HM Revenue & Customs' business intelligence systems are being used to bring in more taxes according to the government.

Treasury Minister David Gauke MP told the Institute for Fiscal Studies conference in London this week, that HMRC's system were "improving compliance".

Gauke said HMRC is using "behavioural economics" to catch tax cheats. "HMRC has already made significant strides testing and embedding behaviour change within letters, forms, telephony and digital channels, and the results have been impressive, said Gauke.

"For next to no cost, HMRC has seen up to a 15 percent improvement in payment behaviour in some trials, simply by changing the way in which they communicate."

Using its CONNECT system HMRC can now cross-match over a billion pieces of data to enable them to segment taxpayers according to their behaviour and their past relationship with the exchequer, he said.

By applying analytical techniques, HMRC is able to separate those who are willing and able to pay their taxes, from those "who might need help or those who might push the boundaries of the law", said Gauke.

CONNECT also allows HMRC to ensure that data collected by one part of the department can be used across all of its compliance areas. It means new relationships can quickly and easily be uncovered between people and organisations that would otherwise remain hidden, so that fraud can be detected and stopped in a way that was not possible before.

Gauke used the example of inheritance tax. HMRC receives around 300,000 paper returns on bequeathed estates every year – including around 200,000 from estates claiming to be below the tax-paying threshold. Due to the very large number of returns received, it was very difficult to identify "high risk cases" where more tax was due than what was in fact declared.

To develop a more thorough and efficient means of enforcing compliance, HMRC experts developed a single risk code that sifts over 50 million lines of data to spot where estates might have been falsely submitted as exempt.

HMRC used the massive amount of information it held on property ownership and transactions, company ownerships, loans, bank accounts, employment history, and self assessment records that had previously been unmanageable. All of this was turned into a single code that indicated when the return was likely to be inaccurate and why.

As a result, said Gauke, HMRC interventions on non-taxpaying estates increased "many times over" and in the first year of operation an additional £26 million was raised from inheritance tax through the use of CONNECT.

Other examples cited were preventing hundreds of  millions of pounds in fraudulent VAT claims, and enabling identification of offshore non-compliance to yield over £50 million. Gauke said CONNECT has "already generated around £1.4 billion in additional tax yield".

Last year HMRC said it would be reducing its 31 business intelligence (BI) systems to just one, that it would use for the next four years. The move was part of a wider consolidation of applications used by HMRC to cut running costs.

Earlier this year the government was told it could "save up to £33 billion a year - the equivalent of £500 for each person in the country – by using public Big Data more effectively".

Free market think tank Policy Exchange said "£2–8 billion" of that £33 billion could come from collecting a greater share of unpaid taxes. It said HMRC could collect more unpaid tax by "accelerating the use of algorithms to mine data" and flag cases that need to be investigated to ensure the correct tax is paid.