Page 24 - GSTL_16 April 2020_Vol 35_Part 3

J44 GST LAW TIMES [ Vol. 35
GSTR-2A and Annual returns under GSTR-9/9C when more data and matching
would take place.
Initial data mining showed that companies were purchasing products at
high prices but there was a mismatch with sales indicating that the liability was
discharged using ITC in excess of 90-95% in some cases. In other words, payment
of tax from PLA/Cash Ledger was less than 5%, It was also noticed that in such
products the value addition ought to be higher and therefore more tax should
have been paid through the account current. Many of the companies recorded
large raw material purchases but little sales, suggesting that they may be either
under-billing or selling their products at lower prices officially and having finan-
cial flow back from other channels.
It is believed that other reasons such as export or SEZ supply or credit
accumulation would have been considered before issuing the Notice. Without
going into the merits or demerits of the case nor the outcome of the case but what
is significant is the detection of revenue leakage through the data mining.
Whereas, the department at present have a robust system of risk analysis
in Customs it has now been effectively introduced the same in GST by establish-
ing a full fledged Directorate of Analytics and Management (DGARM). It is due
to the constant endeavours of the officers that a large number of tax evasion and
fraud cases have been detected by the Directorate and subsequently action was
taken by the field formation. But the efforts are limited to the taxpayers who are regis-
tered with the department and who have filed returns or somehow have mapped into the
system. However, no mechanism seems to be in place to tap the revenue leakage for the
entities that are liable to tax but still chose not to pay. Similar is the case when the
assessee deliberately suppresses his actual turnover by not issuing the in-
voice/bill. In absence of any data these unscrupulous suppliers remain elusive.
In these cases, either the taxability needs to be detected by using some alternate
information platforms including manual/electronic surveillance or any other
suitable source. An Artificial Intelligence (AI) driven approach can fill up the
void as it is not possible for any tax agency to control, manage and check every
single taxpayer.
It is also a waste of scarce enforcement resources to routinely examine
low-risk, more or less compliant taxpayers, not to mention an inconvenience that
is likely to be felt by various stakeholders alike.
As Artificial Intelligence (AI) is the technique where data, which is dug
out with various tools, is used, and the system builds algorithms that help it de-
termine the right way towards better tax compliance and enforcement of new tax
regime.
Therefore, the use of analytics to manage these challenges is bound to in-
crease in the coming years which may begin with identifying suitable cases for
audit or for detailed scrutiny of records by the enforcement branches, return fil-
ing and payment compliance.
The source of intelligence/information can be utilised in designing some
risk parameters in GST regime. For example when a person receives a sales pro-
motion message about CCD (Coffee Cafe Day : a popular eating joint) or from a
popular chain of restaurants or for that matter from a rental car company then it
GST LAW TIMES  16th April 2020  144