WE need to start using “big data” to mine and isolate people who have money but refuse to pay taxes. We have excellent databases in form of the National Database and Registration Authority, immigration and travel data, car registration data, utilities/cell phone data, and property ownership.

First, using data already with the Federal Board of Revenue (FBR), we need to create a profile of people who are non-salaried but pay taxes. This exercise is called constructing a “look-alike filter”. Attributes that make up this filter would have the following elements: travels more than twice outside the country, has car registered under name or family-tree name, owns property in own name or under family-tree name, and has phone bill in excess of Rs5,000 and electricity bill in excess of Rs 25,000.

People with the above profile are paying taxes. Next, we scroll through the existing four to five databases regarding travel, utilities, car and property ownership, looking for people who fit the above criteria but have never paid taxes.

However, a caveat: I believe that if this exercise is undertaken under the FBR, it will fail. We need to create a independent body under the private sector that will be responsible for mining this data, and staffed by people who have been involved in large-scale behavioural data analysis (very few available in Pakistan, but lots available in our non-resident population living in advanced markets such the US and the UK).

I firmly believe that this is only way to bring about a paradigm shift in our tax base. If done independently of the FBR, we can move almost overnight from one million to 10 million tax payers.

Mirza Kazim Namazi

Karachi