Credit Risk Analysis Using Credit Scoring Models

There is an element of financial risk to the landlord’s business in the installation of a new tenant. Tenant screenings provide risk management tools to assess credit risk by estimating the probability of tenant default on rent obligations. With credit risk analysis, the landlord can make an informed decision acceptable to the landlord’s business model. With adequate due diligence using screenings for a credit report and credit score, a landlord can reduce potential financial business risk by objectively analyzing the applicant’s credit history and credit management.

There is good reason for the landlord to protect his business from credit risk. Credit risk occurs when a tenant fails to meet his debt obligations. Rent defaults disrupt the landlord’s cash flow and can increase operational costs if the debt is sent for collection.

Credit risk analysis uses credit scoring models as marketed by the three national credit reporting companies, Equifax, Experian, and Transunion. The credit reporting companies analyze information in the consumer’s credit file as of the date of the credit score request. The analysis estimates the probability of a credit default triggered by an event such as consumer failure to pay as agreed. The probability of default is expressed as a three-digit number that represents creditworthiness. A lower score indicates a more likely credit risk while a higher score indicates a lower risk of credit default. A landlord may utilize credit scores to help evaluate an applicant’s ability to meet rent obligations in compliance with lease terms and conditions.

Credit Risk Scoring Models

How a credit score is calculated is dependent upon the credit scoring model used and the applicable version of the scoring model. There are many different credit scoring models targeting various credit populations, industry applications, and specific credit products.

Each credit reporting company generates its own scores by running the consumer’s credit data through a proprietary modeling process. Credit risk analysis models are predictive tools that can be based on either financial statement analysis, default probability, or machine learning.

As a basic discussion of a conventional scoring model, model development takes into consideration a sample population of consumer accounts large enough to make the model statistically valid yet characteristic of the population to which the model’s scorecard will be applied. Accounts in the selected population would be statistically analyzed to identify the characteristics and attributes that relate to creditworthiness. These characteristics would then be further refined into a smaller group of predictive variables which could best indicate how a credit applicant could be categorized as a credit risk. Predictive variables could include, but would not be limited to, prior credit performance, current level of debt, amount of time that credit has been in use, and new credit. Any characteristic or attribute that is prohibited by law for credit decisioning or that lacks predictive value would be excluded from scoring.

The scoring model would summarize the relevant available consumer credit data into a set of ordered categories that predict an outcome. This ordered set, a numerical score, is a snapshot of estimated credit risk at a specific date in time. The credit score is a statistical assessment of the consumer’s risk within the context of the total risk for the credit population being scored.

However there are limitations in the effectiveness of any credit scoring model. Models are developed, calibrated, and validated using lengthy historical data. If relevant un-modeled conditions change, models may not correctly predict credit behaviors out of sample. And, while a model forecasts the probability of credit default, it is not necessarily a predictor of the level of risk, i.e., the magnitude of loss. For tenant screening purposes landlords, as a matter of business policy for risk reduction, still need to assess the amount of risk posed by the rental applicant in relation to the amount of risk the landlord is willing to absorb.

The sample population size, changing economic conditions, global and domestic business environments, and the reactive nature of consumer credit behaviors can also be constraining factors influencing model effectiveness.

This has led developers to refresh credit scoring models periodically to reflect changes in the industry, consumer behavioral data, and product trends to provide relevant data for credit modeling.

As part of this relevancy, credit scoring models have been optimized to align with the National Consumer Assistance Plan (NCAP) to make credit reports more accurate, transparent, and easier for the consumer, including consumer ability to correct errors on the consumer’s report. As examples, tax liens and public records data reporting were historically used for conventional scoring. Now, tax liens and public records data have been removed from consumer credit files if they failed to meet the enhanced data quality standards as set out in the NCAP guidelines.

While there are many credit risk scoring models and many versions of the scoring models, the scoring models most commonly used to determine credit scores are FICO and VantageScore Solutions.

The current versions of FICO and VantageScore models incorporate trended credit data in their modeling process. Trended data is credit data that reflects patterns in a consumer’s behavior, as example, how the consumer borrows credit and repays credit over time. This is different than conventional credit scoring models that captured static credit events, e.g., looking to the most recently reported utilization rate to calculate the credit score.

FICO and VantageScore use a point credit scoring scale from 300 to 850. The scoring scale has five credit score ranges which determine the likelihood of creditworthiness and what credit products could be made available to the consumer, such as loans and interest rates.

The ranges differ between the two models, and also have different descriptive labels associated with each range. For example, a credit score range above 780 may be labelled exceptional by the FICO model and excellent by the VantageScore model. Other credit scoring models such as those of Experian and Equifax have their own proprietary scoring models and likewise credit scoring scales and descriptive credit range labeling.

The most common credit behaviors that influence credit scores are:

  • Payment history
  • Credit utilization
  • Length of credit history
  • Type of credit
  • New credit

A FICO scoring model may assign score factors such as:

  • 35% Payment History
  • 30% Credit Utilization
  • 15% Credit History
  • 10% Credit Use
  • 10% New Credit

A VantageScore scoring model may assign score factors such as :

  • 40% Payment history
  • 21% Age and type of credit
  • 20% Percentage of credit limit used
  • 11% Total balances
  • 5% Recent credit behavior
  • 3% Available credit

Differences in scoring models will affect only a credit score, not a credit report. Information on a credit report is based on data contained in the consumer’s credit file at the consumer reporting company as of a specific date. The credit report is not a calculated report derived from a credit scoring model. The credit report contains only information about the consumer’s credit usage such as:

  • A summary of the applicant’s positive and negative credit accounts
  • Payment history
  • Prior credit inquiries by date
  • Total estimated past due and monthly debts
  • Listing of credit accounts

A credit score is not an absolute statement of a consumer’s credit risk, financial strength, or stability. However credit reports and credit scores when used in conjunction with other tenant screenings and risk reduction policies may help a landlord to become more confident in tenant selection decisions.

Landlords are advised to conduct their own research and due diligence regarding credit risk analysis and credit scoring models. Those landlords utilizing third party tenant screening services may wish to consult with their screening partner to determine how the screening process is conducted, including the source of credit reporting, the type of credit report, and the availability of credit scoring.

Comments are closed.