Information about http://www.epaynews.com/resources/BasePoint_Road_to_Profitability_White_Paper.pdf

White Paper REBUILDING THE MORTGAGE INDUSTRY � THE PATH BACK TO …

Tags: analytics, circumstances, current mortgage, executive summary, investment banking firms, lenders, loan programs, mortgage fraud, mortgage industry, organizational structures, portfolios, private equity firms, profitability, risk management technology, single reason, squeeze,
Pages: 11
Language: english
Created: Fri Jan 25 15:51:07 2008

Display cached document

Page 1

Page 2

Page 3

Page 4

Page 5

Page 6

Page 7

Page 8

Page 9

Page 10

Page 11

White Paper
REBUILDING THE MORTGAGE
INDUSTRY � THE PATH BACK TO
PROFITABILITY

� 2008 BasePoint Analytics LLC. All Rights Reserved
Executive Summary
There has been a lot written in the media recently about the current
mortgage crisis, and what impact it has had on its victims. Certainly there
is not a single reason or cause; rather a complex web of circumstances that
created this situation. Consequently, the road to recovery will comprise
many components to help ensure this situation doesn't recur.

As we've seen, this market has put a real profitability squeeze on
originators and investment banking firms, and many lenders could not
survive. There has been a considerable shake-out in the industry. Many
fewer originators exist. Some lenders are now owned by investment
banking or private equity firms. Many loan programs have been eliminated,
and many changes have been made in underwriting policy. There is a lot of
talk about what is needed as the industry rebuilds itself, including
discussions of new regulations, "back to basics" underwriting, and more
resources for investigating and prosecuting mortgage fraud. While there is
merit in many of these ideas, BasePoint engenders an approach founded in
using fraud and risk management technology to its full advantage, and
building the right organizational structures, as well as the most appropriate
policies and processes to profitably rebuild the market.

BasePoint recommends a multi-step process to help originators and
investment banking firms improve the quality of their portfolios going
forward. These steps include:
Understanding and measuring mortgage fraud and early
payment default (EPD),
Realigning risk management functions and refocusing them on
protecting the quality of loan portfolios,
Systematically monitoring brokers and appraisers,
Leveraging analytic technology to filter applications and loan
pools for high risk for fraud and EPD, and
Using a standardized process to review high risk applications and
loans.

This paper will explore these recommendations in detail and illuminate the
road to successful mortgage industry recovery. Now is the time for lenders
and investment banking firms to take full advantage of next generation
technology solutions to support rebuilding their profitability in this changed
market.

2
The Road Back to Growth and Profitability
It's hard to escape the dim news, performance and predictions US mortgage
lenders and investors are facing every day. What a difference one year can
make. A year ago mortgage originators were funding almost $3 trillion in Now is the time
new loans a year, and many were making record profits. As we've now for lenders and
seen, there has been a persistent rise in mortgage fraud and EPD that, investment
banking firms to
along with other economic factors, threaten this market. Origination take full
volumes have dropped substantially, which adds to the profitability pressure advantage of
for lenders both large and small. In addition, many investment banks have next generation
become more aggressive about monitoring for fraud and early payment technology
solutions to
defaults over the past year, and kicking back loans to originators as allowed
support
by their purchase agreements. In addition, investment banks have scaled rebuilding their
back or ceased buying mortgages for mortgage backed securities all profitability in
together. These factors drove many lenders out of business. The road this changed
toward rebuilding mortgage industry profitability requires getting back to market.
the basics of a solid risk management strategy, a comprehensive fraud
management and risk management organizational structure, and sound
policies, processes and use of technology.

Now that we have experienced a crisis of profitability across the US
mortgage industry, there is a lot of discussion about what is needed to
revitalize industry growth and how to rebuild the origination and secondary
markets to better protect them from future risk. New state and federal
regulations, a "back to basics" underwriting approach, training and
education, and more resources to investigate and prosecute mortgage fraud
perpetrators are among the tactics being discussed. There is merit in many
of these concepts, however at BasePoint we see several key points missing
from these discussions. BasePoint supports an approach based on using
fraud and risk management technology to its full advantage, and building
the right organizational structures, as well as the most appropriate policies
and practices to profitably rebuild the market. Let's review these points in
greater detail.

Step 1 � Don't be afraid to use the word "fraud"
Historically, many in the mortgage industry have been hesitant to call
misrepresentations "fraud". If there is suspicion of misrepresentation in a
mortgage application, there is suspicion of fraud by definition. While many
in the industry jokingly referred to stated income loans as "liar loans", there
was little appetite to stop the "lies" from happening over the past several
years. Many stated income loans originated between 2004 and 2006
contained grossly unreasonable incomes but the loans were approved.
Fraud does not have to be proven in a court of law before a loan can be
declined for suspicion of fraud. For the industry to move forward there
needs to be a common recognition that fraud is occurring and that fraud
that was suspected over the last few years, funded anyway and not
disclosed to the investment community, has contributed greatly to the
default problems causing the so called credit crunch today. For all the

3
resetting loans going into default today, it would be interesting to review
how many of them had legitimate incomes and employers listed on the
application of the funded loan.

Step 2 � Understand that all fraud and risk matters
The mortgage industry has been dividing fraud, which has historically been
unmeasured, into two categories: fraud for housing and fraud for profit.
Fraud for housing generally involves borrowers who exaggerate specific
details on their mortgage applications to qualify for their dream homes, and
at the time have every intention of repaying the loan as agreed. Fraud for
profit involves egregious scams to con the lender out of money. Fraud for
housing in many cases was considered by lenders to be somewhat
harmless, while fraud for profit schemes grabbed front page headlines. Is
fraud for housing really as harmless as some lenders seem to think?
80% of
rejected loans The common thread of many fraud for profit schemes is property value.
and fraud due Fraud for profit schemes exploit the lender by leaving them holding a
to early pay property worth less than they loaned to the borrower. Fraud for housing
default were schemes, on the other hand, do not typically exploit the lender but rather
associated fool the lender into thinking a borrower is a better risk than they are. This
with borrower generally means misrepresenting the borrower's income, employment,
misrepresent-
intent to occupy, assets or credit profile. The data suggests that fraud for
tations, while
only 20% of
housing isn't harmless. When BasePoint analyzed a large sample of
fraud schemes investor rejects and early payment default losses, it became apparent that
were most of the cost of fraud was the result of fraud for housing schemes rather
associated than fraud for profit. In fact 80% of rejected loans and fraud due to early
with property pay default were associated with borrower misrepresentations, while only
valuation. 20% of fraud schemes were associated with property valuation.

Mortgage originators need to acknowledge that "fraud for profit" is not the
only mortgage fraud that impacts the bottom line. Loan risk coupled with
false application information provided by borrowers and/or brokers makes
all types of fraud matter. While loan volume was the historical yardstick of
success, we can see where this got lenders into trouble and how loan
quality is also paramount.

Step 3 � Bring underwriting functions back to risk management
The role of underwriting needs to be revisited by originators. During the
mortgage boom, underwriting became too closely aligned with the sales
and production operations. The focus was on closing as many loans as
possible, without adequate regard for risk management. This caused
checks and balances between risk management and sales to either be
absent, or to be weighed too heavily toward sales. With the recent collapse
of the sub-prime industry, lenders should place emphasis on the role of
underwriting to approve good loans while protecting from unacceptable
risk.

In addition to a renewed strategic focus on fraud, early payment default
and underwriting goals, lenders also benefit from examining the processes
they use to evaluate repurchase requests. Lenders should put a vetting
process in place to ensure that repurchase demands are reviewed, and
challenged, if appropriate.

4
Step 4 � Create an organization to measure and track fraud
Building a robust risk and fraud management organization is critical in
rebuilding profitability. Previously within the mortgage industry, most
lenders lacked a formal fraud management team with fraud managers
accountable for understanding, measuring, and controlling each lender's
fraud losses. Consequently, fraud was not well-defined internally and
measurement suffered dramatically. Without this important function, there
is a lack of focus and a diffusion of fraud reduction efforts.

It is recommended that lenders create a formal fraud organizational
structure and appoint or hire a fraud manager that will be responsible for:
quality control, repurchase review and challenge, fraud strategy and
analytics, fraud reporting and management information systems. This team
should establish a formal definition for fraud that is well understood by
management and is measured consistently.

In addition, lenders need a renewed focus on overall credit risk
management. The credit risk management organization must take a
leadership role in setting policies and choosing the right tools and
technologies to ensure that lenders are booking the right loans to achieve
their production goals while protecting the lender's profitability goals.
While the risk management team must work in partnership with the sales
and production teams, they must retain their responsibility to ensure the
lender's profitability objectives are reached.

Step 5 � Accepting that guideline changes alone do not fully protect against
fraud and early payment default
The fear of EPD and foreclosures, as well as changes in what the secondary
market is willing to purchase, has caused some lenders to make sweeping
changes to underwriting policies. Although originators are tightening
standards, mortgage fraud persists.

The following table examines several recent changes many lenders have
made, as well as examples of how the fraud trends shift to try and get past
the new guidelines. In the last column you'll see how BasePoint approaches
the solution to these issues.

New Lending Resulting Fraud BasePoint Approach
Standards Patterns
Stated loans Greater use of forged or Assess relationships
are being manipulated income between income, age,
replaced by full and/or asset professional years/job
or "lite" documentation. years, regardless of
document loans document type.

5
New Lending Resulting Fraud BasePoint Approach
Standards Patterns
100% CLTV Property valuation Assess property
loans are being misrepresentations valuation by
replaced by (e.g. outdated comps, comparing to the
maximum 75% out of area comps, property valuation
CLTV inflated value) distribution
Suspicious down determined using
payments (e.g. consortium data.
unseasoned assets,
questionable HUD
transactions)
Greater use of
incentives (>5%)

Minimum Greater credit Credit scores can
credit score bureau manipulation have an inverse
cut-offs have (e.g. authorized relationship to fraud
risen by 40 - users, disputes, co- BasePoint assesses
60 points borrowers) this risk in the
Straw buyers/Straw predictive scoring
lien holders The BasePoint
Dynamic TRAITS
process assesses
changes in income/
credit score
relationship.

Restrictions on Increased demand for Assess income and
investment "second home" loans. property value,
property loans common areas of
investment fraud.

In short, improved underwriting can help reduce some elements of credit
risk, but not fraud risk. Tightening underwriting guidelines will change
fraud patterns, and fraud will exhibit different characteristics. The
BasePoint solutions contain variables to assess fraud risk from these
different perspectives.

One of the most popular modifications has been increasing credit score
cutoffs. For non-prime programs, this can actually increase fraud and EPD
rates. By simply raising the minimum credit score, many performing loans
will be eliminated from these portfolios. While it may lower the overall
number of EPD loans, the default rate can actually get worse because the
loan volume has been cut drastically. However, many of these higher credit
score borrowers should be able to qualify for a prime product. This is a
risky population. They may actually carry more risk than the segment
that's been eliminated.

6
Curtailing good loan volume through dramatic and overarching policy changes
can exacerbate the profit squeeze in the short-term. Rather than reduce Rather than
origination volumes through underwriting policy changes or loan program reduce
eliminations, lenders should adopt more targeted, data-driven tools to retain origination
the good origination volume, while substantially reducing mortgage fraud and volumes
through
EPD risk. Those tools are technology-based. For example, BasePoint has
underwriting
found that the credit score does not predict early payment default as well as policy changes
an EPD-specific model such as BasePoint EPDTM Alert. For one lender studied or loan
the BasePoint EPD model performed 2� times better than the traditional program
credit score at a 10% review rate, resulting in an additional $10MM in eliminations,
detection of early payment defaults. Given that approximately 27% of credit lenders should
scores fall in the mid 600-699 range, above many lenders' new credit score adopt more
cut-off, EPD-specific predictive models can provide significant lift. targeted,
data-driven
tools to retain
Step 6 � Monitor brokers and appraisers analytically
the good
Broker monitoring is another best practice that should be a standard for all origination
wholesale lenders. BasePoint has extensively studied broker-facilitated fraud. volume, while
Most brokers submit only good loans, but a small segment of the brokers substantially
BasePoint analyzed proved to have a higher than average rate of reducing
mortgage
misrepresentations and early payment defaults. In most cases, lenders fraud and EPD
experience most of their fraud and EPD from a small overall percentage of risk.
brokers, typically 10% or fewer brokers.

The underlying catalyst for broker fraud during the boom appears to have
been the incentives those lenders offered to brokers to submit loans to
them. Broker incentives such as yield spread premiums and rebates from
lenders create an environment where brokers may be tempted to put
unqualified borrowers into homes they cannot afford, or at higher than
feasible mortgage rates. BasePoint's analysis indicated that there was a
correlation between the fees and points that a broker charged, and the
corresponding level of misrepresentation in those loan packages. For
example, if brokers were charging higher fees, there was a higher likelihood
that information in the package could contain some material
misrepresentations that would lead to financial loss for the lender.

Step 7 � Use analytic scores to stop fraud and other elements of risk the
way other parts of the financial services industry do
Mortgage originators can learn from the credit card industry of the early
`90s. Credit card fraud rates were accelerating at a tremendous rate in
1993. Accordingly, the credit card industry invested heavily in technology
to combat fraud and is now experiencing some of the lowest fraud levels in
history. Credit card fraud losses in the US have been reduced about 70%
since their peak in 1993 and 1994. To achieve this dramatic improvement,
credit card issuers adopted pattern recognition technology, which is highly
accurate in detecting fraudulent credit card transactions. Similar
technology has proven equally effective at identifying fraud and EPD risk for
the mortgage industry.

While millions of dollars have been spent in the mortgage industry to
automate loan processing and lead generation, adopting technology to

7
specifically manage fraud and EPD can immediately help prevent future
losses and boost profitability. Most traditional fraud products in the
mortgage industry are based on comparisons between application
information and third-party data sources. Data validation will still have its
role, but it's obvious from the current state of delinquency rates and
defaults that these practices alone are not sufficient. And relying on credit
scores is misleading since the credit score is only effective in predicting risk
when the facts of the application are real. Also, the credit score does not
take into account important characteristics related to the borrower, loan
amount, and loan program. Fortunately, there is a proven predictive
analytic technology designed specifically to curb mortgage fraud and early
payment default. Pattern recognition fraud and EPD scores enable more
timely fraud decisions and more targeted loan reviews.

Predictive analytics use science in conjunction with historical application
and performance data to accurately predict the likelihood of fraud and EPD.
A comprehensive study by BasePoint analyzed over three million loans
originated between 1997 and 2006 and found that if predictive models were
deployed early in the loan process, it would help lenders identify which
loans were most likely to contain fraud and/or have a high risk of early
payment default. This enables originators to review suspect loan
applications using an enhanced fraud review process and to reject
confirmed fraud pre-funding. In fact, BasePoint's predictive models can
correctly identify 40% or more of a lender's loans pre-funding that, if
booked, would result in EPD, by reviewing just 10% of total applications.

This filtering approach helps both lenders and investment banks target high
risk loans and use their limited resources to investigate those loans with the
Another
highest risk. These pattern recognition models are proven effective for
advantage of detecting fraud, identifying high risk or EPD, and evaluating broker risk.
these pattern Another advantage of these pattern recognition models is a very low false
recognition positive ratio. False positives are high scoring applications that contain no
models is a
fraud. They are "false alarms". BasePoint's predictive models demonstrate
very low false
positive ratio. less that 6:1 false positive ratios in production. This means that for every
seven high scoring applications that are reviewed, more than one is shown
to contain fraudulent misrepresentation. This is a significant advantage over
the high false positives triggered by data validation tools. Low false positive
ratios enable lenders to easily incorporate this enhanced fraud detection
process into their production environment without placing an undue burden
on investigative resources. The savings from detecting the fraudulent loan
far outweigh the time to review.

Step 8 � Use a standard process to find the fraud and risk
Once pattern-recognition models are used to identify those loans with the
highest risk of fraud and EPD, lenders and investment banks should be
using an enhanced loan review process to effectively confirm loans that
should not be funded or purchased on the secondary market. The review
process differs for reviewing high risk of fraud as opposed to high risk of
EPD or other credit risk factors. Fraudulent misrepresentation(s) in a loan
or loan application is typically confirmed through a review of the loan file,

8
with occasional use of external data validation. Risk of early payment
default and other credit risk factors requires research into the financial
viability of the applicant as it pertains to the particulars of the loan. The
income stability and accuracy/existence of assets should be confirmed.
Total debt should be reviewed to see if there are obvious expenses missing,
or indications that debt is rising. Analysis of the credit report often
provides a lot of insight into the trend of the debt burden, and alerts you to
indications that the applicant's financial viability might be worsening. Loan
program values such as a loan without impounds, high fees and/or large
cash-out amounts can also be indications of payment risk.

Step 9 � Share data across the industry to improve fraud and risk
management efforts
Finally, contributing data to a cross-industry mortgage consortium is an
emerging best practice. BasePoint pools data across lenders and
investment banks to provide several key advantages:
Access to a breadth of multi-client data to create more robust pattern
recognition models
Enhanced ability to predict mortgage fraud and early payment default
risk due to significant depth of mortgage data
Identification of new trends more quickly by leveraging data across
multiple lenders and investment banks
Ability to determine statistical norms across broader populations and
measure deviations
Ability to exploit links and anomalies

This combination of technology, policies and processes helps originators
and investment banking firms better navigate the new mortgage industry.
Significant reductions in fraud rates and early payment default are
achievable by adopting the recommendations discussed throughout this
paper. Forward-thinking mortgage leaders have already begun to have
success using these methods. The road back to high profitability and
growth isn't always easy, but implementing these ideas can make the
journey shorter and less painful.

9
Summary

In summary, BasePoint Analytics believes the road back to profitability will
require the mortgage industry to adopt the following changes:

1. Call a fraud a fraud and do something about it

2. Realize all fraud matters

3. Change underwriting goal to risk management

4. Create a fraud organization and focus the risk management function

5. Adjust guidelines appropriately, but realize that they will not eliminate
fraud and early payment default

6. Monitor brokers and appraisers systematically

7. Use predictive scores like other industries to prevent fraud and early
payment default

8. Establish a standard fraud and credit risk review process

9. Share data across the industry to catch more fraud and better identify
early payment default risk

While the mortgage industry redefines itself in the coming months,
BasePoint believes a new approach to analyzing fraud and early payment
default risk based on its characteristics will become the standard. It is
sound science to build predictions based on data and the patterns within
that data. By applying science to the problem and making rational
decisions based on data, the industry will undoubtedly make better risk-
based decisions in the future.

10
About BasePoint Analytics
BasePoint Analytics is a leading provider of predictive analytic fraud and risk
management solutions for global banks and the mortgage industry. Using
proprietary pattern-recognition technology, BasePoint quickly identifies
potentially fraudulent or high risk activity, minimizing losses while
accelerating the processing of non-risky records. The company's fraud
experts have deep, real-world domain expertise and have successfully
solved fraud and risk challenges for many high profile lending institutions.
BasePoint works with industry-leading customers in mortgage origination,
investment banking and payment cards. Leveraging a client's existing
technology, BasePoint provides clients with immediate results and quick
return on investment. BasePoint clients achieve dramatic improvements in
fraud and loss detection performance using a predictive analytic approach
rather than traditional methods.

Leading Scientists and Top Industry Consultants
BasePoint's team of renowned scientists is dedicated to building state-of-
the-art predictive models using the latest advanced techniques. Our fraud
and EPD specialists have innovated transactional, application and account-
based models that are detecting fraud and EPD in some of the largest
organizations across the globe. BasePoint is committed to investing heavily
in research and development to provide you with the industry's most
effective defense against fraud.

Having successfully managed fraud and risk operations for dozens of the
world's highest profile organizations, BasePoint's elite team of consultants
has deep domain expertise. Clients can have confidence these professionals
will successfully guide you to significant fraud and risk reduction through the
integration of analytic models, tools, strategy alignment, and operational
best-in-class processes.

A Global Focus
Fraud and risk do not have geographic boundaries and neither does
BasePoint. Our experts have spent years understanding the global nature of
fraud and risk migration, and more than a decade researching fraud, risk
trends and management throughout the world. Whether your organization is
local, national, or spans many continents, we have the expertise and
solutions to help.