Connected Lending Checks: Flag Related-Party Loans Early
Last reviewed: November 2026
Connected lending checks are the controls a lender runs before sanction to find out whether a borrower, co-applicant, guarantor or linked entity is related to its own promoters, directors, key managerial personnel (KMP) or senior staff. Without them, related-party loans pass through normal delegation, interested people stay in the decision and the Board hears about it from the auditor.
This guide is for chief compliance officers and chief credit officers (CCOs) at non-banking financial companies (NBFCs), including housing finance companies (HFCs). It covers who counts as a related party under the Reserve Bank of India (RBI) rules, a seven-step pre-sanction workflow, eight match checks and the metrics that show the control works.
Key Takeaways
- From 1 April 2026, NBFC related-party lending follows Section A.3 of the RBI Credit Risk Management Directions, with per-transaction materiality thresholds of ₹1 crore, ₹5 crore and ₹10 crore by layer.
- "Related party" reaches well beyond directors' relatives, to entities where a related person is a partner, director, KMP, guarantor or holds over 10% of equity.
- Specified employees, everyone up to two levels below the Board, and their relatives need their own policy provisions.
- The hard part is data, not rules: name-only matching produces either noise or misses.
What Are Connected Lending Checks?
Connected lending checks compare every party to a loan application with the lender's related-party register before sanction. A confirmed match changes who may approve the loan, who must leave the room and what must be reported.
These checks feed the approval route but do not replace it. Our guide to the loan approval delegation matrix covers how sanctioning powers are set; this guide covers how a related-party match takes a loan out of normal delegation.
Who Counts as a Related Party Under RBI Rules?
Three groups count under the RBI rules: related persons, their relatives and related parties. According to the RBI (Non-Banking Financial Companies – Credit Risk Management) Directions, 2025 (RBI/DOR/2025-26/350, 28 November 2025, as amended), the definitions are wide.
| Group | Who Is Included | Where the Data Comes From |
|---|---|---|
| Related persons | Promoters, directors, KMP, holders of over 5% equity or voting rights, persons able to nominate a director or controlling the NBFC | Board records, shareholding register |
| Relatives | As defined in Section 2(77) of the Companies Act, 2013 | Declarations from each related person |
| Related parties | Entities where a related person is partner, manager, KMP, director or promoter, holds over 10% of equity, controls it, or is a guarantor | Director interest disclosures, company master data |
| Specified employees | Employees up to two levels below the Board, plus others the policy designates | Human resources (HR) system |
Why Do Connected Lending Checks Fail?
These checks fail for five reasons: a stale register, name-only matching, ignored guarantor roles, missed entity links and reciprocal arrangements. Each one has a prevention check.
| Cause | Effect | Prevention Check |
|---|---|---|
| Register updated once a year | New directors' relatives missed | Event-driven updates on appointments and exits |
| Matching on name only | Hundreds of false alerts, or misses on spelling | Permanent Account Number (PAN)-first, multi-attribute match |
| Only the main applicant checked | Related guarantor or co-applicant missed | Check every party on the application |
| Entity checked by name only | Company where a director's spouse holds 15% missed | Match shareholders, partners and directors |
| Reciprocal deals with another lender | Rules circumvented | Track loans to other lenders' insiders |
Why Is the Specified Employee List Harder Than It Looks?
The specified employee list is harder than it looks because "two levels below the Board" usually means dozens of people, not a handful. Each brings a spouse, parents, children, their spouses and siblings into scope.
For example, a Middle Layer NBFC with 60 specified employees can easily have 500 or more relatives to track. Staff loan schemes, home loans and loans to employees' family businesses all need to pass the same check.
What Is the Right Connected Lending Checks Workflow?
The right connected lending checks workflow has seven steps: build the register, keep it fresh, capture declarations, match every party, route matches, enforce recusal and report. The first two steps decide whether the rest work at all.
- Build a related-party register of persons, relatives and entities.
- Refresh it on every appointment, exit and disclosure.
- Capture a relationship declaration from each applicant and guarantor.
- Match every party on the application against the register.
- Route confirmed matches by amount against the materiality threshold.
- Recuse interested persons from the decision and later changes.
- Report to the Board, Audit Committee and internal audit.
Steps 1 and 2: How Do You Build and Refresh a Related-Party Register?
Build the register from four inputs: Board and KMP records, the shareholding register, director interest disclosures and the HR system for specified employees. Directors already disclose their interests in other entities under Section 184 of the Companies Act, 2013, so start there.
According to paragraph 13K of the Credit Risk Management Directions, NBFCs must maintain and periodically update lists of related persons, related parties and the loans sanctioned to them. Treat a director appointment or a specified employee promotion as a trigger, not something for the annual refresh.
Clean customer data makes matching possible. Our guide to customer master data cleanup in the LMS explains how to fix duplicate and incomplete records in the loan management system (LMS).
Steps 3 and 4: How Should Matching Work?
Matching should start with PAN, then use name plus father's name, date of birth, address, mobile number and bank account to score near-matches. The RBI rules do not require a borrower declaration, but asking "Are you related to any director or senior officer of this company?" costs nothing and creates evidence.
Check every party: applicant, co-applicants, guarantors and the promoters of entity borrowers. Our guides to loan application deduplication and guarantor and co-applicant verification cover the same matching logic used for other purposes.
Steps 5 and 6: How Should Related-Party Loans Be Approved?
Related-party loans above the materiality threshold must be sanctioned by the Board or a Board committee other than the Audit Committee. According to the RBI (NBFC – Credit Risk Management) Amendment Directions, 2026 (RBI/2025-26/179, 5 January 2026), the threshold cannot exceed these ceilings per transaction.
| NBFC Layer | Maximum Materiality Threshold |
|---|---|
| Base Layer | ₹1 crore |
| Middle Layer | ₹5 crore |
| Upper and Top Layer | ₹10 crore |
Loans below the threshold may be sanctioned under delegated authority. Directors, KMP and specified employees must recuse themselves from decisions on their own or their related parties' loans, including later settlements, write-offs, waivers and enforcement.
Company law adds a pricing floor for loans to directors. Under Section 185(3)(b) of the Companies Act, 2013, a company that lends in the ordinary course of business escapes the restriction on loans to directors only if it charges interest not below the prevailing yield of the Government security closest to the loan's tenor.
Step 7: What Must Be Reported?
Three reports are required: credit facilities to specified employees and their relatives go to the Board every year, internal auditors review related-party lending at least quarterly, and deviations from the policy go to the Audit Committee with reasons. Keep each match decision, approval and recusal in one audit trail.
Which Match Checks Should Run Before Sanction?
Eight match checks should run before sanction, each on every party to the application.
| Match Check | What It Catches |
|---|---|
| PAN against register | Direct related persons and relatives |
| Name plus father's name and date of birth | Relatives without PAN on file |
| Residential address | Family members at the same address |
| Mobile number and email | Shared contact details with staff |
| Bank account for disbursal or repayment | Funds routed to an insider's account |
| Directors and shareholders of entity borrowers | Related entities |
| Guarantor and co-applicant roles | Related persons standing behind a loan |
| Borrowers linked to other lenders' insiders | Reciprocal lending |
Would your sanction system catch a loan to your director's brother-in-law's company?
Talk to Our Lending Platform Specialists
Visit Digiqt to see how a custom lending platform can screen every application against your related-party register.
How Much Review Effort Do Weak Connected Lending Checks Create?
Weak matching creates a flood of false alerts, and that flood is how real matches get missed. In this illustrative scenario, a lender screens 20,000 applications a month with three parties each, or 60,000 party checks.
Assume name-only fuzzy matching flags 1% of parties, while PAN-first, multi-attribute matching flags 0.1%. Assume each alert takes 15 minutes to clear.
| Measure (Illustrative) | Name-Only Matching | Multi-Attribute Matching |
|---|---|---|
| Alerts a month | 600 | 60 |
| Review hours a month | 150 | 15 |
| Alerts per true match (if 3 real matches) | 200 | 20 |
The lesson for leadership: reviewers who clear 200 false alerts for every real one start clearing them by habit. Better matching data is a control, not a convenience.
Where Does the Digiqt Lending Platform Fit in Connected Lending Checks?
The Digiqt Lending Platform helps lenders initiate a custom lending product or platform in a matter of weeks. It handles everything from onboarding and disbursement to repayments through application programming interfaces (APIs), and its NBFC use case offers APIs for onboarding, payments and other financial products.
The platform page does not name a related-party module, so these controls would be custom builds on the platform:
- A related-party register with event-driven updates can be built alongside the onboarding APIs.
- Matching at onboarding can be built to screen every party before an application reaches credit.
- Threshold-based routing and recusal-aware approvals can be configured so matched loans never reach a conflicted approver.
Digiqt's delivery process runs from problem understanding, through a suggested solution with return on investment (ROI) validation, to prototyping, agile sprints, launch and ongoing optimisation. Run the matching rules on a year of your own past applications before switching them on.
Exposure to related parties also counts toward concentration. Digiqt's portfolio concentration monitoring AI agent shows how exposure across linked borrowers can be tracked.
Which Control Fits Each Related-Party Situation?
The right control depends on who the related person is, the amount and whether the loan is new.
| Situation | Recommended Approach |
|---|---|
| Loan to a director's relative below threshold | Delegated sanction, recusal, entry in related-party loan list |
| Any related-party loan above threshold | Board or non-Audit Board committee sanction |
| Staff loan to a specified employee | Policy provisions for specified employees, annual Board report |
| Entity where a director is guarantor | Treat entity as related party, full match on promoters |
| Existing non-compliant loan due for renewal | No renewal or enhancement; let it run off |
| Loan to another lender's director | Check for reciprocal arrangements |
Exceptions approved outside policy should be visible too. Our guide to credit policy deviation tracking explains how to record and review them.
Which Metrics Show Connected Lending Checks Are Working?
Five metrics show whether connected lending checks are working: register freshness, alert precision, routing compliance, recusal completeness and post-sanction findings. Review them each quarter with internal audit.
| Metric | Formula | What a Weak Number Signals |
|---|---|---|
| Register freshness | Days between a board or staff change and register update | Matches missed for weeks |
| Alert precision | Confirmed matches ÷ alerts raised | Matching data too thin |
| Routing compliance | Above-threshold related loans sanctioned by Board or committee ÷ all such loans | Delegation bypassed |
| Recusal completeness | Related-loan decisions with recorded recusal ÷ related-loan decisions | Weak audit trail |
| Post-sanction findings | Related loans found by audit, not at sanction ÷ all related loans | Pre-sanction check failing |
Inspectors ask for exactly this evidence. Our guide to RBI inspection readiness for NBFCs covers how to keep it ready.
Which Mistakes Weaken Connected Lending Controls?
Four mistakes weaken connected lending controls, and each is avoidable.
- Checking only the main applicant. Guarantors and co-applicants are where links hide.
- Using last year's register. One new director can bring dozens of new related persons.
- Applying another layer's threshold. Your layer decides the ceiling; see our guide to scale-based regulation for NBFCs.
- Renewing legacy related-party loans by default. Automatic renewal clauses do not override the run-off rule.
What Is Changing Next?
The renewal rule will bite as legacy facilities mature. According to RBI Amendment Directions RBI/2025-26/179 (5 January 2026), existing non-compliant related-party loans may run off to maturity but cannot be renewed, reviewed or enhanced unless they meet the new rules, even if the contract allows renewal.
Working capital lines and annual-review facilities will reach that point first. Flag every legacy related-party facility in the loan system now, so renewal workflows stop it before a relationship manager does not.
Want related-party matches caught at onboarding, not at audit?
Talk to Our Lending Platform Specialists
Visit Digiqt to see how the Digiqt Lending Platform can support related-party screening, routing and audit trails.



