Analytical maestros with a mastery of mathematical models and how they weave through financial crime compliance programs are finding their specialized skill set in high demand, even more than other compliance jobs in an overall frothy market, say analysts and recruiters in the field.
The current demand for “model risk validation” experts arises from a confluence of several events. In April 2011, the U.S. Treasury’s Office of the Comptroller of the Currency (OCC) and Federal Reserve issued guidance on the expectations for validating risk models and the need for independence in their review. Recent years have also seen tougher regulatory expectations and high-profile enforcement actions highlighting compliance failures partially tied to model missteps.
As a result, banks in recent years have scrambled to find individuals who understand anti-money laundering (AML) programs and validation procedures for related models – for example, those tied to transaction monitoring and customer risk assessments – pushing demand for the creation of separate, more independent AML validation teams.
“Federal bank examiners are more worried about models across the board, including those used in AML,” said Ralph Dahm, president of Wheaton, Il.-based recruiting company I.T. Audit Search. This has increased the demand for individuals that understand both AML and model risk management, and can validate those processes, Dahm says. “Banks also seem to be looking outside more than internally,” such as hiring outsiders to be part of a full-time model review team or contractors.
Compliance more broadly has already become a priority issue for large banks. Earlier this week, Citigroup Co-President Jamie Forese stated that compliance costs could soar to as high as $10 billion annually for financial institutions, noting as well that the bank has made “significant investments” in AML and know-your-customer programs. The New York-based bank has 30,000 employees working in control functions, Forese said in a prepared statement.
At larger banks, OCC’s ‘CRAD’ comes calling
Central to internal controls, however, is uncovering weaknesses in complex modeling stratagems earlier, rather than later. Such a philosophy is behind the OCC’s decision in 2011 to shift resources in its Compliance Risk Analysis Division (CRAD) from analyzing the workings and testing of models for credit risk and other areas to that of AML models.
Banks large and small routinely use analytical models for a broad range of activities, including underwriting credit; valuing exposures, instruments, and positions; measuring risk; managing and safeguarding client assets; and determining capital and reserve adequacy, according to the OCC.
Along with major institutions, the OCC, which examines the nation’s largest domestic and foreign banks, bolstered its recruiting processes for model risk management roles. The agency has been hiring PhD economists with expertise in quantitative statistical and mathematical modeling within the financial services sector, according to examiners and compliance officers.
For the OCC, a “model” consists of three main components: an information input component, such as data, judgments or assumptions; a processing component, which applies mathematics to transform the inputs into a quantitative estimate; and a reporting component that translates the estimate into business information that factors into broader decision-making and backend and downstream processes.
An Ernst and Young (EY) report on AML risk model validation released last year concluded that after the joint guidance on model validation – guidance that never specifically mentioned the application to AML – regulators in subsequent exam cycles have been “placing a much greater emphasis on statistically valid processes and methods.”
Moreover, in many of the largest enforcement actions over the past two years, examiners have included specific warnings citing deficiencies tied to AML models, “challenging whether banks are using statistically valid processes, controls and validation techniques,” EY stated in the report.
Model risk management duties creep into other job roles
The validation procedures are crucial for AML models because they can ensure transaction monitoring and screening systems are “picking up the right names, kicking out the right information and helping the institution eliminate false positives,” said Bob Pasley, a former OCC assistant director of compliance and enforcement.
Apart from the interagency validation guidance, examiners are realizing it’s “not enough to just hire a good BSA officer, have good training and good audit,” said Pasley, now an independent AML consultant in Alexandria, Virginia. “The bank has much greater reliance in recent years on IT models as well and the AML officer has to rely on what the computer spits out.”
That has generated a significant demand, for “experts in this particular area that know the complexities and nuances of AML and the related models,” he said. “One thing is to buy software off the shelf, but that is not enough. Each vendor has to tinker with the model to fit the needs and profile of the particular [institution]” and follow that up with later validation to uncover quirks or not readily apparent bugs.
To better keep systems free of problems, some institutions have decided to build their own, separate teams to have on premises.
On Sept. 24, Morgan Stanley stated the company was looking for an AML model validation associate in New York to be part of the “newly formed” AML Model Validation Team in its operational risk department. The institution, “recognizing the importance of meeting heightened regulatory scrutiny and enhanced rigor in the firm’s AML efforts…is creating a new team to oversee the Model Validation of AML surveillance efforts.”
The job responsibilities include performing independent validations of AML models across various business lines and communicating those findings and results to the compliance team.
A scanning of job postings suggests others are hiring as well. In one job posting released last month for Analytic Recruiting Inc., a firm is seeking a full time position in New York for an AML model development and validation director. The candidate is expected to have a Ph. D. in a quantitative discipline and five or more years of practical experience in the area.
As well, in a position Citibank posted this week for a New York director of AML and sanctions, the institution stated the person must also oversee “model validation initiatives to support the CITI AML platform” and building a quality assurance process for transaction monitoring and enhanced due diligence processes.
Institutions seek both in-house experts and independent validation
Examiners are “going line by line and asking the bank is this the model we want,” said Jeff Sklar, managing director of SHC Consulting Group in Bellmore, NY. “They also realize one size does not fit all. Even if the model is good for the bank, it still may not be acceptable because the institution didn’t customize it to the particular risks of the operation.”
Moreover, if the model works and is customized, regulators can still fault the bank for not having the proper “independence” because someone in-house in AML checked the work, Sklar said. He noted that banks are trying to be creative with the solution, in some cases using outside consultants or creating separate teams that possess AML and model knowledge, but are insulated from compliance reporting structures.
The challenge is that seasoned statisticians who also have a deep understanding of AML program goals and gaps are “rare, there are not a lot of people around,” Sklar said. And with regulators more ready to dissect systems to ensure they are properly processing and spitting out the right data, the particular skill set is “more in demand.”