The use of predictive analytics is
becoming standard practice for credit issuers as they return to
expansion following the banking crisis. While things are on the up
for lenders, they are operating in a very different landscape and
require new tools, writes Charles Davis
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
US credit card issuers never stopped issuing
credit cards, even in the teeth of the Great Recession but they did
slow down, and they certainly ramped up lending standards
significantly.
With the economy improving, albeit slightly,
issuers are returning to expansion mode but with a new risk
dynamic. Consumers are paying their credit card bills on time again
– five of the top six card issuers said rates at which customers
defaulted on their accounts fell in June – enabling issuers to
reinvest and innovate.
Much of that innovation is focused on ensuring
that issuers make good decisions. And demand for risk assessment
technology, predictive analytics, data mining and risk scoring
tools is increasing.
Chisoo Lyons, vice president at FICO, the
credit scoring and risk management titan, said that the economic
downturn has heavily constrained the creativity of risk and
profitability management.
US Tariffs are shifting - will you react or anticipate?
Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.
By GlobalData“Credit card profitability is deeply
constrained by the risk environment, and issuers are spending more
time explaining to other people what they are doing on compliance,
rather than to use analytics to get themselves out of the rabbit
hole,” she says.
“There is a real competitive opportunity in
analytics, but people are so bogged down with the day-to-day.”
Exacerbating the usual challenges is the
uncertainty of far-reaching regulatory changes under Dodd-Frank and
the CARD Act, Lyons says. This has many seeking balance between
their business needs and a nearly paralyzing caution. The new
regulations also place a new premium on extending the proper amount
of credit as a card is issued, because the CARD Act requires
issuers to receive written requests before they can increase a
card’s credit line.
Shifting scores
Though the economy is stabilising, the
fluidity of consumer finance these days is still reason enough for
renewed focus on credit risk management. FICO recently conducted a
study of a bankcard issuer, taking a proxy of the US population
that found significant score movement over time. The study showed
that 27% of consumers scores change plus or minus 20 points over a
three-month period; almost 40% of consumers see a score shift of
plus or minus 20 points over a six-month time period.
For issuers, this means that even the best
customers may not be risk-averse, warranting stricter account
management processes to help identify early delinquency issues
before they become unrecoverable. And it demonstrates that risk
management no longer is a matter of tracking high-risk accounts, as
the fluidity of credit scores these days means that today’s stable
account can quickly mushroom into an account worth watching
closely.
“Everything about risk management revolves
around historical models, and this does not account for these new
defaulters,” says Deron Weston, a principal in the Financial
Services Banking practice of Deloitte Consulting LLP.
“How do we assess risk in this scenario? Can
we incorporate variables beyond the credit history, like real
estate data, the type of mortgage they hold, their occupation, that
differ from region to region?”
On the heels of a devastating financial crisis
and ever mindful of examiners’ increased scrutiny of risk, issuers
largely remain reticent about pursuing new opportunities that
ultimately could make for a stronger and more profitable portfolio.
Issuers are deepening their investments in credit-risk management
tools. Instead of just considering a customer’s propensity to pay
or credit history, issuers can use the technology to gauge how
external events might affect a loan.
Positive trend
It seems to be working: the latest credit
trend report from Equifax showed that banks are issuing more credit
to consumers as repayment behavior improves. Equifax’s recent
credit trend revealed an increase in the amount of available US
consumer credit. New lines of credit for auto, bankcard, consumer
finance, and home equity revolving lines have risen from $145bn to
$167bn from March 2010 to March 2011. Total consumer debt fell from
$12.4trn in October 2008 to $11.4trn – an 8.7% drop.
“Despite concerns of the economy relapsing,
several current metrics indicate the credit cycle is stabilizing –
even growing somewhat as consumer payment behavior improves,” says
Michael Koukonas, Equifax’s senior vice president of client
services.
In May 2011, the average Equifax Risk Score
has risen to 695. This score predicts the likelihood of a serious
delinquency in the two years after scoring. The Equifax Risk Score
ranges from 280 to 850 where the odds of becoming severely
delinquent doubles every 33 points – even odds of severe
delinquency at 590.
Card issuers approved 1.8m more credit cards
in May, while consumer finance credit limits experienced a slight
increase of 1.5%. Equifax attributes the increase in new card
credit to the increase of new cards and lending competition –
issuers are racing to get as many reasonably issued cards in the
hands of consumers as possible as they fear the new regulatory
environment will again shrink the pool of potential cardholders
worth their trouble.
“We are starting to see a shift toward
growth,” says Lyons. “It’s extremely cautious, and it is slow
moving, but issuers are realising that in the new environment in
which they operate, growth is an imperative.”
The regulatory environment is another driver
for credit risk management. Bank regulators want more details on
risk management procedures. The New York Stock Exchange now
requires listing firms to share risk assessment policies, and
Standard & Poor’s factors risk management capabilities into its
credit ratings. In December, the Securities and Exchange Commission
began requiring disclosures on a board’s involvement in risk
oversight.
Regulators are leaning on risk officers more
too. Pushed by the Federal Reserve’s proposed guidelines on
incentive compensation practices, they have asked some risk
officers to “opine” on the plans at their banks.
CARD Act implications
The recent passage of the CARD Act has
affected issuers’ credit risk management in two ways. First, the
new regulations lowered profit potential, forcing lenders to shift
business strategies to stay profitable. Second, with rising
interest rates at traditional card issuers, there is more
opportunity than in years past for niche players to offer cards
with more attractive terms than the big issuers – if they can
manage the risk. An increase in the number of cards in circulation
is generally positive, but issuers must also be mindful to select
customers who, despite various income levels, represent the lowest
levels of credit risk.
New technologies, no longer the purview of
only large lending institutions, can provide a critical difference
to issuers caught between hyper-caution and long-term risk.
It is dexterity that’s needed these days.
FICO’s post-crisis research yielded three major lessons about
credit risk management:
- Risk is dynamic, and all elements of risk management must be
treated dynamically. - Rapid and significant changes in economic forces and market
forces can render traditional risk management approaches less
reliable. - Credit providers need better economic forecasting relative to
risk management for loan origination and portfolio management.
While most scoring models are able to rank
order consumer risk accurately during turbulent times, there is
empirical evidence that economic upheaval can have a significant
impact on default rates even when credit scores stay the same. In
other words, immediate past default experience may be a poor
indicator of future payment performance when economic conditions
deteriorate rapidly.
For example, in 2005 and 2006 a 2% default
rate was associated with a FICO score of 650 to 660. By 2007, a 2%
default rate was associated with a score of about 710 as rapidly
worsening economic conditions (and the impact of prior weak
underwriting standards) affected loan performance.
Weston says that Deloitte’s research shows a
similar fluidity to creditworthiness.
“Our research showed us that the economic
crisis swallowed individuals that had nothing in common with
at-risk people in the past,” he says. “Somewhere between 11 and 12%
the defaulters we studies had never even had a blip in their credit
before the bottom simply fell out. That is a very different
pattern.”
The bigger picture
Risk estimates today should incorporate the
potential future impact of changes in key economic indicators
(unemployment, gross domestic product, housing prices, per-capita
income) on credit risk. Such economic forecasts can augment
credit-risk prediction in two ways, FICO says.
First, it can be used to improve predictions
for payment performance associated with any given score. These
improved predictions can be incorporated into individual lending
decisions and be used at the aggregate level to predict portfolio
performance. Second, economic data can be used to predict the
migration of assets between risk grades.
The ultimate benefit of adding this type of
forecasting capability into credit-risk management is that lenders
will have a better ability to limit losses by tightening credit
policies sooner and targeting appropriate customer segments more
precisely.
Issuers must strike a workable balance between
risk-assessment tools and quality underwriting. With higher default
rates, that means relying on more precise risk-assessment tools to
understand the risk represented by a given borrower. With
widespread adoption, new analytics, updated and tuned to changing
environments, issuers are able to more accurately pinpoint less
risky customers at the time of origination. The bottom line, Lyons
says, is that improved risk assessment through newly developed
scores can help lenders approve more loans without increasing
risk.
Analytics also fuels early intervention. Given
the volatility of unemployment and other economic shifts, it’s more
important than ever for issuers to continually assess risk across
their accounts. By pulling credit scores more frequently,
segmenting member populations for special treatments and running
standard risk-analysis reports, credit unions will be able to
provide early intervention and assistance for members exhibiting
signs of early delinquency.
Lyons says that FICO is in the launch stage of
its new Bankcard Growth Solution suite, which includes what the
company calls the Analytic Learning Hub – a data mart that compiles
all the information from across product lines into one analytic
engine.
“The Hub gathers everything necessary to
generate campaigns that meet risk levels while coordinating the
various functions of issuance, from marketing to customer
acquisition to account management,” she says.
“The tracking of the customer begins even
before onboarding, and then the account is part of a constant
feedback loop that improves analysis of the overall portfolio.”
Analytic learning loops
Lyons says that in such dynamic markets as
card issuance these days, there is a direct relationship between
profitable portfolio growth and an issuer’s ability to quickly
understand and adapt to changing consumer behavior. Rapid risk
management can be achieved by creating what FICO calls “analytic
learning loops.”
These loops are created through the analytic
leaning hub, and comprised of data marts continually refreshed from
internal and external data sources, then threaded through both
acquisition and origination, with decision-making strategies built
in by the issuer.
This approach provides constant learning based
on what consumers are doing that day – not 90 days in the past.
The hub cuts across product silos, allowing
managers from marketing, lending and risk management to hone card
offerings in real time. For example, the marketing department
launches a new card product, and while they do that, they give the
risk management people a profile of the customer they are after
demographically. The risk management analysts create a risk profile
based on the desired outcome and desired response rate. The risk
department knows what they are looking for, and the marketing
department agrees. If the response rate comes in a bit low, the
risk department looks and adjusts accordingly.
The data generated from the campaign, in turn,
can be used to cross-sell other products to selected consumers.
“We can use the Analytic Learning Hub to
extend to other products in the bank. If you are mining all that
card data to extend to lending in other areas, then you grow that –
it’s a customer profitability management solution, really,” Lyons
says.
Lenders now have the analytic tools to enable
safer, measured growth while simultaneously preparing for a
lingering recession. The use of such forward-looking analytic tools
is fast becoming a best practice in risk management. With risk
predictions better aligned to current and future economic
conditions, lenders can adjust more quickly to a dynamic market and
steer their portfolios through uncertain times.
