GST Without Handcuffs: Can Data Do the Work That Fear Did?

The Arrest Power Was Never About Arrests. It Was About Leverage.

The headline number in this week's GST proposals is the prosecution threshold, which goes from Rs 1 crore to Rs 5 crore. The more consequential change is quieter. Officials want to remove the arrest power under Section 69 of the CGST Act.

Their own stated reason is revealing: the power gave officers leverage over businesses before any arrest actually happened. That admission is the whole story. A deterrent works through its threat more than through its use.

The arrest power was a behavioural tool. It made a trader think twice before inflating input tax credit and it made a distributor think twice before buying an invoice that came with no goods. Remove it and the state has to replace that pressure with something else.

The proposals name that replacement: data.

Who Gains Immediately The winners are easy to identify.

• The small trader with a Rs 8,000 dispute. Responding to that notice costs more in a consultant's fees and lost working days than the amount at stake. Cases below Rs 10,000 make up about a fifth of all cases by number but a negligible share of the money involved. The proposal would drop them, both new and pending.
• The mid-sized firm with a genuine classification dispute. Its disagreement with the department is about interpretation, not fraud. Today that dispute carries a criminal shadow. Under the proposal, it becomes a civil matter settled through tax, interest and penalty.
• The testing lab paid in dollars. If place of supply follows the customer rather than where the work was done, testing, certification, repair, calibration and research services would qualify as exports.

Who Should Worry: The Fake-Invoice Operator Section 132

exists mainly for one kind of actor: the racket that creates shell firms, issues invoices for goods that never move and passes on input tax credit the state never collected. This is the test the reform has to pass. Consider how such a racket is built.

Its invoices are spread across many registrations, so the amount on any single one tends to be small. With a prosecution floor of Rs 5 crore, an operator who splits activity across enough shell registrations can keep each one under the line.

Without the arrest power, the person behind those registrations no longer faces a quick, personal consequence. That is why the "data" in this reform cannot be a slogan. It has to work at the network level. The system must link e-invoices, e-way bills and return filings across registrations, so that a cluster of small shell firms shows up as one large fraud.

If analytics can do that and then pursue the beneficiary, the threshold does not matter. If they cannot, the Rs 5 crore floor works like a design specification for evaders.

The Exam Insight This is not

simply "ease of doing business versus revenue." The reform replaces deterrence by fear with deterrence by detection probability. A rational evader weighs two things: the chance of being caught and the cost if caught.

These proposals lower the cost if caught and promise to raise the chance of being caught. The reform succeeds only if that second promise is kept and fake-invoice networks will show quickly whether it has been.