Banking Analytics Blog

Regulations Amplify Demands on Model Management

2013-05-23T09:06:32-07:00

Since the global recession, regulators worldwide have become increasingly concerned about the soundness of decision making and capital adequacy within financial services. As a result, they are taking an even closer look at how financial institutions manage risk and use scoring models.

To address this greater regulatory scrutiny of model risk management, financial institutions face many challenges, among them:

Clear guidance. While the Fed/OCC and Basel regulators have issued guidelines on what they expect, they are often just that—guidelines. The onus falls on financial institutions to develop a rigorous model management framework to satisfy audit requirements.

More models. Predictive models are being used on an increasingly broader scale—often in the thousands for large lenders. This makes it difficult to manage and track models, and determine whether they are still performing well. It's also a struggle to provide required documentation and administration.

Transparency. Models are growing in both number and complexity, but they must still be easy to understand, defend and explain. This is important in internal discussions, for regulatory reviews and when explaining to customers why, for instance, a loan was denied.

Bandwidth. Much of the compliance burden falls to IT departments, who in these lean economic times, have had to cut back on staff and resources.

For evidence of the growing focus on model management and governance, check out a new white paper from IDC Financial Insights, "A Framework for Model Governance: Opinions and Insights from Bankers Across North America." In it, IDC shares findings from in-depth interviews with financial institutions across North America and Latin America. These institutions identified regulatory requirements as the number 1 driver behind their model management investments, however many continue to be challenged in establishing effective programs around model risk. IDC also defines the characteristics of model governance "laggards" versus "leaders," and shares several best practices for creating a strong model governance framework.

I invite you to download the IDC report, and join us for a June 4 webinar where Mike Versace, Research Director at IDC Financial Insights, and FICO's own Sharon O'Connor Clarke will discuss findings.

France Takes Unwelcome Spot at Top of Europe’s Fraud Loss Table

2013-05-20T09:05:30-07:00

European countries vie for various sports prizes, but one prize no one wants has just been claimed by France. In terms of the highest card fraud losses for 2012, France is at the top of the league — edging out longtime champion the UK.

As revealed in FICO’s updated map of card fraud in Europe, France now has 29% of the region’s losses, compared to the UK’s 27%. It’s a sour place to be for any country, but particularly for France, which was an early adopter of the chip and PIN strategy that drove down fraud rates across Europe.

As my colleague Martin Warwick notes, despite a rise of 14% last year, UK card fraud losses were still 36% lower in 2012 than at their peak in 2008. By contrast, France’s overall fraud losses grew 65% between 2007 and 2012, which translated to an additional €174 million of card fraud losses over the period.

However, France’s “win” is no cause for celebration in London or other European capitals — unfortunately, nearly all the European countries experienced a rise in fraud in 2012. According to data from Euromonitor International, a provider of global strategic market intelligence, European card fraud losses were 6% higher last year than in 2011. France, Russia and the UK made up more than 80% of this increase. France alone generated nearly half of the total increase in losses, but Russia’s losses are growing the fastest, with 2012 totals reaching three times the level of those reported in 2010.

Compared to most regions, Europe has a sophisticated approach to fighting card fraud. But as these new numbers show, criminals aren’t throwing in the towel. Every card issuer needs to maintain a vigilant approach and use the latest technology to combat the threat — or their country could well wind up at the top of next year’s fraud loss table.

Best Practices for Modeling Regulations

2013-05-16T16:53:31-07:00

Financial institutions have had a difficult time adapting to the latest regulatory guidance regarding model validation and management. But making the right improvements can also translate into better analytic performance and risk management.

To both comply and compete, it's critical to build an organizational policy for comprehensive model and credit policy management. This framework should include the following tried-and-true practices:

Have clearly stated credit policies; review these regularly. We recommend reviewing these every six months since they have a direct impact on your bottom line. In the US, the Fed and OCC require a review of policies at least annually.
Prepare a suitable data sample. Regulators require you demonstrate your model validation sampling techniques are complete, responsible and relevant, since incorrect or inaccurate sampling can impact model performance.
Ensure model segmentation transparency. In general, you’ll need to clearly document how you segmented subpopulations and how this supports business objectives. Automated tools can help ensure transparency of segmentation logic for regulators, while enabling performance improvements.
Choose the right model type. It’s important to not only select a model type that's appropriate for the decision type and available data, but one that enables transparency and palatability to both regulators and customers.
Validate model effectiveness. Revalidate models on an ongoing basis—minimum once a year, but more often in a dynamic economy and/or where account volumes are sufficient to give reliable results.
Track performance. Employ standard reporting and analysis that provide insight into the health of models that drive your critical decisions.
Defend decision strategies. Regulators will ask for empirical evidence to justify your decision strategies, and they’ll want to know your realized gains, losses and exposure. Since decision strategies have become increasingly complex, interactive strategy exploration functionality is essential for tracking strategies, strategy changes and results.
Monitor overrides. Be able to prove that overrides are based on clear and consistent guidelines.
Document thoroughly. Track everything in your model development and monitoring processes. Build a detailed inventory of every model, along with their purposes, usage, restrictions, inputs, performance, updates, owners within the organization and audit history. By automating production and review of standard reports, you'll free up analysts to focus on ad hoc regulator queries and new model developments.

For more details on these best practices, I encourage you to download our Insights white paper, "Comply and Compete: Model Management Best Practices" (#55--registration required).

Credit Line Increase Trends: A Mixed Bag

2013-05-08T10:00:00-07:00

Since the recession of 2008-2009, there has been quite a bit of discussion about credit lines. Are they going up? Are they going down? Are banks still aggressively managing their risk?

Based on the US data I’ve seen, it looks like a mixed bag. On one hand, the number of credit line increases is clearly going up.

As the table above illustrates, banks granted significantly more credit line increases in 2012 than they did at the dawn of the recession in 2008. Despite the predictable dip in line increases during 2010, this data suggests that both lender and consumer appetites for line increases were quite robust in 2012.

However, I did a little more digging and found that not all credit line increases are created equal.

The size of the average increase was roughly 57% lower in 2012 than it had been in 2008. In fact, the average line increase was about 35% lower in 2012 than it was when lenders were supposedly “super cautious” in 2010. In addition, from 2008 to 2012, there was a dramatic decline in the percent of consumers receiving sizeable line increases, which we defined as at or above $7500.

In terms of actual dollars, lenders gave out approximately $10.5 billion in credit line increases in 2008. In 2012, that figure was only $5.3 billion despite hundreds of thousands more consumers receiving line increases.

So…if 300,000 more US consumers received credit line increases in 2012 than in 2008, but the total dollar amount was nearly 50% lower, are bankers tightening or loosening their purse strings?

It’s an interesting question. As the economic recovery rolls along, credit line increases could have a significant impact on how quickly consumer spending rises. This is definitely a topic that I’ll be following closely.

US Bank Priorities for 2013: Big Data, Customer Experience

2013-05-07T08:45:00-07:00

I've been blogging about results from our latest quarterly survey of US bank risk professionals. This quarter, we took the survey in a new direction by also asking about business priorities in 2013. Two related responses tied for the top spot, each with 35% of respondents: utilizing Big Data analytics to gain greater insight into customers and improving the customer experience—both key themes at last week's FICO World 2013 conference.

These responses align closely with the anecdotal evidence I’ve gathered while talking with FICO banking clients, both within and outside the US. Almost without fail, they tell me that enhancing the customer experience is a strategic goal this year, and they increasingly recognize the value of Big Data analytics in executing against these initiatives. We’re helping clients worldwide apply Big Data analytics to everything from marketing and originations to account management and collections. In fact, I share best practices on this very subject in my new Insights white paper: "When Is Big Data the Way to Customer Centricity?" (#67, registration required), which I invite you to download.

You’ll note that strengthening fraud prevention was cited as the top priority by 20% of respondents, and 9% said it was increasing utilization of mobile technology. For those of you with like-minded priorities, stay tuned to this blog where my colleagues and I will continue to share innovations and best practices in both these areas.

Even Risk Managers Put Customer Experience as Priority

2013-05-07T04:02:49-07:00

92% of bankers surveyed at FICO World 2013 last week in Miami said improving the customer experience at their company is a high priority or the top priority. Given that the majority of those attending are in risk management, fraud and collections, this shows how serious banks are about regaining customer trust and loyalty.

Despite this lack of trust, customers may be more willing today to share information with their bank, if there’s a definite benefit. We asked attendees what percentage of their customers would be willing to share more information in exchange for a more personalized experience. More than three-quarters of respondents said they would — here’s the breakdown:

• Approximately 50% - 48% of attendees
• Approximately 75% - 29% of attendees
• Approximately 25% - 17% of attendees
• Close to 100% - 3% of attendees
• Close to 0% - 3% of attendees

The “customer revolution” was one of the conference’s themes, and Friday's general session included a quartet of C-suite bankers from the US, Europe, Latin America and Asia discussing how their customers feel about the banking experience, and what banks are doing to improve it. Here were some of the interesting observations:

“We’ve had a number of own goals, and shot ourselves in the foot with issues like LIBOR and mortgage settlements. Building customer trust is critical.”

“In China, banking trust is high, since the banks are primarily government-run. If you don’t trust the banks, it means you don’t trust the government.”

“If people have a bad service experience, they can tweet about it immediately. If we respond to the tweet immediately, we can create a positive experience.”

“We’ve clustered clients in terms of their banking transactions, and created very specific clusters we can treat differently. These clusters do a good job of reflecting the personality of clients. For example, here’s one segment called Living Large.”

“We have a program to offer university graduates their first cards, with smaller instalments and lower credit lines. If you help these people to get and build their credit, they’re more loyal to the bank.”

When retailers and other companies are mining vast stores of Big Data to personalize the experience, many banks find themselves playing catch-up in a sport they pioneered — targeted, analytics-based marketing and customer management. As our panel showed, different regions have very different challenges, but all of them face competitive pressures that make positive customer engagement a priority.

Secrets of Big Data Success – Live from FICO World 2013

2013-05-02T19:15:20-07:00

At this week’s FICO World conference in Miami, Big Data has taken center stage. Our keynote panellist yesterday, Kenneth Cukier, presented a captivating talk on how Big Data will transform our lives. Based on his new book, Cukier, the data editor of The Economist, emphasized that the secret of succeeding with Big Data goes beyond just having access to the data, or even the analytic technology to make sense of it.

Cukier posited that the mindset is the most important thing that differentiates successful companies — the ability to think differently about data, to see its possibilities. He cited numerous examples, including the company Farecast, which used passenger manifests and other travel records to build a system that helped passengers save money on airline tickets by picking the optimal time to buy.

A similar note was sounded in the panel discussion I hosted yesterday on Big Data: From Vision to Value. When the discussion turned to what makes Big Data applications successful, the consensus was that the most important aspect is executive sponsorship. If you don’t have the buy-in and support from the top, you’ll end up simply mucking about in the weeds.

Like so much else about Big Data, this is not new. Back in 1985, FICO published an article on “14 Secrets of Scoring Success.” The first step was “Obtain Management Support.”

I’d like to thank my fellow panellists for an engaging discussion: Mark Luber, VP, Analytics and Data Acquisition, LexisNexis, Risk Solutions; Mark Warren, Director, Decision Science, Barclays; Juan Huerta, Vice President, Advanced Analytics, Global Decision Management, Citibank; and Rod Nelsestuen, CEB TowerGroup.

FICO World Video: “It’s All About the Customer”

2013-05-02T16:50:00-07:00

At FICO World 2013, we caught up with FICO blogger Joanne Gaskin after she participated in a full day of sessions on mortgage lending, covering topics ranging from best practices to compliance challenges. Here, she shares her thoughts on the conference so far:

For regular news and commentary from FICO World 2013, stay tuned to this blog, check out the FICO Labs Blog for additional videos and views, and get up-to-the minute tweets from @FICONews and @FICOWorld.

FICO World Survey: Bucking Analytic Trends

2013-05-02T15:15:00-07:00

Each day during FICO World 2013, we’re surveying attendees. The theme of yesterday’s survey was “Implementing Analytics”—quite fitting on the day that Kenneth Cukier of The Economist gave a keynote address on the “datafying” of society and the rapid growth of analytics. (View responses from the previous day’s Big Data survey.)

We began the FICO World survey by asking: “How quickly can you change your decision-making strategies and systems to incorporate a new analytic model or a new insight about your customers?”

The most common answers by respondents, who primarily work in banking, were “a month or less” and “up to three months,” both receiving 26% of responses. This was followed by “up to six months” (20%), “up to a year” (13%), “a week or less” (8%) and “more than a year” (6%).

Clearly, those who want to incorporate new analytic insights into decision strategies felt they could do so rather quickly. Some 60% said they could do so in 90 days or less.

Our second survey question was very simple: “Do you plan to use any cloud-based analytic software within the next 12-24 months?” Nearly 1/3 said yes (32.5%), while the rest (67.5%) said no.

We expect the pendulum will shift toward greater adoption of cloud-based analytics—and soon. Indeed among recent analyst projections, industry statistics, expert commentary and media coverage, the general consensus is that cloud computing will grow significantly in the near future. Perhaps we should ask the same question of FICO World 2014 attendees to see if responses change.

For our third and final survey question, we asked: “What is typically the biggest challenge to implementing analytic projects at your company?"

The top answer was “limited bandwidth of our IT team” (48%), followed by “limited bandwidth of our analytic experts” (21%), “cost” (18%), and “concerns about data security and privacy” (13%).

These responses are somewhat surprising. Despite noise in the media and elsewhere about an emerging analytic talent shortage, only about 1 in 5 respondents said the bandwidth of their analytic teams is the main limitation to implementing analytic projects. Limited IT resources appear to be the real culprit.

FICO World Video: Key to Big Data? Start Small

2013-05-02T09:19:39-07:00

At FICO World 2013, we caught up with FICO blogger Andrew Jennings after he moderated the panel discussion “Big Data—From Vision to Value,” which included representatives from Barclays, LexisNexis, Citibank and TowerGroup. Here, he summarizes a few key learnings from the panel: