Tag: models

In this article...What is fair lending?Understanding machine learning modelsThe pitfalls: bias and fairness in ML modelsFairness metricsRegulatory frameworks and complianceHow Experian® can help As the financial sector continues to embrace technological innovations, machine learning models are becoming indispensable tools for credit decisioning. These models offer enhanced efficiency and predictive power, but they also introduce new challenges. These challenges particularly concern fairness and bias, as complex machine learning models can be difficult to explain. Understanding how to ensure fair lending practices while leveraging machine learning models is crucial for organizations committed to ethical and compliant operations. What is fair lending? Fair lending is a cornerstone of ethical financial practices, prohibiting discrimination based on race, color, national origin, religion, sex, familial status, age, disability, or public assistance status during the lending process. This principle is enshrined in regulations such as the Equal Credit Opportunity Act (ECOA) and the Fair Housing Act (FHA). Overall, fair lending is essential for promoting economic opportunity, preventing discrimination, and fostering financial inclusion. Key components of fair lending include: Equal treatment: Lenders must treat all applicants fairly and consistently throughout the lending process, regardless of their personal characteristics. This means evaluating applicants based on their creditworthiness and financial qualifications rather than discriminatory factors. Non-discrimination: Lenders are prohibited from discriminating against individuals or businesses on the basis of race, color, religion, national origin, sex, marital status, age, or other protected characteristics. Discriminatory practices include redlining (denying credit to applicants based on their location) and steering (channeling applicants into less favorable loan products based on discriminatory factors). Fair credit practices: Lenders must adhere to fair and transparent credit practices, such as providing clear information about loan terms and conditions, offering reasonable interest rates, and ensuring that borrowers have the ability to repay their loans. Compliance: Financial institutions are required to comply with fair lending laws and regulations, which are enforced by government agencies such as the Consumer Financial Protection Bureau (CFPB) in the United States. Compliance efforts include conducting fair lending risk assessments, monitoring lending practices for potential discrimination, and implementing policies and procedures to prevent unfair treatment. Model governance: Financial institutions should establish robust governance frameworks to oversee the development, implementation and monitoring of lending models and algorithms. This includes ensuring that models are fair, transparent, and free from biases that could lead to discriminatory outcomes. Data integrity and privacy: Lenders must ensure the accuracy, completeness, and integrity of the data used in lending decisions, including traditional credit and alternative credit data. They should also uphold borrowers’ privacy rights and adhere to data protection regulations when collecting, storing, and using personal information. Understanding machine learning models and their application in lending Machine learning in lending has revolutionized how financial institutions assess creditworthiness and manage risk. By analyzing vast amounts of data, machine learning models can identify patterns and trends that traditional methods might overlook, thereby enabling more accurate and efficient lending decisions. However, with these advancements come new challenges, particularly in the realms of model risk management and financial regulatory compliance. The complexity of machine learning models requires rigorous evaluation to ensure fair lending. Let’s explore why. The pitfalls: bias and fairness in machine learning lending models Despite their advantages, machine learning models can inadvertently introduce or perpetuate biases, especially when trained on historical data that reflects past prejudices. One of the primary concerns with machine learning models is their potential lack of transparency, often referred to as the "black box" problem. Model explainability aims to address this by providing clear and understandable explanations of how models make decisions. This transparency is crucial for building trust with consumers and regulators and for ensuring that lending practices are fair and non-discriminatory. Fairness metrics Key metrics used to evaluate fairness in models can include standardized mean difference (SMD), information value (IV), and disparate impact (DI). Each of these metrics offers insights into potential biases but also has limitations. Standardized mean difference (SMD). SMD quantifies the difference between two groups' score averages, divided by the pooled standard deviation. However, this metric may not fully capture the nuances of fairness when used in isolation. Information value (IV). IV compares distributions between control and protected groups across score bins. While useful, IV can sometimes mask deeper biases present in the data. Disparate impact (DI). DI, or the adverse impact ratio (AIR), measures the ratio of approval rates between protected and control classes. Although DI is widely used, it can oversimplify the complex interplay of factors influencing credit decisions. Regulatory frameworks and compliance in fair lending Ensuring compliance with fair lending regulations involves more than just implementing fairness metrics. It requires a comprehensive end-to-end approach, including regular audits, transparent reporting, and continuous monitoring and governance of machine learning models. Financial institutions must be vigilant in aligning their practices with regulatory standards to avoid legal repercussions and maintain ethical standards. Read more: Journey of a machine learning model How Experian® can help By remaining committed to regulatory compliance and fair lending practices, organizations can balance technological advancements with ethical responsibility. Partnering with Experian gives organizations a unique advantage in the rapidly evolving landscape of AI and machine learning in lending. As an industry leader, Experian offers state-of-the-art analytics and machine learning solutions that are designed to drive efficiency and accuracy in lending decisions while ensuring compliance with regulatory standards. Our expertise in model risk management and machine learning model governance empowers lenders to deploy robust and transparent models, mitigating potential biases and aligning with fair lending practices. When it comes to machine learning model explainability, Experian’s clear and proven methodology assesses the relative contribution and level of influence of each variable to the overall score — enabling organizations to demonstrate transparency and fair treatment to auditors, regulators, and customers. Interested in learning more about ensuring fair lending practices in your machine learning models?    Learn More This article includes content created by an AI language model and is intended to provide general information.

This article was updated on February 28, 2024. There's always a risk that a borrower will miss or completely stop making payments. And when lending is your business, quantifying that credit risk is imperative. However, your credit risk analysts need the right tools and resources to perform at the highest level — which is why understanding the latest developments in credit risk analytics and finding the right partner are important. What is credit risk analytics? Credit risk analytics help turn historical and forecast data into actionable analytical insights, enabling financial institutions to assess risk and make lending and account management decisions. One way organizations do this is by incorporating credit risk modeling into their decisions. Credit risk modeling Financial institutions can use credit risk modeling tools in different ways. They might use one credit risk model, also called a scorecard, to assess credit risk (the likelihood that you won't be repaid) at the time of application. Its output helps you determine whether to approve or deny an application and set the terms of approved accounts. Later in the customer lifecycle, a behavior scorecard might help you understand the risk in your portfolio, adjust credit lines and identify up- or cross-selling opportunities. Risk modeling can also go beyond individual account management to help drive high-level portfolio and strategic decisions. However, managing risk models is an ongoing task. As market conditions and business goals change, monitoring, testing and recalibrating your models is important for accurately assessing credit risk. Credit scoring models Application credit scoring models are one of the most popular applications for credit risk modeling. Designed to predict the probability of default (PD) when making lending decisions, conventional credit risk scoring models focus on the likelihood that a borrower will become 90 days past due (DPD) on a credit obligation in the following 24 months. These risk scores are traditionally logistic regression models built on historical credit bureau data. They often have a 300 to 850 scoring range, and they rank-order consumers so people with higher scores are less likely to go 90 DPD than those with lower scores. However, credit risk models can have different score ranges and be developed to predict different outcomes over varying horizons, such as 60 DPD in the next 12 months. In addition to the conventional credit risk scores, organizations can use in-house and custom credit risk models that incorporate additional data points to better predict PD for their target market. However, they need to have the resources to manage the entire development and deployment or find an experienced partner who can help. The latest trends in credit risk scoring Organizations have used statistical and mathematical tools to measure risk and predict outcomes for decades. But the future of credit underwriting is playing out as big data meets advanced data analytics and increased computing power. Some of the recent trends that we see are: Machine learning credit risk models: Machine learning (ML) is a type of artificial intelligence (AI) that's proven to be especially helpful in evaluating credit risk. ML models can outperform traditional models by 10 to 15 percent.1 Experian survey data from September 2021 found that about 80 percent of businesses are confident in AI and cloud-based credit risk decisioning, and 70 percent frequently discuss using advanced analytics and AI for determining credit risk and collection efforts.2 Expanding data sources: The ML models' performance lift is due, in part, to their ability to incorporate internal and alternative credit data* (or expanded FCRA-regulated data), such as credit data from alternative financial services, rental payments and Buy Now Pay Later loans. Cognitively countering bias: Lenders have a regulatory and moral imperative to remove biases from their lending decisions. They need to beware of how biased training data could influence their credit risk models (ML or otherwise) and monitor the outcomes for unintentionally discriminatory results. This is also why lenders need to be certain that their ML-driven models are fully explainable — there are no black boxes. A focus on agility: The pandemic highlighted the need to have credit risk models and systems that you can quickly adjust to account for unexpected world events and changes in consumer behavior. Real-time analytical insights can increase accuracy during these transitory periods. Financial institutions that can efficiently incorporate the latest developments in credit risk analytics have a lot to gain. For instance, a digital-first lending platform coupled with ML models allows lenders to increasingly automate loan underwriting, which can help them manage rising loan volumes, improve customer satisfaction and free up resources for other growth opportunities. READ: The getting AI-driven decisioning right in financial services white paper to learn more about the current AI decisioning landscape. Why does getting credit risk right matter? Getting credit risk right is at the heart of what lenders do and accurately predicting the likelihood that a borrower won't repay a loan is the starting point. From there, you can look for ways to more accurately score a wider population of consumers, and focus on how to automate and efficiently scale your system. Credit risk analysis also goes beyond simply using the output from a scoring model. Organizations must make lending decisions within the constraints of their internal resources, goals and policies, as well as the external regulatory requirements and market conditions. Analytics and modeling are essential tools, but as credit analysts will tell you, there's also an art to the practice. CASE STUDY: Atlas Credit, a small-dollar lender, worked with Experian's analytics experts to create a custom explainable ML-powered model using various data sources. After reworking the prequalification and credit decisioning processes and optimizing their score cutoffs and business rules, the company can now make instant decisions. It also doubled its approval rate while reducing risk by 15 to 20 percent. How Experian helps clients With decades of experience in credit risk analytics and data management, Experian offers a variety of products and services for financial services firms. Ascend Intelligence Services™ is an award-winning, end-to-end suite of analytics solutions. At a high level, the offering set can rapidly develop new credit risk models, seamlessly deploy them into production and optimize decisioning strategies. It also has the capability to continuously monitor and retrain models to improve performance over time. For organizations that have the experience and resources to develop new credit risk models on their own, Experian can give you access to data and expertise to help guide and improve the process. But there are also off-the-shelf options for organizations that want to quickly benefit from the latest developments in credit risk modeling. Learn more 1Experian (2020). Machine Learning Decisions in Milliseconds 2Experian (2021). Global Insights Report September/October 2021

Model explainability has become a hot topic as lenders look for ways to use artificial intelligence (AI) to improve their decision-making. Within credit decisioning, machine learning (ML) models can often outperform traditional models at predicting credit risk.  ML models can also be helpful throughout the customer lifecycle, from marketing and fraud detection to collections optimization. However, without explainability, using ML models may result in unethical and illegal business practices.  What is model explainability?  Broadly defined, model explainability is the ability to understand and explain a model's outputs at either a high level (global explainability) or for a specific output (local explainability).1  Local vs global explanation: Global explanations attempt to explain the main factors that determine a model's outputs, such as what causes a credit score to rise or fall. Local explanations attempt to explain specific outputs, such as what leads to a consumer's credit score being 688. But it's not an either-or decision — you may need to explain both.  Model explainability can also have varying definitions depending on who asks you to explain a model and how detailed of a definition they require. For example, a model developer may require a different explanation than a regulator.  Model explainability vs interpretability  Some people use model explainability and interpretability interchangeably. But when the two terms are distinguished, model interpretability may refer to how easily a person can understand and explain a model's decisions.2 We might call a model interpretable if a person can clearly understand:  The features or inputs that the model uses to make a decision.  The relative importance of the features in determining the outputs.  What conditions can lead to specific outputs.  Both explainability and interpretability are important, especially for credit risk models used in credit underwriting. However, we will use model explainability as an overarching term that encompasses an explanation of a model's outputs and interpretability of its internal workings below.  ML models highlight the need for explainability in finance  Lenders have used credit risk models for decades. Many of these models have a clear set of rules and limited inputs, and they might be described as self-explanatory. These include traditional linear and logistic regression models, scorecards and small decision trees.3  AI analytics solutions, such as ML-powered credit models, have been shown to better predict credit risk. And most financial institutions are increasing their budgets for advanced analytics solutions and see their implementation as a top priority.4  However, ML models can be more complex than traditional models and they introduce the potential of a “black box." In short, even if someone knows what goes into and comes out of the model, it's difficult to explain what's happening without an in-depth analysis.  Lenders now have to navigate a necessary trade-off. ML-powered models may be more predictive, but regulatory requirements and fair lending goals require lenders to use explainable models.  READ MORE: Explainability: ML and AI in credit decisioning  Why is model explainability required?  Model explainability is necessary for several reasons:  To comply with regulatory requirements: Decisions made using ML models need to comply with lending and credit-related, including the Fair Credit Reporting Act (FCRA) and Equal Credit Opportunity Act (ECOA). Lenders may also need to ensure their ML-driven models comply with newer AI-focused regulations, such as the AI Bill of Rights in the U.S. and the E.U. AI Act.  To improve long-term credit risk management: Model developers and risk managers may want to understand why decisions are being made to audit, manage and recalibrate models.  To avoid bias: Model explainability is important for ensuring that lenders aren't discriminating against groups of consumers.  To build trust: Lenders also want to be able to explain to consumers why a decision was made, which is only possible if they understand how the model comes to its conclusions.  There's a real potential for growth if you can create and deploy explainable ML models. In addition to offering a more predictive output, ML models can incorporate alternative credit data* (also known as expanded FCRA-regulated data) and score more consumers than traditional risk models. As a result, the explainable ML models could increase financial inclusion and allow you to expand your lending universe.  READ MORE: Raising the AI Bar  How can you implement ML model explainability?  Navigating the trade-off and worries about explainability can keep financial institutions from deploying ML models. As of early 2023, only 14 percent of banks and 19 percent of credit unions have deployed ML models. Over a third (35 percent) list explainability of machine learning models as one of the main barriers to adopting ML.5  Although a cautious approach is understandable and advisable, there are various ways to tackle the explainability problem. One major differentiator is whether you build explainability into the model or try to explain it post hoc—after it's trained.  Using post hoc explainability  Complex ML models are, by their nature, not self-explanatory. However, several post hoc explainability techniques are model agnostic (they don't depend on the model being analyzed) and they don't require model developers to add specific constraints during training.  Shapley Additive Explanations (SHAP) is one used approach. It can help you understand the average marginal contribution features to an output. For instance, how much each feature (input) affected the resulting credit score.  The analysis can be time-consuming and expensive, but it works with black box models even if you only know the inputs and outputs. You can also use the Shapley values for local explanations, and then extrapolate the results for a global explanation.  Other post hoc approaches also might help shine a light into a black box model, including partial dependence plots and local interpretable model-agnostic explanations (LIME).  READ MORE: Getting AI-driven decisioning right in financial services  Build explainability into model development  Post hoc explainability techniques have limitations and might not be sufficient to address some regulators' explainability and transparency concerns.6 Alternatively, you can try to build explainability into your models. Although you might give up some predictive power, the approach can be a safer option.  For instance, you can identify features that could potentially lead to biased outcomes and limit their influence on the model. You can also compare the explainability of various ML-based models to see which may be more or less inherently explainable. For example, gradient boosting machines (GBMs) may be preferable to neural networks for this reason.7  You can also use ML to blend traditional and alternative credit data, which may provide a significant lift — around 60 to 70 percent compared to traditional scorecards — while maintaining explainability.8  READ MORE: Journey of an ML Model  How Experian can help  As a leader in machine learning and analytics, Experian partners with financial institutions to create, test, validate, deploy and monitor ML-driven models. Learn how you can build explainable ML-powered models using credit bureau, alternative credit, third-party and proprietary data. And monitor all your ML models with a web-based platform that helps you track performance, improve drift and prepare for compliance and audit requests. *When we refer to “Alternative Credit Data," this refers to the use of alternative data and its appropriate use in consumer credit lending decisions, as regulated by the Fair Credit Reporting Act. Hence, the term “Expanded FCRA Data" may also apply and can be used interchangeably.  1-3. FinRegLab (2021). The Use of Machine Learning for Credit Underwriting  4. Experian (2022). Explainability: ML and AI in credit decisioning  5. Experian (2023). Finding the Lending Diamonds in the Rough  6. FinRegLab (2021). The Use of Machine Learning for Credit Underwriting  7. Experian (2022). Explainability: ML and AI in credit decisioning  8. Experian (2023). Raising the AI Bar 

From chatbots to image generators, artificial intelligence (AI) has captured consumers' attention and spurred joy — and sometimes a little fear. It's not too different in the business world. There are amazing opportunities and lenders are increasingly turning to AI-driven lending decision engines and processes. But there are also open questions about how AI can work within existing regulatory requirements, how new regulations will impact its use and how to implement advanced analytics in a way that increases equitable inclusion rather than further embedding disparities. How are lenders using AI today? Many financial institutions have implemented — or at least tested — AI-driven tools throughout the customer lifecycle to: Target the right consumers: With tools like Ascend Intelligence ServicesTM Target (AIS Target), lenders can better identify consumers who match their credit criteria and send right-sized offers, which enables them to maximize their acceptance rates. Detect and prevent fraud: Fraud detection tools have used AI and machine learning techniques to detect and prevent fraud for years. These systems may be even more important as new fraud risks emerge, from tried-and-true methods to generative AI (GenAI) fraud. Assess creditworthiness: ML-based models can incorporate a range of internal and external data points to more precisely evaluate creditworthiness. When combined with traditional and alternative credit data*, some lenders can even see a Gini uplift of 60 to 70 percent compared to a traditional credit risk model. Manage portfolios: Lenders can also use a more complete picture of their current customers to make better decisions. For example, AI-driven models can help lenders set initial credit limits and suggest when a change could help them increase wallet share or reduce risk. Lenders can also use AI to help determine which up- and cross-selling offers to present and when (and how) to reach out. Improve collections: Models can be built to ease debt collection processes, such as choosing where to assign accounts, which accounts to prioritize and how to contact the consumer. Additionally, businesses can implement AI-powered tools to increase their organizations' productivity and agility. GenAI solutions like Experian Assistant accelerate the modeling lifecycle by providing immediate responses to questions, enhancing model transparency and parsing through multiple model iterations quickly, resulting in streamlined workflows, improved data visibility and reduced expenses. WATCH: Explore best practices for building, fine-tuning and deploying robust machine learning models for credit risk. The benefits of AI in lending Although lenders can use machine learning models in many ways, the primary drivers for adoption in underwriting include: Improving credit risk assessment Faster development and deployment cycles for new or recalibrated models Unlocking the possibilities within large datasets Keeping up with competing lenders Some of the use cases for machine learning solutions have a direct impact on the bottom line — improving credit risk assessment can decrease charge-offs. Others are less direct but still meaningful. For instance, machine learning models might increase efficiency and allow further automation. This takes the pressure off your underwriting team, even when application volume is extremely high, and results in faster decisions for applicants, which can improve your customer experience. Incorporating large data sets into their decisions also allows lenders to expand their lending universe without taking on additional risk. For example, they may now be able to offer risk-appropriate credit lines to consumers that traditional scoring models can't score. And machine learning solutions can increase customer lifetime value when they're incorporated throughout the customer lifecycle by stopping fraud, improving retention, increasing up- or cross-selling and streamlining collections. Hurdles to adoption of machine learning in lending There are clear benefits and interest in machine learning and analytics, but adoption can be difficult, especially within credit underwriting. A recent Forrester Consulting study commissioned by Experian found that the top pain points for technology decision makers in financial services were reported to be automation and availability of data. Explainability comes down to transparency and trust. Financial institutions have to trust that machine learning models will continue to outperform traditional models to make them a worthwhile investment. The models also have to be transparent and explainable for financial institutions to meet regulatory fair lending requirements. A lack of resources and expertise could hinder model development and deployment. It can take a long time to build and deploy a custom model, and there's a lot of overhead to cover during the process. Large lenders might have in-house credit modeling teams that can take on the workload, but they also face barriers when integrating new models into legacy systems. Small- and mid-sized institutions may be more nimble, but they rarely have the in-house expertise to build or deploy models on their own. The models also have to be trained on appropriate data sets. Similar to model building and deployment, organizations might not have the human or financial resources to clean and organize internal data. And although vendors offer access to a lot of external data, sometimes sorting through and using the data requires a large commitment. How Experian is shaping the future of AI in lending Lenders are finding new ways to use AI throughout the customer lifecycle and with varying types of financial products. However, while the cost to create custom machine learning models is dropping, the complexities and unknowns are still too great for some lenders to manage. But that's changing. Experian built the Ascend Intelligence Services™ to help smaller and mid-market lenders access the most advanced analytics tools. The managed service platform can significantly reduce the cost and deployment time for lenders who want to incorporate AI-driven strategies and machine learning models into their lending process. The end-to-end managed analytics service gives lenders access to Experian's vast data sets and can incorporate internal data to build and seamlessly deploy custom machine learning models. The platform can also continually monitor and retrain models to increase lift, and there's no “black box" to obscure how the model works. Everything is fully explainable, and the platform bakes regulatory constraints into the data curation and model development to ensure lenders stay compliant. Learn more * When we refer to “Alternative Credit Data," this refers to the use of alternative data and its appropriate use in consumer credit lending decisions as regulated by the Fair Credit Reporting Act (FCRA). Hence, the term “Expanded FCRA Data" may also apply in this instance and both can be used interchangeably.

This is the third in a series of blog posts highlighting optimization, artificial intelligence, predictive analytics, and decisioning for lending operations in times of extreme uncertainty. The first post dealt with optimization under uncertainty and the second with predicting consumer payment behavior. In this post I will discuss how well credit scores will work for consumer lenders during and after the COVID-19 crisis and offer some recommendations for what lenders can be doing to measure and manage that model risk in a time like this. Perhaps no analytics innovation has created opportunity for more individuals than the credit score has. The first commercially available credit score was developed by MDS (now part of Experian) in 1987. Soon afterwards FICO® popularized the use of scores that evaluate the risk that a consumer would default on a loan. Prior to that, lending decisions were made by loan officers largely on the basis on their personal familiarity with credit applicants. Using data and analytics to assess risk not only created economic opportunity for millions of borrowers, but it also greatly improved the financial soundness of lending institutions worldwide. Predictive models such as credit scores have become the most critical tools for consumer lending businesses. They determine, among other things, who gets a loan and at what price and how an account such as a credit line is managed through its life cycle. Predictive models are in many cases critical for calculating loan and loss reserves, for stress testing, and for complying with accounting standards. Nearly all lenders rely on generic scores such as the FICO® score and VantageScore® credit score. Most larger companies also have a portfolio of custom scorecards that better predict particular aspects of payment behavior for the customers of interest. So how well are these scorecards likely to perform during and after the current pandemic? The models need to predict consumer credit risk even as: Nearly all consumers change their behaviors in response to the health crisis, Millions of people—in America and internationally—find their income suddenly reduced, and Consumers receive large numbers of accommodations from creditors, who have in turn temporarily changed some of their credit reporting practices in response to guidelines in the federal CARES Act. In an earlier post, I pointed out that there is good reason to believe that credit scores will tend to continue to rank order consumers from most likely to least likely to repay their debts even as we move from the longest economic expansion in history to a period of unforeseen and unexpected challenges. But the interpretation of the score (for example, the log odds or the bad rate) may need to be adjusted. Furthermore, that assumes that the model was working well on a lender’s population before this crisis started. If it has been a long time since a scorecard was validated, that assumption needs to be questioned. Because experts are considering several different scenarios regarding both the immediate and long-term economic impacts of COVID-19, it’s important to have a plan for ongoing monitoring as long as necessary. Some lenders have strong Model Risk Management (MRM) teams complying with requirements from the Federal Reserve, Federal Deposit Insurance Corporation (FDIC), the Office of the Comptroller of the Currency (OCC). Those resources are now stretched thin. Other institutions, with fewer resources for MRM, are now discovering gaps in their model inventories as they implement operational changes. In either case, now’s the time to reassess how well scorecards are working. Good model validation practices are especially critical now if lenders are to continue to make the sound data-driven decisions that promote fairness for consumers and financial soundness for the institution. If you’re a credit risk manager responsible for the generic or custom models driving your lending, servicing, or capital allocation policies, there are several things you can do--starting now--to be sure that your organization can continue to make fair and sound lending decisions throughout this volatile period: Assess your model inventory. Do you have good documentation showing when each of the models in your organization was built? When was it last validated? Assign a level of criticality to each model in use. Starting with your most critical models, perform a baseline validation to determine how the model was performing prior to the global health crisis. It may be prudent to conduct not only your routine validation (verifying that the model was continuing to perform at the beginning of the period) but also a baseline validation with a shortened performance window (such as 6-12 months). That baseline validation will be useful if the downturn becomes a protracted one—in which case your scorecard models should be validated more frequently than usual. A shorter outcome window will allow a timelier assessment of the relationship between the score and the bad rate—which will help you update your lending and servicing policies to prevent losses. Determine if any of your scorecards had deteriorated even before the global pandemic. Consider recalibrating or rebuilding those scorecards. (Use metrics such as the Population Stability Index, the K-S statistic and the Gini Coefficient to help with that decision.) Many lenders chose not to prioritize rebuilding their behavioral scorecards for account management or collections during the longest period of economic growth in memory. Those models may soon be among the most critical models in your organization as you work to maintain the trust of your accountholders while also maintaining your institution’s financial soundness. Once the CARES accommodation period has expired, it will be important to revalidate your models more frequently than in the past—for as long as it takes until consumer behavior normalizes and the economy finds its footing. When you find it appropriate to rebuild a scorecard model, consider whether now is the time to implement ethical and explainable AI. Some of our clients are finding that Machine Learned models are more predictive than traditional scorecards. Early Experian research using data from the last recession indicates this will continue to be true for the foreseeable future. Furthermore, Experian has invested in Research and Development to help these clients deliver FCRA-compliant Adverse Action reasons to their consumers and to make the models explainable and transparent for model risk governance and compliance purposes. The sudden economic volatility that has resulted from this global health crisis has been a shock to all organizations. It is important for lenders to take the pulse of their predictive models now and throughout the downturn. They are especially critical tools for making sound data-driven business decisions until the economy is less volatile. Experian is committed to helping your organization during times of uncertainty. For more resources, visit our Look Ahead 2020 Hub. Learn more