Utility of GST data for lending is proven, especially when it comes to lending to small businesses. The availability of data over APIs, recency, relevance, and authenticity of data source i.e. the GST system itself, makes it the best fit for large scale digital lending programs.
Small businessesare often in the need of funds for their routine operations as well as capital expenditure. However, historically they have remained underserved because of the high cost of acquisition and servicing. But an absolute digital lending journey completely changes the game. And in a digital-first approach, GST data, along with other emerging frameworks like Account Aggregators, makes credit evaluation and underwriting efficient and comprehensive.
GST Data is vast and there are steps to be followed to get the same in underwriting applications. It is important to understand the finer points as these may directly or indirectly influence the way digital lending process is designed and underwriting models are defined.
Here are some points to be noted while consuming GST data for lending
1. GST data is of two types – Public and Private
Public GST Data: For any GST registered taxpayer, the basic registration details, current status and history of return filed are easily available for public consumption. The public data includes information such as legal name, the constitution of business i.e. whether it’s a proprietor, a firm, company etc. The main location and additional locations from where the business operates is also public data. The filing history gives details about the GST returns filed, date and mode of filing (whether online, SMS etc.)
Private GST Data: Private data consists of the details filed in the GST returns and requires taxpayer to give consent. Invoice information, vendors, customers, tax liability etc. are typically reported in GST Returns, which can be either on monthly or quarterly frequency as opted by the taxpayer. Read more about GST Data and its emerging use cases.
Public data itself is a good data set for initial screening for loan eligibility. Whereas, private data from GST Returns can be used for credit underwriting.
2. Taxpayer’s Consent for GST Data– ‘one time’ and ‘recurring’
Taxpayer’s consent is needed to access the private data. For API-based GST data fetch, there are two types of consent needed:
One-time consent by enabling API access. This consent means that taxpayer’s data can be fetched using APIs by any application but via a GSP and after a secured session is established via OTP. Unless API access is enabled, none of the applications are authorized to get the data not even GSPs. While enabling API access, taxpayers can also set the duration of the secured session. The session duration can be set as either 6 hours or 30 days. This duration will be applicable on an on-going basis. OTP based consent on need basis
Just enabling API access does not mean your data can be fetched by any system. To get data from GST system or send data, authenticity of taxpayer and taxpayers’ systems needs to be established. This helps to avoid misuse of data and keeps in check unauthorized access.
OTP is sent to the number registered on GST Portal. Once the authenticity between the taxpayer and the requesting application is established via OTP, any interaction with GST system can be performed for the duration as set for secured session. After the duration expiry, a new session needs to be established via OTP. Read more about Factors to consider while building the Credit models
3. Helps To keep a check on fake GST invoices and ITC claims
To keep a check on fake GST invoices and ITC claims, GST Returns and taxpayers’ activities are now monitored by the Government. Anomalies, if any, result in temporary suspension or cancellation the GSTIN. Taxpayers are also getting notices for discrepancies in data reported in GST returns, which again could result in change in their GSTIN status till the time the matter is not explained and resolved.
From public data, one can know the status of GSTIN, however the purpose or whether it was self-initiated or because of regulatory action, whether temporary etc. cannot be deciphered.
Also, the filing status can change within minutes of filing without a previous trail. Hence, the lending companies need to see that even if only active GSTINs are considered during evaluation, they still need to keep track of any change in status subsequently. The non-compliance may indicate potential risk of default or delayed payments.
4. Everything is at GSTIN level
Be it data or consent, any interaction with GST system is at GSTIN level. An important point to note is, under GST regime businesses had to opt for state-level registration. Thus, a business entity (say having a PAN) may have more than one GSTINs registered in different states, or even within the same state may have multiple GSTINs for different business verticals.
While it is possible to identify whether a PAN has more than one GST registrations obtained, it is fair to assume small businesses are most likely to have one GST registration.
Nonetheless, it is good to check out how many registrations have been taken by the business and importantly how compliant they are. Using public data itself, a holistic view of business can be observed as we do in our Peridot Compliance Report.
5. Keep an eye on GSTIN status
GST Data APIs are modular in nature. In order to make optimum API calls, it is essential to know beforehand what data to request. Availability of GST data is determined by factors such as:
- GST Registration type – The returns to be filed by taxpayers are determined by their nature of registration. 85% of the taxpayers registered as regular or normal type need to file GSTR 1 and GSTR3B. The next 13% who have opted for composition, currently file CMP 08. Remaining registration types are select type of operations such as ecommerce, ISD etc.
- Filing Frequency – To ease the reporting for small businesses, quarterly return and monthly payment schemes have been introduced from time to time. Before Dec 2020, it was only GSTR 1 which could be filed quarterly, while from Jan 2021, both GSTR 1 and 3B can be filed quarterly. The filing preference can change every quarter and is selected preference is available as public data.
- Returns filed – While the registration type and frequency will tell which returns are expected and at what periodicity, the data is available only if returns have been filed. The filing history provides details of Returns filed, which can be used to determine the data to be fetched.
It is useful to analyze the profiles of borrowers before deciding the GST data to be requested instead of having blanket rule for all borrower profiles.
6. Varying levels of GST Data for lending
While the GST Returns are standardized across all taxpayers, the data and its granularity levels differ across returns as well as within returns. If we look at the most common return filed by majority of the taxpayers, these would be:
- GSTR 1 – outward supplies by regular taxpayers
- GSTR 3B – statement of tax liability and settlement
- CMP 08 – quarterly statement for composition dealer
- GSTR 2A /GSTR 2B (and 4A) – auto-drafted statements created by GST system.
Across these returns, GSTR 1 and GSTR 2A/2B have invoice level details for B2B transactions and its corresponding debit and credit notes. Exports and large B2C transactions (in GSTR 1) and Import (GSTR 2A) are also available at invoicelevel, however there is no data field to identify the counterparty. Other sections of GSTR 1 like NIL, HSN summary, B2C are aggregated but again at various levels. GSTR 3B and CMP 08 contain only aggregated numbers.
While defining evaluation models, understanding of how much information will be available for different transaction types should be well understood.
7. Branch transfers are also reported as sales
As per GST rules, in cases where a business has obtained multiple GST registrations, any supply transaction between own GSTINs is also to be reported as sales. Inter-state branch transfer is one classic example.
As a result of this, the turnover mentioned in GSTR 3B will include intra-business transaction, even though it may not be actual sale. However such cases can be identified using GSTR 1.
This case may or may not arise in case of small businesses where the likeliness of single GSTIN is higher. Still it is good to analyze intra-business transactions and its impact on total turnover.
8. Upto date Filings may not always imply full liability discharged
In case of regular taxpayers, the tax liability (or sales transactions) is declared in GSTR 1 at a granular level, while in GSTR 3B where actual payment is made, only total values of sales are declared.
GSTR 1 and GSTR 3B are loosely linked. Taxpayers do have option to edit amounts. There could be genuine cases where edits could be needed. Nil GSTR 3B could be also filed, as it was observed in some fraud invoicing cases.
Whether the filings were NIL or if there was any discrepancy between GSTR 1 and 3B can be known only after private data is fetched. Hence, there are risks involved if evaluationis done purely based on public data without looking at the GST returns filed.
9. Inferring error responses
As seen above, GST data lending is based on many parameters. Once you decide which ‘data sets’ you want to fetch, next watch out for the responses, especially error responses. There will be certain cases where though an error response is received; it actually doesn’t mean there is an issue in getting data.
For example, GSTR 1 is made of many sections such as B2B invoices, B2C aggregate data, Export invoices, Credit and Debit notes etc. Data fetch happens section wise and if there are no export transactions, you may get a specific message. While, if GSTR 1 itself is not filed for the period, a different error is received.
It is important to classify the errors and focus on the ones which indicate issues requiring further action
10. Be prepared for updates in GST Returns and the APIs
There have been numerous changes in GST rules and reporting in GST Returns from the time GST was rolled out. From a data perspective, most of the core data fields have remained steady. There have been new additions and updates to the existing ones.
The changes and the future are not in anyone’s control, but provisioning for changes is something to be factored in. Also in certain cases, backward compatibility might be needed. For example, the HSN summary was updated from May 2021. If that’s the data set being used and trend is to be drawn, then data for earlier periods is also needed, where the data will be available in a slightly different way.
For GST data based lending , the source is the GST system and the APIs are released for applications to use. There are certain areas such as Challan and actual payment which are not released for GSPs/applications to consume. Likewise, certain information can be found on the GST portal and not in the public or private APIs. While we hope such differences will eventually be ironed out, overall the available GST data is still vast and valuable to be considered in digital lending programs.
These are some common and general aspects of getting started. As we go deep in understanding GST Returns and structures, the definitions of data, what is included in every category etc. needs to be understood well.
IRIS Credixo – Your GST Data Lending Partner
IRIS Credixo offers set of APIs covering publicly available GST data and consent based data filed in GST returns. Additionally pre-processed data sets, reports and insights are also available as enhanced data APIs which can quickly plug into any system. Banks, lending institutions and fintechs can integrate these APIs in their credit models and provide faster and seamless lending experience to the MSMEs.
IRIS Credixo APIs can embed into your existing credit appraisal or loan origination systems. Alternative, institutions can also build new applications and expand to new use cases based on the GST data available.
A full-blown web-application is already set-up for IRIS Credixo, which help data users i.e. the banks and fintechs to visualise the spread of GST data for lending.
Credixo stands at the intersection of Data and Lending Experience. It allows lenders to grow volumes, stand out from the competition and gather meaningful customer data to provide seamless “lending in the box” credit experience and design tailored customer journeys.