuses solely on procurement.
One risk indicator is short delivery times. Vitaly explains that there were many cases where government procuring entities would announce a tender to supply, for instance, thousands of computers in just two weeks. In such cases, the affiliated supplier likely knew about the procurement in advance, giving them an unfair advantage over potential competitors. For the first time, oversight authorities could now quickly identify and respond to these potentially fraudulent contracts in real time.
While there’s nothing stopping other countries from emulating the Datanomix approach, many countries lack the necessary data to calculate, for instance, median prices and pricing anomalies. Vitaly Trenkenshu applauds the Center for Electronic Finance of Kazakhstan for making their data available in a machine-readable format, without which the Datanomix system wouldn’t be able to function.
“In addition, there is a Unified Directory of Goods, Works and Services in Kazakhstan. It is much more detailed than the Common Procurement Vocabulary of the European Union, and helps to build reasonably accurate price analysis in the system,” he says.
New challenges
No system is perfect. Vitaly highlights three challenges that Datanomix is currently working to contain: incentives for law enforcement officials to catch criminals red-handed, the ingenuity of criminals, and corruption among law enforcement officials themselves.
No system is perfect. Vitaly highlights three challenges that Datanomix is currently working to contain: incentives for law enforcement officials to catch criminals red-handed, the ingenuity of criminals, and corruption among law enforcement officials themselves.
- 1. Incentives for law enforcement officersThe Datanomix system distinguishes between risks at the planning stage and risks at the award or contract implementation stage. While it is much more beneficial to identify risks at the planning stage, law enforcement officers are often more focused on trying to catch people “red-handed”, Vitaly explains:
“Identifying risks in the planning phase is better because that has a preventive effect. The money has not gone to the supplier yet, and it is possible to intervene by, for instance, conducting a competitive purchase, which saves money”.
“But the system of rewards among law enforcement incentivizes officers to try to catch people red-handed. For them, the main goal is to prove the damage done to the state and open a criminal case,” Vitaly says. - 2. Loopholes and criminal ingenuityHuman imagination is limitless. And since criminals are also human, it goes to reason that they will attempt to find and exploit loopholes in public procurement systems.
Datanomix has a list of known criminal schemes that the company identified through public procurement monitoring and were able to terminate in cooperation with the authorities. But new schemes keep emerging.
At the beginning of their collaboration with the financial police, Vitaly recalls how they learned how bad-faith actors in business and government used shell companies to launder money and evade taxes by registering the companies in the names of unwitting homeless people. When this scheme was exposed, they instead began registering companies in the names of stage four cancer patients. When the supervisory authorities caught on to them and arranged an inspection, the “owners” of the companies had often passed away. Now, fictitious companies are registered in the name of foreign citizens who leave far fewer financial traces. The Datanomix team developed risk-indicators to identify and prevent such scenarios.
Now, there is a popular scheme in Kazakhstan involving the author certificate. If a supplier has the author certificate for a unique product or service that the procuring entity needs, it is allowed to single-source. The trouble is that you can get the author certificate for anything. It is issued on the same day upon request with no expert review; you just need to pay the state fee. Therefore, suppliers abuse it and register IP for anything. So far, there have been cases where suppliers obtained author certificates for a perpetual motion machine and even the Linux operating system.
Datanomix flagged this issue to the National Chamber of Entrepreneurs, which protects the interests of business in Kazakhstan, and the Prosecutor General’s Office.
“As a result, the Ministry of Finance is putting this problem on their agenda, and this loophole will probably be patched up in the new version of the Law,” says Vitaly. “Collaborating with government bodies is an opportunity to impact policies and stop systemic violations quite quickly.” - 3. Flipside of risk indicatorsThe final risk involves the law enforcement officers themselves, Vitaly explains.
There have been instances where field operatives — upon seeing a triggered risk indicator in a specific procurement — have extorted companies in return for turning a blind eye to the triggered indicator.
Vitaly does not yet have a clear solution for how to prevent this ‘catch-and-release’, but highlights the importance of leadership in supervisory and law enforcement agencies, who need to monitor work on the ground to root out ‘shadow business’. Making inspection results completely transparent and open to the public could be one solution, but according to Vitaly this is unlikely to work, since most of the risks in Datanomix are ‘closed’ and can’t be opened to the public, as they contain confidential commercial and tax information. This information is protected by law, and therefore can’t be published.
A similar risk is associated with auditors. Procuring entities carry out thousands of procurement transactions a year, and an auditor can approach companies and offer to overlook certain stipulations for a bribe. Now, Kazakhstan is moving towards defining the scope for audits in an automatic and data-driven way, which according to Vitaly should significantly limit corruption risks.
In addition, Datanomix signed several contracts to analyze public procurement in municipalities. This enables local authorities to control the purchases of their subordinate organizations more effectively, work to prevent errors and quickly fix them until there is public interest and supervisory authorities show up.
So far, the collaboration has been fruitful, Vitaly says, but new challenges emerge:
“Municipal leaders have been positive, but public procurement officers have entirely different headaches and KPIs. Their main job is to use all budgetary funds on time. Therefore, they would rather organize a purchase without savings than not conduct it at all,” explains Vitaly. And he sees that even if municipalities do buy their system to make reporting easier, they do not use the risk indicators.
Competition matters
Next up on Datanomix’s agenda is tackling the high rate of non-competitive purchases in Kazakhstan.
As an average over the last five years, the share of non-competitive purchases has been at a staggering 63.8%. Non-competitive procurements often lead to overpricing, lobbying the interests of one supplier and corruption risks.
To meet this challenge, Datanomix is further developing their “Eye” business intelligence system to increase the level of competition in Kazakhstan’s public procurement.
This year, the company participated in the Open Contracting Innovation Challenge, an innovation contest to support the development of tools based on open data on public procurement in Kazakhstan, Kyrgyzstan, and Moldova. Datanomix was one of the three teams from Kazakhstan that won the main prize of this challenge: $15,000 USD to implement their idea.
Next up on Datanomix’s agenda is tackling the high rate of non-competitive purchases in Kazakhstan.
As an average over the last five years, the share of non-competitive purchases has been at a staggering 63.8%. Non-competitive procurements often lead to overpricing, lobbying the interests of one supplier and corruption risks.
To meet this challenge, Datanomix is further developing their “Eye” business intelligence system to increase the level of competition in Kazakhstan’s public procurement.
This year, the company participated in the Open Contracting Innovation Challenge, an innovation contest to support the development of tools based on open data on public procurement in Kazakhstan, Kyrgyzstan, and Moldova. Datanomix was one of the three teams from Kazakhstan that won the main prize of this challenge: $15,000 USD to implement their idea.
Vitaly explains their concept with an example: procurement of office paper. The median price in competitive purchases is $2,82 USD per ream of paper. The median price in direct agreements is $3,14 USD per ream — 10% more than in competitive purchases. Currently, 45% procurements for paper are done through direct agreements. While the difference per sheet of paper is marginal, it quickly adds up.
So what’s the solution?
First, the system identifies unreasonable restrictions on competition in procurements. For example, dividing procurements of goods and services so they don’t exceed certain price thresholds, which can be abused to procure through single-sourcing. There are 55 types of justifications to conduct single source procurements in Kazakhstan, and each is a source of procurement risks that should be taken into account.
Then, the system automatically generates and sends official letters to procuring entities and their governing bodies, supervisory and law enforcement agencies to inform them of the triggered risk indicators.
But according to Vitaly, this is unlikely to close the loophole by itself, since it will require more work from officials at the procuring entities — work that they are currently not paid to do. However, the letter is also copied to the procuring entity’s governing officials — the municipal and regional executive bodies. They are obliged by law to respond. The expectation is that they will turn to the procuring entity and demand an open competitive procedure.
The letters will also go to the Economic Investigation Service, which will be able to monitor the movement of funds for these tenders, as well as the Prosecutor General’s Office, the Anti-Corruption Service, and the Internal Audit Committee under the Ministry of Finance. And the procuring entity will be able to see these recipients cc’d in the email.
So what’s the solution?
First, the system identifies unreasonable restrictions on competition in procurements. For example, dividing procurements of goods and services so they don’t exceed certain price thresholds, which can be abused to procure through single-sourcing. There are 55 types of justifications to conduct single source procurements in Kazakhstan, and each is a source of procurement risks that should be taken into account.
Then, the system automatically generates and sends official letters to procuring entities and their governing bodies, supervisory and law enforcement agencies to inform them of the triggered risk indicators.
But according to Vitaly, this is unlikely to close the loophole by itself, since it will require more work from officials at the procuring entities — work that they are currently not paid to do. However, the letter is also copied to the procuring entity’s governing officials — the municipal and regional executive bodies. They are obliged by law to respond. The expectation is that they will turn to the procuring entity and demand an open competitive procedure.
The letters will also go to the Economic Investigation Service, which will be able to monitor the movement of funds for these tenders, as well as the Prosecutor General’s Office, the Anti-Corruption Service, and the Internal Audit Committee under the Ministry of Finance. And the procuring entity will be able to see these recipients cc’d in the email.
A dollar saved is a dollar earned
Over the last decade, the collaboration between Datanomix and government agencies in Kazakhstan has saved millions of dollars in taxpayers’ money, according to Datanomix. These savings, in turn, can be spent to procure more medicine, schooling equipment, and overhaul road infrastructure in Kazakhstan.
“We are planning to organize pressure and get a conversion this way. Procuring entities will cancel these purchases, making them competitive instead. This way, we will achieve cost savings. If even 5% of transactions switch to open tenders or requests for proposals, this will yield savings of at least 10 billion tenges per year for budgets at all levels,” says Vitaly.
And this can become another story about how the Datanomix team made public procurement in Kazakhstan more efficient—big time.
Source: open-contracting.org
Over the last decade, the collaboration between Datanomix and government agencies in Kazakhstan has saved millions of dollars in taxpayers’ money, according to Datanomix. These savings, in turn, can be spent to procure more medicine, schooling equipment, and overhaul road infrastructure in Kazakhstan.
“We are planning to organize pressure and get a conversion this way. Procuring entities will cancel these purchases, making them competitive instead. This way, we will achieve cost savings. If even 5% of transactions switch to open tenders or requests for proposals, this will yield savings of at least 10 billion tenges per year for budgets at all levels,” says Vitaly.
And this can become another story about how the Datanomix team made public procurement in Kazakhstan more efficient—big time.
Source: open-contracting.org
Request a Demo
Resources
Company
See the advantages of using our platform for your procurement
See our award-winning solutions in Action
Recognized by
