The Credit Reporting Industry is About to Experience the Biggest Change in Decades... Are You Prepared?

Similar documents
Confidently Assess Risk Using Public Records Data with Scalable Automated Linking Technology (SALT)

INSIGHT ADVANCING. Lexis Advance. Find just what you re looking for faster with research innovations inspired by legal professionals like you.

FALSE PROMISE. That high-value prospect could be a synthetic identity

Analogy Engine. November Jay Ulfelder. Mark Pipes. Quantitative Geo-Analyst

NCRIS Capability 5.7: Population Health and Clinical Data Linkage

Using Iterative Automation in Utility Analytics

White Paper. What is Enterprise MDM and Why You Need It

Patent portfolio audits. Cost-effective IP management. Vashe Kanesarajah Manager, Europe & Asia Clarivate Analytics

The A.I. Revolution Begins With Augmented Intelligence. White Paper January 2018

IBM SPSS Neural Networks

Pan-Canadian Trust Framework Overview

UNIT-III LIFE-CYCLE PHASES

HARNESSING TECHNOLOGY

HOW AI BOOSTS INDUSTRY SPOTLIGHTS INDUSTRY PROFITS AND INNOVATION

Comparison: On-Device and Drive Test Measurements

CREDITING-RELATED READINESS ACTIVITIES UNDER THE PMR: UPDATE AND SUGGESTED NEXT STEPS

Technology forecasting used in European Commission's policy designs is enhanced with Scopus and LexisNexis datasets

Achieving Desirable Gameplay Objectives by Niched Evolution of Game Parameters

Removing Duplication from the 2002 Census of Agriculture

Connecting Australia. How the nbn broadband access network is changing Australia. An economic study of the way we work, live and connect.

Executive Summary FUTURE SYSTEMS. Thriving in a world of constant change

5G, IoT, UN-SDG OMA LwM2M, IPSO

Annotated Chapter Outline

Where tax and science meet part 2*

2005 Modelithics Inc.

Visual Analytics in the New Normal: Past, Present & Future. geologic Technology Showcase Adapting to the New Normal, Nov 16 th, 2017

CONSENT IN THE TIME OF BIG DATA. Richard Austin February 1, 2017

Privacy and the EU GDPR US and UK Privacy Professionals

COMPREHENSIVE COMPETITIVE INTELLIGENCE MONITORING IN REAL TIME

Upstream Oil and Gas. Spill Prevention, Preparedness, Response, and Recovery. March 2013

ENGINEERING SERVICES CONSULTANCY

Visualization in automotive product development workflow

YOUR OWN HEADHUNTING BUSINESS

Economic and Social Council

An ESRI White Paper May 2009 ArcGIS 9.3 Geocoding Technology

Imagine your future lab. Designed using Virtual Reality and Computer Simulation

Central Cancer Registry Geocoding Needs

PEAK GAMES IMPLEMENTS VOLTDB FOR REAL-TIME SEGMENTATION & PERSONALIZATION

Assessing Geocoding Solutions

An Intellectual Property Whitepaper by Katy Wood of Minesoft in association with Kogan Page

Years 9 and 10 standard elaborations Australian Curriculum: Digital Technologies

ATLAS. P25 Systems. LMR communications made simple.

Our position. ICDPPC declaration on ethics and data protection in artificial intelligence

Assessing the Welfare of Farm Animals

Patent Mining: Use of Data/Text Mining for Supporting Patent Retrieval and Analysis

Interview Guidance for Hiring Managers. Page 1 of 14

Supplementary Data for

Smart cities: A human-centered approach Engineering and Construction Conference June 20 22, 2018

Intelligent, Rapid Discovery of Audio, Video and Text Documents for Legal Teams

Phase One: Determine Top 5 Teams

Helping good businesses become great businesses

Plan Your Financial Future

Digital without Compromise.

Is Texas Ready for Mileage Fees? Results from Exploratory Study Presentation to the Texas Transportation Commission December 15, 2010.

Instrumentation and Control

SURVEY ON USE OF INFORMATION AND COMMUNICATION TECHNOLOGY (ICT)

WHAT SMALL AND GROWING BUSINESSES NEED TO SCALE UP

Lexisnexis PatentOptimizer Streamline your patent analysis and applications

CircumSpect TM 360 Degree Label Verification and Inspection Technology

Serco Awarded $95 Million Patent Classi cation Contract with the U.S. Patent and Trademark Of ce

Powering Human Capability

Pharmaceutical Sector Inquiry Presentation of the Preliminary Report. 28 November 2008

PRODUCT INFORMATION FORM (PIF TM )

Writing a Business Plan

Digital Engineering Support to Mission Engineering

Smarter oil and gas exploration with IBM

Tommy W. Gaulden, Jane D. Sandusky, Elizabeth Ann Vacca, U.S. Bureau of the Census Tommy W. Gaulden, U.S. Bureau of the Census, Washington, D.C.

ANALYTICS EXCELLENCE WEBINAR SERIES

MORE POWER TO THE ENERGY AND UTILITIES BUSINESS, FROM AI.

Infringement and Enforcement Panel How can you identify infringement and enforce your rights?

L A N D R A Y P R O D U C T 1 BREAKTHROUGH PERFORMANCE BY GROUND PENETRATING RADAR

in the New Zealand Curriculum

APX TM TWO-WAY RADIOS NORTH AMERICA REGION. MotorolaSolutions.com/Services

The Savvy Survey #3: Successful Sampling 1

Implementing BIM for infrastructure: a guide to the essential steps

The real impact of using artificial intelligence in legal research. A study conducted by the attorneys of the National Legal Research Group, Inc.

The robots are coming, but the humans aren't leaving

A REAL ESTATE INVESTOR S GUIDE TO MANAGING CONTRACTORS FROM START TO FINISH. LendingOne YOUR LOGO

Executive Summary Industry s Responsibility in Promoting Responsible Development and Use:

ZoneFox Augmented Intelligence (A.I.)

TRUSTING THE MIND OF A MACHINE

DESIGN THINKING AND THE ENTERPRISE

Find and analyse the most relevant patents for your research

OECD WORK ON ARTIFICIAL INTELLIGENCE

ARGUING THE SAFETY OF MACHINE LEARNING FOR HIGHLY AUTOMATED DRIVING USING ASSURANCE CASES LYDIA GAUERHOF BOSCH CORPORATE RESEARCH

Castan Centre for Human Rights Law Faculty of Law, Monash University. Submission to Senate Standing Committee on Economics

BUREAU OF LAND MANAGEMENT INFORMATION QUALITY GUIDELINES

Novatek NT35695 LCD Driver IC

OUR VISION. Explaining the WHY, the WHAT and the HOW of SBM Offshore.

The Profitable Side Project Handbook

Instrumentation, Controls, and Automation - Program 68

STRATEGIC FRAMEWORK Updated August 2017

Our Quality Promise WHITE PAPER

SPECIAL REPORT. The Smart Home Gender Gap. What it is and how to bridge it

Violent Intent Modeling System

MASTER DATA MANAGEMENT 7 QUESTIONS TO CONSIDER

Where does architecture end and technology begin? Rami Razouk The Aerospace Corporation

Digital Identity Innovation Canada s Opportunity to Lead the World. Digital ID and Authentication Council of Canada Pre-Budget Submission

2. What is Text Mining? There is no single definition of text mining. In general, text mining is a subdomain of data mining that primarily deals with

Standing Committee on the Law of Patents

Transcription:

The LexisNexis RiskView Liens & Judgments Report The Credit Reporting Industry is About to Experience the Biggest Change in Decades... Are You Prepared? When access to most lien & civil judgment data is cut off later this year, it won t be business as usual.

It s time to reconsider your default position The clock is ticking: In July 2017, the three Nationwide Credit Reporting Agencies will eliminate about 50% of tax liens and almost all civil judgments from their credit reports as part of their National Consumer Assistance Plan. You have a lot to consider especially since lien and civil judgment data has been a key piece of credit decisioning for decades. Uncertainty isn t a good look in lending These changes will certainly have an impact on portfolio performance and profitability. The question is, to what extent? You ll need to update critical decision workflows, models and scorecards, somewhat in a vacuum. Maybe you figure you ll just weather the storm and see how everything plays out. After all, your competitors are in the same boat. How bad can it be? Rose-colored glasses hide reality You ve been dealt lemons, and you think you can make lemonade. But is it that easy? Going back to the old ways of doing things evaluating credit applicants based purely on their credit bureau tradeline history could severely decrease your predictive power. There is little room for error. Your boss wants to know the plan now. What s it going to be: Ride it out and hope for the best? Not a good idea. Testing and analysis shows it s too risky. What about tweaking your models and processes? Again, risky. You know from past experience that tinkering with this even a little can potentially increase default rates. Regulators won t like that. Let s look at this another way: Liens and civil judgments are still out there What if you could actually continue to reliably and confidently access this data? Rather than potentially taking on greater financial losses and regulatory risk, continued use of timely lien and civil judgment content could allow you to sustain growth and profitability. And give you leg-up on unprepared competitors. Why lean on liens & civil judgments to assess risk? About 11% of the U.S. population has either a tax lien or civil judgment. 1 Almost 1/3 have multiple liens and judgments. A consumer who has a lien or civil judgment removed from their credit file experiences an increase of 10 points to their credit score on average. 2 1 The LexisNexis RiskView Liens & Judgments Report

A consumer who has a lien or civil judgment removed from their credit file experiences an increase of 10 points to their credit We provide what the credit bureaus can t LexisNexis Risk Solutions is the answer to the lien and civil judgment data gap. The LexisNexis RiskView Liens & Judgments Report delivers technology advancements that bolster the reliability and currency of lien and civil judgment content. Our solution offers: score on average. Advanced Linking Technology: Identity information from nationwide credit bureaus, expansive public records, hundreds of other data sources, and supercomputer technology are leveraged to provide our industry leading LexID identity linking technology Proven Results: LexID exceeds 99% linking precision, so you can better connect dots between identities and records Consumers with a lien or judgment record on file are about 2x more likely to default on a debt obligation than those without a lien or judgment. 3 Current Insights: Nationwide network of court runners provides the most current public record content available 1 VantageScore Solutions, The Impact to Credit Scores & Credit Score Models From Data Suppression, 2016 2 Ibid. 3 LexisNexis data The LexisNexis RiskView Liens & Judgments Report 2

July is just around the corner and you ve got decisions to make Don t give up the lien and civil judgment data you depend on for credit risk decisioning. Book an assessment to learn more about how we can incorporate this content into your process. Call 800.869.0751 or visit lexisnexis.com/creditrisk About LexisNexis Risk Solutions LexisNexis Risk Solutions (www.lexisnexis.com/risk) is a leader in providing essential information that helps customers across all industries and government assess, predict, and manage risk. Combining cutting-edge technology, unique data and advanced scoring analytics, we provide products and services that address evolving client needs in the risk sector while upholding the highest standards of security and privacy. LexisNexis Risk Solutions is part of RELX Group plc, a world-leading provider of information and analytics for professional and business customers across industries. LexisNexis, the Knowledge Burst logo, and LexID are registered trademarks of RELX Inc., used under license. HPCC Systems is a registered trademark of LexisNexis Risk Data Management Inc. Copyright 2017 LexisNexis Risk Solutions. All rights reserved. NXR11549-00-0117

WHITE PAPER Linking Liens and Civil Judgments Data Confidently Assess Risk Using Public Records Data with Scalable Automated Linking Technology (SALT)

Table of Contents Executive Summary... 3 Collecting Liens & Civil Judgments Public Records Data... 3 Extensive Data Coverage The Basis for Reliable Linking... 3 Superior Linking Technology... 4 Traditional Linking vs. SALT-based Linking... 6 Effect of SSN in Linking Lien & Civil Judgment Records... 6 Assessing Reliability of Liens & Judgments Linking... 7 Conclusion... 8 Glossary... 8 2

Executive Summary Public record lien and civil judgment data has been used by the financial services industry for decades. Liens and civil judgments help financial institutions assess the creditworthiness of consumers, and inform account receivable strategy to avoid unnecessary and unproductive lawsuits. Given the value and sensitivity of this content, it is constructive to understand how LexisNexis Risk Solutions collects and links this data to provide the most current and reliable data available in the market. This paper provides a description of the processes, procedures, and technology that enable LexisNexis to collect this data and reliably link it to the correct consumer, including the company s proprietary Scalable Automated Linking Technology (SALT), cutting-edge big data technology, and vast data resources. Collecting Liens & Civil Judgments Public Records Data The ability to provide solutions containing public lien and civil judgment data begins with dependable data collection. LexisNexis Risk Solutions has robust policies and procedures to maximize data reliability and ensure our records accurately reflect the underlying public records. We obtain lien and civil judgment information from government sources, which vary in their systems and policies regarding the availability of public record data. LexisNexis collects public records data from over 3,000 counties, boroughs, and parishes in the United States, representing over 98% of the U.S. population 1. We utilize digital access methods where viable, in combination with a national network of collection vendors. We collect from each jurisdiction at a frequency of 90 days or less, with a majority of jurisdictions visited on a monthly cycle. Also, beginning in July 2017, LexisNexis is implementing system upgrades to enable more robust collection performance management across hundreds of data points to discern sources of latency, including vendor collection performance and government record-reporting latency. LexisNexis is particularly focused on the collection of judgment and lien dispositions of paid, vacated or filed in error documents, and continues to invest in audit and collection methodologies to ensure confidence in our comprehensive and timely collection of this information. The results of these systems and investments is a broad dataset of public records data, including liens and judgments, that is current, reliable, and available for linking to consumers through LexisNexis Risk Solutions state-of-the-art linking technology. Extensive Data Coverage The Basis for Reliable Linking LexisNexis compiles the largest collection of U.S. consumer identity information available today. We leverage approximately 65 billion public and proprietary records, which are updated regularly. This rich dataset allows LexisNexis to understand the identity information for almost the entire U.S. consumer population (not just a subset of the credit active population), and provides the foundation for our ability to link consumers to public records accurately. This broad coverage enables us to assess the 1 Based on census data from July 2015. 3

uniqueness of different combinations of identity information and to evaluate whether a given record belongs to a given consumer. Superior Linking Technology LexisNexis utilizes a proprietary and patented data linking approach called Scalable Automated Linking Technology (SALT) to draw upon our extensive data and turn disparate information into meaningful insights. SALT allows the analysis of large data sets more easily, reliably, and efficiently due to the High Performance Computing Cluster (HPCC) Big Data technology platform. HPCC provides the processing power necessary to run billions of complex statistical analyses and data comparisons. These statistical clustering algorithms run iteratively, learning through repetitive analyses and as incremental data is updated. Core to the linking algorithms is specificity. Specificity is the measure of uniqueness assigned to each value of data, in each field of the database. Every relevant identity element and combination thereof is assigned a specificity; this allows the system to identify when records have sufficient evidence to confidently match. The automated analysis of data, measuring specificity, informs the matching function for each unique pair of records. The outcome of the SALT linking algorithms is the assignment of our patented LexID to each identity, a unique identification number for individuals that is a reliable and secure indicator of an individual. Once created, the LexID is assigned to the set of records associated with each unique individual. LexisNexis performs ongoing analysis of the results of this linking system, periodically auditing a significant sample of the linked identities. Additionally, with each iteration, a sample of newly-linked records is analyzed by a group of engineers and engineering managers with linking expertise. These procedures are fine-tuned over time, as product teams, quality assurance teams, consumers, and customers provide feedback. Patented LexisNexis Linking Technology Leverages a Unique Identifier Creating a reliable view of an individual or business. When we acquire a new record, LexisNexis assigns our own unique identifier, LexID. To ensure the security of the consumer identities, LexID is not derived from any personally identifiable information, such as a social security number or name. Records with a common LexID are linked together using LexisNexis Scalable Automated Linking Technology, a proprietary and patented method of linking and clustering data. Identity profiles are continuously updated to ingest new data sets and records. Having this persistent link across multiple touchpoints and data silos eliminates false positives and builds an extremely comprehensive and accurate representation of the identities that matter to our customers. Figure 1: LexisNexis Linking Technology. Our high-precision data linking technology drives linking effectiveness by building extremely comprehensive and precise profiles of consumer identities. 4

Leveraging multiple sources and linking these together using proprietary and patented linking technology, LexID delivers a more complete view of individuals, especially those with multiple names and addresses. By combining data integrity with linking algorithms, we make connections and identify relationships with high reliability. This dynamic process results in a consumer profile that accounts for identity changes over time, since LexisNexis continually updates profiles with new records. By aggregating, cross-linking, and analyzing large volumes of identity data, LexisNexis is able to build complete identity portraits that connect otherwise fragmented information and disparate touchpoints across time. Even when there are misspellings, typographical errors, or other errors in the originating data sources, LexisNexis can associate records with the same individual. We are also able to recognize when someone may have presented himself in different ways over time (i.e., Alexander Jonathan Marks vs. A.J. Marks, Alex Marks, or John Marks ). 65+ billion public records distilled down to... 10+ billion unique name/adress combinations, which we cross-link, refine, and map to isolate 300+ million unique identities, which are further enhanced to remove deceased and inactive identities, resulting in a precise view of 279+ million current active identities Figure 2: Distillation of address data. LexisNexis turns massive raw data into precise, crosslinked, and actionable results. Figure 2 graphically depicts what LexisNexis linking technology means for end users. The breadth of LexisNexis public records coverage is demonstrated by the more than 10 billion unique name/address combinations contained within our databases, which 5

we have cross-linked and mapped to more than 300 million identities and to which each has been assigned a unique LexID. This linking process affords insight into more than 279 million currently active identities. Traditional Linking vs. SALT-based Linking Traditional, deterministic record linkage also known as data matching generally uses if/then logic to assign a series of confidence factors that indicate the degree to which the content in each field can be considered the same as, similar to, or different from content in an existing consumer file. When this process finds a record that meets the match rules, it links the record. As such, attempts to improve linking accuracy rely in large part on increasing the number of minimum fields in source data records in order to establish suitable confidence in the match. In contrast, SALT uses advanced concepts such as term specificity to determine the relevance/weight of a particular field in the scope of the linking process, and proprietary computer algorithms based on the input data rather than the need for hand-coded user rules, which is key to overall linking effectiveness. SALT routines also measure the statistical strength of a match, allowing LexisNexis to require strong matches before linking records to existing consumer files. Compared to traditional if/then logic, SALT-based linking allows for consideration of a greater number of factors across more dimensions, which increases accuracy and reliability in a statistically significant manner. The combination of the ability to evaluate the uniqueness of a set of input values combined with the logic to prevent a link that is not sufficiently strong allows our linking technology to overcome many of the challenges associated with linking public record data. This includes the scenario when two people with the same name but a different name suffix live at the same address (what is commonly referred to as the Junior/Senior scenario ). In such cases, our linking technology can often distinguish between such individuals, but in the rare instance when there is not enough data to accurately assign ownership of a record to the right individual, we will not make the link. The end result of the SALT-based approach is a linking technology that uses analytics and machine learning to avoid both false positive and false negative results, and to deliver highly reliable linking outcomes without imposing rigid, hand-coded rules. Effect of SSN in Linking Lien & Civil Judgment Records One of the critical aspects of linking liens and judgments records is the ability to match records to consumers reliably without the presence of a Social Security Number (SSN) on the public record. This is critical because approximately 50% of tax lien records and approximately 96% of civil judgment records do not contain a SSN. In order to verify the reliability of our linking technology as it relates to the data available within liens and civil judgments, LexisNexis conducted a study to test the consistency of linking results in the absence of SSN. For this study, a random sample of approximately 26M lien and civil judgment records containing a full SSN or partial SSN (i.e. containing 6

only the last four digits) was processed twice through our linking processes. In one round of processing, the SSN values in the records were available for use in linking. In the other round of processing, the SSN values on the records were suppressed from the data and not available for use in the linking. Following the two rounds of linking, the results were investigated to determine, for the set of records that were matched to a LexID in both rounds of linking, how often a different LexID was assigned when the SSN was suppressed compared to the LexID assigned when the SSN was not suppressed. As shown in the table below, the results of the study indicate that the presence of a SSN on the source record led to the assignment of a different LexID approximately 0.09% and 0.10% of the time in civil judgments and liens, respectively. We would note that the focus of this study was on the consistency of our linking with and without a SSN, and the different linking results were not reviewed in detail for accuracy the issue of linking accuracy will be discussed in the next section. However, by examining the differences in LexID assignment with and without a SSN on the record, we can demonstrate the high degree of consistency and robustness of our linking technology. Public Record Type Different LexID match with the SSN suppressed (%) LIENS 0.10% CIVIL JUDGMENTS 0.09% Table 1. Linking study with SSN suppression. Assessing Reliability of Liens & Judgments Linking LexisNexis also conducted a separate study to assess SALT links within lien and civil judgment data. A team of engineers with linking expertise manually reviewed a sample of 1000 liens and judgments records (500 randomly sampled liens and 500 randomly sampled judgments). In the manual review process, the linking engineers reviewed the public record, the assigned LexID, and the other closest LexID matches (e.g. the second and third closest candidates) to confirm that the selected LexID was the appropriate match for the record. Each record being reviewed was labeled Correct where the assigned LexID was the appropriate identity for the record, Incorrect when the assigned LexID was not appropriate for the record, and Maybe when the validity of the match was unclear or ambiguous. Our experience has shown that in the Maybe cases, where it is still unclear if the match is correct after deep review, it is conservative to estimate that half of the Maybe results are incorrect. Given this assumption, the final accuracy of the sample can be estimated as the sum of the Correct results and half of the Maybe results divided by the total number of records included in the review. Based on this approach, the linking accuracy for liens and judgments can be estimated at 99.75%, 99% CI [98.90%, 99.94%] (see Table 2). 7

Match Quality Lien & Judgment Records CORRECT 995 MAYBE 5 INCORRECT 0 TOTAL 1000 SAMPLE ACCURACY 99.75% Table 2. Results from manual accuracy study of liens and judgments linking. Conclusion LexisNexis Risk Solutions leverages proprietary linking algorithms, cutting-edge big data technology, and vast data resources to reliably link lien and civil judgment public record content to the correct consumer files. This enables our customers to continue to rely on this valuable content to make informed and confident fact-based decisions. Glossary Consumer file: A collection of records associated with a given consumer. Linking: The process of matching a record or inquiry to a unique consumer identity. Record: A discrete unit of data that has defined values within fields. Fields are the critical data elements used for record linkage. Specificity: A measure of uniqueness assigned to each value in each field of a database. 8

For More Information Call 800.715.0959 or visit lexisnexis.com/creditrisk. About LexisNexis Risk Solutions LexisNexis Risk Solutions (www.lexisnexis.com/risk) is a leader in providing essential information that helps customers across all industries and government assess, predict and manage risk. Combining cutting-edge technology, unique data and advanced analytics, LexisNexis Risk Solutions provides products and services that address evolving client needs in the risk sector while upholding the highest standards of security and privacy. LexisNexis Risk Solutions is part of RELX Group plc, a world-leading provider of information solutions for professional customers across industries. LexisNexis and the Knowledge Burst logo are registered trademarks of RELX Inc., used under license. Copyright 2017 LexisNexis. NXR12085-00-0617-EN-US