Presentation on theme: "Equifax Confidential and Proprietary Understanding Credit Scores Customer Tutorial."— Presentation transcript:
Equifax Confidential and Proprietary Understanding Credit Scores Customer Tutorial
Equifax Confidential and Proprietary 2 Important Legal Note The information in this presentation is not to be relied upon, is not intended to be, nor should it be used or construed as, legal advice. Equifax assumes no liability for any errors or omissions in the information in this presentation. Compliance with the Fair Credit Reporting Act (FCRA), the Equal Credit Opportunity Act (ECOA) or their respective regulations is the responsibility of each entity to which such laws apply. All specific consumer, customer and other third-party information in this presentation is fictitious. This presentation contains Equifax proprietary and confidential information. Do not distribute or copy.
Equifax Confidential and Proprietary 3 What is Credit Scoring? The application of statistical methods to credit data with the intent of predicting the likelihood of some credit-related event taking place. Makes use of credit history information Developed using analytically derived, demonstrably and statistically sound statistical techniques A credit score does not tell how an individual will act. Rather, it tells the probability or likelihood that the individual will act a certain way.
Equifax Confidential and Proprietary 4 Developing a Scoring Model Commonly used terms An Attribute (aka Characteristic or Variable) is an aspect of an individuals credit history. Some examples might be Age of Oldest Trade or Utilization Rate on Open Bankcard Trades. A Bad Definition is what the model is developed to predict. A common bad definition for the Financial industry is 90+ Days Past Due; common bad definitions for the Telecommunication industry are No Pay, Involuntary Disconnect or 60+ Days Past Due.
Equifax Confidential and Proprietary 5 Developing a Scoring Model (continued) Commonly used terms A Performance Period is the time period for which the bad definition applies. Financial industry scorecards are designed to predict the likelihood of some event occurring over the next months. Telecommunication industry scorecards have a performance window of 6-12 months. The Observation Point is the point from which the model development data was taken. It is the starting point of the Performance Period. A Bad Rate is the percentage of accounts that meet the bad definition within a certain score range. Typically, bad rates are quoted as interval or cumulative bad rates.
Equifax Confidential and Proprietary 6 Model Development Utilizes historical information to predict future events and outcomes Predictive Model Observation Point 90+ Days Past Due No Pay Prediction Time Frame Performance Window Independent Variables Dependent Variable Credit Attributes Historical Information
Equifax Confidential and Proprietary 7 Developing a Scoring Model General scoring model factors Payment History Has there been delinquency in the recent or historical past? Amount Owed What are the aggregate balances? How high is the credit utilization (balances as a percent of available credit)? Length of Credit History This is a proxy for stability – longer history equates to stability and often more credit information.
Equifax Confidential and Proprietary 8 Developing a Scoring Model (continued) General scoring model factors New Credit Has the consumer escalated their use of credit? Types of Credit in Use Does the consumer have a healthy mix of credit devices? Public Records Publicly available information related to bankruptcies, judgments and liens.
Equifax Confidential and Proprietary 9 Developing a Scoring Model (continued) An example scorecard (for illustrative purposes only)
Equifax Confidential and Proprietary 10 What Factors Affect a Score? Payment History A record of late payments on current and past credit accounts may lower the score. Public Records Matters of public record such as bankruptcies, judgments, and lien items may lower the score. Amount Owed Owing too much may lower the score, especially if the accounts are approaching the total credit limit.
Equifax Confidential and Proprietary 11 What Factors Affect a Score? (continued) Length of Credit History In general, a credit history that dates back for a longer period of time is better. New Accounts Opening multiple new accounts in a short period of time may lower the score.
Equifax Confidential and Proprietary 12 What Factors Affect a Score? (continued) Inquiries Whenever someone else, i.e. a lender, gets a credit report an inquiry is recorded on that credit report. A large number of recent inquiries may lower the score. Open Accounts The presence of too many open accounts can lower the score, regardless of whether the accounts are being used or not. However, closing accounts will likely cause the utilization rate to go up which may lower the score.
Equifax Confidential and Proprietary 13 Performance Charts A gains chart is a tool used to obtain an overall view of how a particular score performs on a specific population. Generally speaking, the population is rank ordered from the lowest risk (top) to the highest risk accounts (bottom), and is then separated into equal-sized groups. Gains charts can be used in a variety of ways and here is an example of information that a risk model gains chart can provide: Isolating high-risk accounts in the low level score ranges. When using risk models, this information is useful to clients who are looking to reduce their overall portfolio delinquency rate. By referencing the Decum % of bads column, the client is able to determine which score ranges capture the most amount of bad accounts in their portfolio. (The higher the number, the higher percentage of bads that are isolated at or below a particular score range.) This information can then be used to drive their acquisition strategy by allowing the client to determine which customers they wish to add to their existing customer base. Accounts Eliminated Cumulative Population Cut off score
Equifax Confidential and Proprietary 14 Gains Chart Explanation Percent of Goods: A calculation that divides the summation of good accounts by the total good population. For example, 11.86% = 23,775 / 200,533. Decum % of Bads: A calculation that divides the summation of bad accounts by the total bad population. For example, 43.58% = 21,239 / 48,741. Note: Calculations for this number starts from the bottom score range and filters to the top. Interval Bad Rate: A calculation that divides the number of bads by the population within that interval. For example, in the upper most decile the interval bad rate of 1.16 = 280 / 24,055. Cumulative Bad Rate: A calculation that divides the summation of bad accounts by the summation of total of accounts. For example, 2.36 = ( ) / (24, ,746). Min / Max Score: The minimum and maximum score ranges within that specific percentile. Total Accounts: The number of accounts present in the portfolio. Goods: The population of accounts the customer is targeting to keep (e.g., paid accounts). Bads: The population of accounts the customer is targeting to eliminate (e.g., non-payers or 90+DPD).
Equifax Confidential and Proprietary 15 Dual Score Matrix- Risk Strategy A dual score matrix offers further risk segmentation. Risk based product offers can be set within each one of the risk segments based on combinations of the General Risk Score and Bankruptcy Navigator Index 3.0 scores value. High Risk Medium Risk Low Risk