A growing loan portfolio can significantly increase interest income, but lenders still need to watch the quality of the loans fueling that growth. TransUnion reported that the average amount financed on new auto loans reached $45,028 in Q1 2026, up 6.6% compared to the previous year. At the same time, the share of consumers who were 60 days or more past due rose to 1.57%.
For lenders focused on improving your loan portfolio, growth and credit performance should be assessed together. This means evaluating the quality of new loans, monitoring how existing loans perform over time, and identifying where risk is building up. The strategies below focus on underwriting, portfolio monitoring, concentration risk, and the technology connecting those activities.
1. Strengthen Underwriting With Better Data
Better underwriting relies on having the right information available before making a credit decision. The exact inputs will vary by loan product, but lenders can start with the following:
| Data Needed | Where to Source It | How It’s Used |
| Credit history | Credit bureaus | Assess payment history and credit risk |
| Income and employment | Payroll, bank data, employer verification | Verify ability to repay |
| Identity | Identity verification providers, government records | Confirm identity and detect discrepancies |
| Fraud signals | Fraud databases, device and identity tools | Flag suspicious applications |
| Collateral data | Valuation providers, vehicle data sources | Assess collateral value and loan structure |
| Alternative credit data | Cash-flow and other permitted data sources | Add context for thin-file applicants |
| Past loan performance | Internal origination and servicing systems | Refine credit rules using actual outcomes |
How the Strategy Works
The quality of a loan portfolio is partly set before a loan is booked. Adding verified information to the underwriting process can help lenders decide if an applicant meets their credit criteria and if the proposed terms match the risk involved.
The value comes from combining the information instead of just gathering more of it. Credit history may show how an applicant managed previous debt, while income and cash-flow data can provide more insight into current repayment ability. Fraud and identity checks answer a different question: whether the applicant and the supplied information are legitimate.
How to Implement It
Improving your portfolio begins with mapping the information needed at each step in the credit decision. Not every application needs the same level of verification, so lenders can set rules that decide when additional data or manual review is necessary.
Next, incorporate those rules into the loan origination workflow. Applications that meet defined criteria can proceed automatically, while exceptions can be sent to the right employee for review.
After funding, compare the original underwriting data against the overall loan performance. If loans with certain traits consistently lead to higher delinquencies or losses, lenders can investigate the cause and see if credit criteria, pricing, verification requirements, or decision rules should change.
Roadblocks to Plan For
The biggest issue may be data quality instead of data availability. Information can be outdated, inconsistent across sources, or tough to match to the correct applicant. Third-party data also brings costs, integration challenges, privacy issues, and compliance concerns.
Lenders should validate data sources before relying on them, create procedures for conflicting information, monitor automated decision rules for unexpected results, and maintain a clear process for exceptions and manual review.
2. Identify Portfolio Risk Earlier
Once loans are funded, servicing and performance data reveal where risk is changing. Lenders need both portfolio-level metrics and enough detail to see which loans or segments are causing the change.
| Data or Metric Needed | Where to Source It | What to Watch For |
| Delinquency rates | Loan servicing system | Increases by credit tier, dealer, product, geography, or vintage |
| Roll rates | Servicing and collections data | More accounts moving into later delinquency stages |
| Charge-offs and recoveries | Servicing and finance systems | Rising net losses or changes within specific segments |
| Vintage performance | Origination and servicing systems | Newer cohorts performing worse than comparable older cohorts |
| Payment behavior | Servicing system | Missed, partial, or increasingly late payments |
| Collateral values | Valuation providers and market data | Declining values that increase loss severity |
| Credit migration | Credit bureaus and account-review data | Borrowers moving into higher-risk credit tiers |
How the Strategy Works
Risk often develops unevenly across a portfolio. The overall delinquency rate can stay relatively stable while loans from a certain vintage, credit tier, dealer, or geography start to deteriorate.
Segmenting performance data helps lenders spot those changes earlier. For example, vintage analysis can reveal whether loans originated during one period are becoming delinquent more quickly than previous groups at the same point in their lifecycle. Roll rates can show whether delinquent borrowers are recovering or moving into more serious stages of delinquency.
The OCC’s 2026 Lending and Loan Portfolio Risk Management guidance similarly calls for ongoing monitoring to identify changes, trends, concerns, and emerging risks.
How to Implement It
Start by choosing the indicators that are most relevant to the portfolio and establishing a baseline for normal performance. Break those metrics down by the segments most likely to explain changes in results, such as credit tier, product, dealer, geography, term, and origination vintage.
Set thresholds for when changes require investigation and define the response in advance. If 30-day delinquencies for a recent vintage exceed the lender’s expected range, for example, the next step could be to review the affected loans for common characteristics and compare them with previous vintages.
The response should depend on what the analysis finds. A servicing issue may need earlier borrower outreach, while deterioration concentrated among newly originated loans may require a review of underwriting criteria or verification procedures.
Roadblocks to Plan For
The main challenge is distinguishing an early warning from normal variation. Setting thresholds too tightly can overwhelm teams with alerts, while thresholds that are too broad may let deterioration continue before anyone investigates it.
Delayed or fragmented data can create another issue. If servicing, collections, and origination information is updated on different schedules or stored separately, lenders may not notice the pattern soon enough to act on it.
Start with a manageable set of indicators, establish baselines using the lender’s own historical performance, and assign responsibility for each trigger. As the lender learns which indicators reliably predict deterioration, thresholds can be refined and less useful alerts removed.
3. Manage Concentration Risk With Exposure Limits and Stress Testing
Concentration risk can increase when too many areas of the portfolio share the same source of risk. Lenders need sufficient detail to pinpoint these concentrations and predict what might occur if a heavily represented segment begins to underperform.
| Data or Metric Needed | Where to Source It | What to Watch For |
| Product exposure | Origination and servicing systems | Growing share of balances in one loan product |
| Credit-tier exposure | Credit and origination data | Large share of balances within one risk tier |
| Dealer exposure | Dealer and origination records | Dependence on a small number of dealers |
| Geographic exposure | Borrower and collateral records | Loans concentrated in the same market or region |
| Collateral exposure | Loan and valuation data | Dependence on similar collateral types or values |
| Term and LTV mix | Origination system | Growing exposure to longer terms or higher LTVs |
| Losses by segment | Servicing and collections systems | Concentrated segments producing disproportionate losses |
How the Strategy Works
Concentration analysis looks for groups of loans that could respond similarly to the same conditions. A portfolio may include thousands of individual loans but still carry significant concentration risk if a large percentage share the same geography, credit tier, dealer, collateral characteristics, or loan structure.
Exposure limits help lenders determine how much of a particular risk they are willing to hold. Stress testing takes this further by estimating potential losses if conditions worsen. For instance, a lender with a large share of high-LTV auto loans could test how that segment performs under assumptions of increased defaults and declining vehicle values.
The results can indicate which concentrations are manageable and which could lead to losses exceeding the lender’s risk tolerance.
How to Implement It
Begin by identifying the characteristics most likely to cause correlated losses in the portfolio. Calculate each material exposure as a percentage of outstanding balances, originations, or another relevant measure of risk.
Next, set internal limits or review thresholds based on the lender’s risk appetite. Monitor these as new loans are originated, rather than reviewing them after a concentration has developed.
Then stress-test the largest or most significant exposures. Scenarios should represent risks likely to affect the segment being tested. If projected losses exceed acceptable levels, lenders can adjust future origination volumes, credit criteria, pricing, or other portfolio parameters to reduce further exposure.
Roadblocks to Plan For
Concentration does not automatically indicate a problem. An auto lender will naturally have significant exposure to auto loans, while a regional lender may have notable geographic concentration. Applying generic diversification targets could conflict with the institution’s business model without significantly reducing risk.
A more effective approach is to set limits based on the lender’s capital, historical losses, market conditions, and risk appetite. Stress tests also need realistic assumptions. Scenarios that are too mild may understate potential losses, while extreme scenarios with little relation to the portfolio may offer limited insight for decision-making.
4. Use Loan Performance Data to Improve Future Decisions
Loans currently held within a lending institution can help lenders improve future decisioning. To achieve this, lenders need to connect the information gathered during origination with later servicing and performance data.
| Data or Metric Needed | Where to Source It | What to Watch For |
| Original credit profile | LOS and credit bureau data | Credit characteristics associated with stronger or weaker performance |
| Loan structure and pricing | LOS and loan records | Terms, rates, and LTVs associated with higher delinquencies or losses |
| Dealer performance | LOS and dealer records | Differences in loan performance by dealer |
| Payment performance | Loan servicing system | Late, partial, and missed payments |
| Delinquencies and losses | Servicing and collections systems | Higher-than-expected losses by origination characteristic |
| Vintage performance | LOS and servicing system | Differences between origination cohorts |
| Exceptions and overrides | LOS and decisioning records | Whether exceptions perform differently from standard approvals |
How the Strategy Works
Underwriting relies on what the lender knows about a borrower and loan during application. Performance data fills in the gaps: it reveals what happened after the loan was approved.
Connecting the two allows lenders to test if their initial assumptions were right. For instance, a lender might find that loans in a specific credit tier tend to perform as expected, but a certain combination of term, LTV, and dealer source leads to higher losses. This information can guide future credit policy reviews.
Over time, this creates a feedback loop. Each new group of loans produces performance data that helps the lender improve how it evaluates and prices future originations.
How to Implement It
Begin by identifying which origination fields should follow a loan throughout its lifecycle. These could include:
- Credit tier
- Score
- Income
- DTI, LTV
- Loan Term
- Pricing
- Dealer
- Decision result
- Any underwriting exceptions.
Link those fields with servicing outcomes such as payment history, delinquency, charge-offs, recoveries, and payoff. Lenders can then compare outcomes across vintages and segments to spot patterns.
When a pattern emerges, investigate the cause before changing policy. For example, if loans approved through a specific exception show consistently higher losses, review the exception criteria and how employees apply them. The solution might be stricter rules, different pricing, additional verification, or no change if the added risk is already reflected in the return.
Make this analysis a regular part of credit-policy reviews, rather than waiting for portfolio performance to decline.
Roadblocks to Plan For
The main challenge is often aligning origination and servicing data. Systems may have different field names, definitions, identifiers, or reporting periods. Historical records may also lack completeness, making comparisons difficult.
Start with a defined set of fields and outcomes instead of trying to link every available data point at once. Standardize definitions, create a common loan identifier, and document how each metric is calculated. Automated reporting can lessen manual effort once the underlying data is consistent.
Lenders should also be cautious about drawing conclusions from small samples or short performance periods. A new vintage needs sufficient time and loan volume to yield meaningful results. Changes to credit policy should come from data-backed patterns and not isolated performance shifts.
Improving Your Loan Portfolio With defi SOLUTIONS
Improving your loan portfolio requires lenders to keep learning from the loans they originate. Underwriting determines which risks enter the portfolio, while servicing and performance data show whether those risks played out as expected. Feeding that information back into credit policy can help lenders make better decisions about future applications.
defi SOLUTIONS connects loan origination and servicing with configurable lending technology designed for the entire loan lifecycle. Lenders can automate workflows, apply credit policies consistently, manage servicing activity, and use portfolio data to guide future decisions. Contact the defi team to learn more.
