spacer

A House Divided: The Role of Technology Strategy,
Organization, and Culture in the Subprime Crisis

In the last issue, we saw how automation may have played a role in the sub-prime meltdown. In this issue, we consider how strategy, organization, and culture contributed to the collapse.

In March 2007, the same year he finished EMTM, Raj Rajkotia joined Countrywide Financial, the nation's largest mortgage lender. At the start of 2007, the company had $500 billion in home loans, 900 offices, and $200 billion in assets. Countrywide helped to fuel the largest housing boom in U.S. history by offering loans to high-risk borrowers. That also made it one of the biggest casualties when the mortgage market collapsed. By August, as defaults rose, there was concern that the company might go bankrupt. Its stock price fell 80 percent and it lost $20 billion in market value. But Bank of America announced a plan in August to purchase the company for more than $4 billion. Rajkotia said he found himself living the issues he had studied in Tom Oser's IT Strategy course in EMTM.

Silos and Lack of Integration
Rajkotia, Senior Vice President of Countrywide, is now working with Bank of America colleagues on the post-merger technology integration team. They are focusing not only on integrating the two companies but integrating the mortgage business itself. Rajkotia said the lack of integration between the different parts of the mortgage business contributed to the sub-prime crisis.

"Three strong silos ran in mortgage banking without strong end-to-end integration," he said. The three major parts of the mortgage business are: origination, where the loans to consumers are made; capital markets, where loans are packaged and sold to investors; and loan servicing, since investors do not want to get involved in taking mortgage payments and handling customer calls. At most mortgage companies, business units had responsibility to provide their own IT, so these three organizations and their IT systems were not well integrated. Divided into these separate worlds, no one within the company could see the big picture. Creation of loans was separated from the risks associated with managing those loans over their lifetimes. This led to a lack of any real sense of responsibility for risk management at the time of load origination, promoting the adverse behaviors in lending that contributed to the collapse.

Oser, an IT strategy consultant and adjunct EMTM faculty member who taught Rajkotia's strategy course, points out that these silos can be seen in the organizational charts of companies in many industries. There are senior vice presidents who handle separate areas such as finance, operations, production, HR, IT, marketing, and sales. While these silos are bound together by a shared set of financial statements (balance sheet, income statement, cash flows), they are typically managed by more local management objectives to maximize some metric related to their services to the company. "Then once they have a green light, they go off and complete their function for the cause, focused locally often without regard to consequences upstream and downstream" he said.

The interfaces between these silos are people, processes, and technologies. Because each silo typically decides on its own interfaces, which are designed to maximize their own localized performance metrics, the end-to-end processes "emerge sometimes as an accident of stringing the silos together. In the finest organizations, these processes are well defined, deliberate and well tuned for interdependence, but in the worst cases they are the product of happenstance."

Blind Trust
The rapid development of technology in some areas, and lag or lack of interest in others, contributed to the development of blind spots and disconnects. "The biggest problem with technology is the way the market has grown so fast but origination has had very little communication with capital markets," Rajkotia said. "Brokers and the sales force sold to sub-prime consumers with poor credit, but didn't realize the impact of not being able to sell to the secondary market." Without tighter regulatory oversight and rules in place to anticipate downstream risks, the earlier stages in the value chain could act irresponsibly, either because they were oblivious to the downstream dangers or intentionally ignored them.

"You could now get a loan without even verifying that a person exists, apart from social security number," Rajkotia said. "This was good for consumers because you processed loans more quickly, but the sub-prime crisis may have been caused in part by the creation of these products and automatic processing."

Oser explains that with technology and high throughput, everything depended upon the quality of information and the algorithms that took the place of human judgment. The mortgage-backed instruments sold into capital markets were based upon statistical techniques that mixed loans of different qualities into a portfolio of loans that distribute the risks, creating a securitized instrument that provides a return that is some "average" of the risk of the loans in the portfolio. "But these instruments were put together with no oversight, and so the quality of the instruments was questionable," Oser said. "These instruments were sold virtually blind into electronically traded markets valued in the billions with no real quality tracking possible."

The players had blind trust that somehow the technology of portfolio construction, the rules over rating practices, and the integrity of electronic trading would protect the fiscally responsible banks that used these instruments. But this trust in technology was misplaced. "They assumed the technology cared about the quality of loans and newly defined modes of risk — without being taught," Oser said.

A Familiar Pattern: S&L on Steroids
The subprime crisis illustrates a common pattern in many markets affected by technological and strategic change. "It is a little bit of history repeating itself," Oser said. Information and communications technologies accelerate the natural process of strategic innovation. New technologies, rules, and business models result in new "degrees of freedom." Companies, particularly new entrants, recognize the opportunities created by these changes and exploit them. Technology and automation speed up the cycles. "Distribution of loan products is instantaneous," Oser said.

By the time regulators wake up, "the leveraging effects of technology may have already produced millions of inter-twined transactions — the unwinding of which undermines all at once," Oser said. "Then positions have to collapse and this causes these crises along the way. At the same time, everyone involved in the run-up is reluctant to pull that trigger — assuming case after case that 'the large numbers in the market will absorb this one'. "

We saw this pattern in the savings & loan (S&L) crisis in the 1970s and 1980s. Savings & loans could lend with greater freedom than regulated banks in terms of capital requirements and restrictions on lending practices. They irresponsibly lent beyond their capacity, seeking outrageous returns in the prevailing environment of interest rates upwards of 18 percent. Each step in a rapid sequence of increasingly compromised rules of engagement becomes the new "norm" in the industry. For savings & loans, lenders felt they would never see single digit interest rates again, and this helped to drive overbuilding in housing and commercial real estate. When the redundantly planned and overbuilt real estate could not be leased in an environment of now falling interest rates, S&L's were forced to refinance their deals at dramatically reduced interest rates — wiping out the profit and driving the deals and institutions into default and insolvency.

Regulators, who may have recognized warning signs, did not act until disaster struck. "Everyone wondered: What went wrong?" Oser said. "While they may sense the danger of such behaviors, regulatory bodies have proven time and again that they will act to clean up only after the ships run aground, but not to avert clear signals of disaster looming. So entrepreneurs and speculators seeking large paydays continue unchecked. Now, in this current technology environment, the disasters can form, build and bust before regulation and common sense can even detect the build up, let alone respond."

The "subprime debacle" could be considered the S&L crisis on steroids. "The technology of collateral-backed securities and their rapid trading through electronic markets, and an immaturity of regulatory control in this new area, provided the perfect storm of opportunity and temptation to recreate the S&L crisis. This environment appeared to offer the ability to dissociate unscrupulously originated loans and lending agents from some distant individual events of defaulting loans. Lenders believed correctly that they'd be buried somewhere down the line in poorly documented and reviewed portfolios. They pretended that, once such loans were securitized, that each over-extended loan would be carried by the design of the portfolio of loans. Oser adds, "This only holds if some qualifications are made: the portfolios must be legitimately designed to contain the risk profiles that they advertise; they must be rated properly by a real review process other than taking the originators and portfolio designers' word; and then sold at properly risk-adjusted rates of return into a market that can verify the preceding two qualifications. And these qualifying conditions are what did not exist in the subprime market run-up." The verified portfolios turned out to be fiction.

Players in the subprime market, as with the S&L crisis, also made outrageous assumptions. "When collateralization depends upon rising real estate valuations in the future, in a credit environment characterized by lax attention to rules, lack of common sense and absence of due diligence, what we have are many of the same factors present in the S&L crisis," Oser said. "Except that today's technology accelerated and deepened the effect. This time around, mortgage brokers and securitizing banks ignored any sense of responsibility to rate, track, manage, and make transparent the origination of loans and securitization and sale off related securities. Unscrupulous lenders and securitizers depended upon absence of technology to effectively bury the trail in the depth of the raw numbers of transactions."

Managing the White Spaces
Technology can help bridge the silos and manage across broader mega-processes, Oser said. "Managers need to look at the white spaces on the organizational chart between the boxes and the silos," he said. "It is across the white space, without regard to territories, where the real interactions and processes are defined that execute the business."

of payments to vendors to generating and managing required working capital in the process. Other mega-processes include planning-to-production (forecasting demand to production and delivery of products to inventory or scheduling service to customers) and quote-to-cash (from customer quote to product delivery and cash receipt from the sale). These processes actually represent the opportunities to create core capabilities of the company — a combination of people, processes and technologies that are able to produce value and are hard to imitate.

Instead of focusing on the silos that deliver these processes, Oser said, companies can begin to look more closely at how the pieces fit together, to create an architecture for modularized Service Oriented Architectures (SOA). This approach focuses on the services delivered as end-to-end processes. "Then, all organizational functions are responsible for the end-to-end performance metrics and not only their own budgets within a silo," Oser said.

Technology can facilitate communication across silos. Using an open platform within silos can help make them more accessible to others. Oser said the technology issues that are important to building Service Oriented Architectures include:

  • Data warehousing and data integrity: Having "one clean data store" for the entire organization is critical in making sure everyone is on the same page; this can start strategically with the customer data, then product data, and build throughout the organization;
  • Technology standards: Standards permit new additions by "plugging in" to common interfaces rather than creating new and unique ones between every pair of systems;
  • Web-enabled software: This allows anyone with a PC anywhere in the world to access programs and functions for which they have access authority;
  • Software as a service: Companies need to be able to obtain service capabilities that are configured to the business' needs through the Internet or private networks; using providers outside the organization can save capital and increase flexibility without having to build the staff and invest capital to develop systems internally.

Limits of Culture
While the silos of the loan origination process may have contributed to the sub-prime meltdown, culture may also have played a role in keeping effective technology from being used. Rajkotia points out that many traders of mortgage-backed securities did their calculations in Excel spreadsheets and made decisions by gut. More transparent methods of analysis, open to review, along with a more systematic or automated process, using technology, could have put constraints or controls on the process. But the trading culture was reluctant to impose such automation, particularly for deals that could run $10 million or $50 million.

"From a cultural perspective, Wall Street is still not ready for automating their trading, especially for a single transaction that could make or break many parts of their business," Rajkotia said. The technology was there, but the culture wasn't. "Not using technology in investment banking was one of the causes of this debacle."

Rebuilding an Industry
Perhaps the deepest questions posed by the subprime crisis are not technological or strategic but moral. Why didn't anyone act sooner? "Too much money being made in a marketplace that cannot be readily explained and understood by a large number of people is sure sign of disaster," Oser said. "The question at the center of all technology-enabled moral stories is: Should something be done just because it can be because no regulation or laws are there yet to stop it?"

As regulators debate the right controls, Rajkotia is working long days with his new colleagues at Bank of America to develop systems that will better manage end-to-end business processes. He said many other major mortgage companies are doing the same thing. They are helping rebuild technology and organizational structures designed to create better communication and integration of broader processes.

In this work, his experience at EMTM has been invaluable. In addition to Oser's strategy sessions, Rajkotia said EMTM courses in leadership made him appreciate the politics involved in organizational change, which can be even more significant than the technology. "When you are going through this large M&A activity, there are a lot of political barriers. You have to be politically savvy and a thought leader. That was something EMTM definitely taught us."

 

Read the first installment of “The Mortgage Market Collapse” in the Spring 2008 edition of InSITE.

“Too much money being made in a marketplace that cannot be readily explained and understood by a large number of people is sure sign of disaster. The question at the center of all technology-enabled moral stories is: Should something be done just because it can be because no regulation or laws are there yet to stop it?”

Tom Oser
Founder of Oser & Company
EMTM adjunct faculty member

“Three strong silos ran in mortgage banking without strong end-to-end integration.’”

Raj Rajkotia, EMTM '07
Senior Vice President
Countrywide Financial

“Managers need to look at the white spaces on the organizational chart between the boxes and the silos. It is across the white space, without regard to territories, where the real interactions and processes are defined that execute the business.”

Tom Oser
Founder of Oser & Company
EMTM adjunct faculty member

spacer spacer

© 2008 University of Pennsylvania. All Rights Reserved. | Sitemap