This is a discussion forum for S for SAS, SAS Primer for Consumer Finance Analysts. SAS certification, use of SAS in lending environments etc are discussed

Sunday, May 27, 2007

S4SAS - A beginners Guide

Welcome to S4SAS

Objective of this guide is to equip an entry level analyst with basic SAS skills in a week’s time.

SAS in Consumer Lending Analytics

SAS System is mainly used for data management, scoring and modeling, statistical analysis and reporting by various departments like marketing, risk management, collections and operations. Further, all these activities fall into one of the two life cycle phases of a lending business; Acquisition and Account Management.

Data Management

Every analytics activity precedes a data preparation stage in which significant amount of time and efforts are spent to consolidate, clean and format data. Consumer lending businesses usually have a complex data environment with large data warehouses built to store application data, billing information and every day transactions. Additionally, data from credit bureaus and other data vendors are also stored for various analytics purposes. SAS has built in modules and functions to handle data from various sources and convert them into a format required for the analytics purposes.

Scoring and Modeling

Scorecard is a statistical model that attributes a number (score) to a customer which corresponds to a probability that the customer exhibits certain behavior in the future. A most widespread type is an application scorecard which predicts the probability of a person default on the borrowing. Objective of the scorecards can vary depends on the department; A behavioral scorecard predicts the probability of an existing account turning bad; A response scorecard predicts the probability of a person likely to respond to an offer; A collection scorecard predicts the probability of a customer pays back the owed money. Development of scorecard is a detailed statistical procedure typically using logistics regression and other procedures in SAS.

Statistical Analysis

A statistical analysis can be classified as a one-time report or diagnostics with a specific objective. For example, identifying the loss drivers in a portfolio or data segmentation for a credit line increase or recognizing a fraud pattern etc. can be considered as a typical analysis project. Analysis involves dealing with large amount of data and drawing relationship out of them using statistical procedures.


An accurate and robust reporting system is an essential part of any consumer lending business. The management relies largely on reporting to get early pointers on growth as well as troubles. Regular reports produced at various frequencies report different metrics like number of applications, approvals, average scores and average credit limits ; Performance metrics like outstanding amounts , sales, charge-offs , bankruptcy, fraud ; Collection related metrics like delinquency stages and amounts , collection efficiency, roll rates and recoveries ; Vintage analysis of performance metrics by acquisition and account management scores ; Strategy tracking dashboards in a champion challenger framework ; Scorecard validation reports etc are some examples of standard reports produced in a lending environment.

Now let us look at the applications of SAS in various life cycles of consumer lending.


This is the initial phase in lending where the analytics is used to target and acquire the right people. One of the main channels of acquisition is through the direct mail campaigns and often, the pre-approved (pre screened) credit cards are mailed to prospects. Mail campaigns are expensive operations and so it’s important that the response rates for such campaigns are sufficient to balance the cost-benefit sides of the equation.

SAS System is used to segment the large address lists and demographic information available in the credit bureaus and the segments that are most likely to respond (based on past experience) to the campaign are selected for the mail solicitation . In cases of through-the-door or internet applications, the business rules developed using SAS – also known as data driven strategies- is used to evaluate the applications, often online, to come out with a decision. Further, acquisitions and strategy tracking dashboards are developed using SAS to monitor the acquisition process and also to compare the performance of various strategies.

Now let us look at how various analytics department use SAS System at the acquisition stage; Risk team scores the prospects based on the risk rules and does the pre-screening to avoid the chances of soliciting bad accounts. Marketing department profiles the prospects and identify the likely segments that are going to bring more sales. Operations team does the segmentation to maximize the response rates for the mail campaign. All these teams develop the monitoring dashboards using SAS to track their objectives once the plan is rolled out and share their finding as reports/analysis to the senior management.

We will discuss the campaign operations and custom score card development process using SAS further in this books These two processes are widely used in the lending environment for the prospecting and acquisition.

Account Management:

Once the accounts are acquired, authorization of repeated credit purchases, credit line management, fraud detection, sales promotion, cross-selling and collections are some of the life-cycle events that occurs in account management phase. Departments use leverage on analytics to effectively manage these business situations.

Risk management closely monitors the delinquency and utilization patterns along with charge-offs, bankruptcies and fraud occurrences. Authorization and credit line policies are modified based on the learning to reduce loss or improve profitability. Behavioral score cards, vintage performance dashboards and reports are developed support the decision process.

Marketing analyzes activation and sales patterns and prompt the customers to buy more through rewards and other offers. Cross-selling of financial products like personal loans and mortgages to existing customers is another area marketing analytics focuses on.
Collections analytics team identifies the likely defaulters using behavioral scorecards and based on the risk profiles, devise strategies for collections. They use the past experiences to improve the collection efficiency and work closely with the operations team to implement the data driven strategies.

Operations analytics team, typically aligned with customer service operations, forecasts the call volumes and analyzes call logs to improve the call routing and support processes. Collection efficiency of the analysts in terms of number of calls and amount collected are some of the metrics monitored by this department.

To Sum-up, SAS System provides a variety of tools and utilities for analysts in consumer lending environment to perform an array of analytics activities ranging from simple reporting to complex statistical modeling . Number crunching abilities of SAS in a voluminous data environment is much appreciated in consumer lending than any other industry.

The U.S.A Lending Market

The U.S. market witnessed an astounding growth in terms revolving credit in the last two decades. Federal Reserve Statistical releases show that in 1985 the average revolving credit outstanding was about $ 115 Billion (20% of the total consumer credit).In 2006, these average numbers soared to 843 Billion (36% of total consumer credit). Survey of Consumer Finances observed that, of the 75% households with a revolving credit line, 58% had a balance outstanding. This growth was noticed by a lot of financial institutions that led to their entry into consumer lending business.

Within the last decade many consolidations happened in consumer lending businesses through acquisitions, mergers and purchases. Bank One’s acquisition of First USA in 1997, Chase’s purchase of Providian in 2002, HSBC’s acquisition of Household in 2003, Citi’s purchase of Sears in 2003, Chase’s purchase of Bank One in 2004, Bank of America’s acquisition of Fleet Bank and Bank of America’s merger with NBNA in 2005 were some of the major deals impacted the market share of consumer lending businesses. Bank of America now leads the bunch with 40 million active accounts and $143 billion outstanding balances. Chase, Citigroup, Capital One and Discover are the other four players in the top five.

Managing by the Odds.

Over the period of time, lending to consumers profitably proved to be one of the most challenging to the businesses involved. Intense competition for market share and large volume of small loans to service are the two major reasons prompted the lenders to approach the way they do business differently. Increased automation in decision processes and managing the portfolios by the odds (in a statically predictable manner) were opted by all the lenders to stay competitive. Large scale investments were made on rule based systems, data warehouses and data mining. The benefits were tangible; increased responses on the expensive mail solicitations, higher approval of applications, consistent underwriting processes, improved fraud detection, data driven portfolio management, comprehensive reporting and MIS , optimization of call center operations, early delinquency detection and efficient collection efforts to name a few. All these efforts, collectively named as ‘analytics’, hence formed an inevitable part of the consumer lending business.

Consumer lending is a volume based business with millions of small loans and transactions handled on a daily basis with thin profit margins. In such a scenario, managers depend on the analytics departments to control the risk-returns statistically. This requires data warehousing, scorecard development, data segmentation and pattern analysis, tracking dashboards and comprehensive reporting systems. SAS System is widely used by these businesses in meeting these requirements.

Wednesday, August 30, 2006

S for SAS Beginner's Guide

Welcome to S4SAS

Objective of this guide is to equip an entry level analyst with basic SAS skills in a week’s time.