FOIR Calculation Automation: Extracting Income from Multiple Document Types
How lending ops teams can automate the Fixed Obligation to Income Ratio calculation by extracting income and obligation data from the full loan document stack.
ParseAI Editorial Team
Document automation research and analysis
FOIR — Fixed Obligation to Income Ratio — is one of the core metrics in retail credit assessment. It measures what percentage of a borrower's net monthly income is already committed to existing obligations. Most lenders set a maximum FOIR threshold of 40 to 55% depending on income band and loan type. An applicant whose existing EMIs consume 60% of take-home pay is unlikely to be approved for a new obligation.
Calculating FOIR accurately requires two things: the correct net income figure, and a complete picture of existing obligations. Both come from documents. Getting both right, consistently, at scale, is a manual data entry challenge that many lending ops teams have not fully solved.
The Income Side: Where Errors Happen
Net income for FOIR is not the same as gross income. For a salaried applicant:
- Gross salary on the salary slip is not the right number — net take-home after all deductions is
- Net taxable income on the ITR is not the same as take-home — it is income after deductions but before tax
- Average salary credits in the bank statement may differ from the salary slip if the applicant has variable pay components
The correct income figure depends on the lender's FOIR policy — some use bank statement average credits, some use ITR net income, some use salary slip take-home. The policy is consistent; the error is in which figure is extracted from which document.
For self-employed applicants, income is calculated from ITR business income or banking turnover, with different policies across lenders. The document to field mapping is more complex, and errors are correspondingly more common.
The Obligation Side: What Gets Missed
Most applicants declare their existing EMIs on the loan application form. But obligations also appear in documents — and those document-sourced obligations are more reliable than self-declarations:
- Bank statement EMI debits — recurring fixed outflows matching EMI patterns (same amount, same date each month)
- Salary slip deductions — EMIs or loan deductions listed in the deductions section of the payslip
- CIBIL report obligations — active loan accounts and their EMI amounts
- Sanction letter references — if the applicant has provided a sanction letter for an existing loan
An applicant who declares no existing EMIs but has three consistent monthly debits of fixed amounts in their bank statement has undisclosed obligations. Automated extraction catches these; manual entry of self-declared figures does not.
FOIR errors are systematically asymmetric — they tend to understate obligations and overstate income. Both errors increase approval rates on applications that should be declined, raising default risk.
How Automated FOIR Calculation Works
Automated FOIR calculation extracts income and obligation data from every document in the loan file and assembles the calculation:
- Salary slips — net take-home extracted from each month, averaged
- ITR — net taxable income extracted per assessment year
- Bank statements — average monthly credits calculated, recurring debit patterns identified as potential EMIs
- Salary slip deductions — existing EMI deductions extracted
- Declared EMIs from application — carried forward
The system presents the underwriter with a FOIR summary: income source and figure, obligation list with source document for each, and the calculated ratio. The underwriter reviews the summary rather than re-reading all source documents.
Integration with Loan Origination Systems
FOIR calculation output integrates directly with the LOS via API. Extracted income and obligation figures push to the relevant LOS fields. The calculated FOIR appears on the application record alongside the source data, with document references for each figure. Underwriters can trace every number back to the source document without hunting through a physical file.
Related solution
See how ParseAI works for Lending & Banking
See ParseAI on your documents
Send us 10 documents from your workflow. We'll show you the extraction output in 24 hours.
Book a demo