Dodd Frank Stress Test Results

So much has been written in the news about the US Banks that failed the Fed’s recent stress test. While it certainly is not a pleasant experience for those institutions, we need to look for the lessons that can be learned in this situation and discuss ways to lessen the stress of stress tests within boardrooms.

Stress Testing must be an integral part to how the bank views and manages its risk and needs to be considered more as a business-as-usual activity rather than ad hoc.  But is that the case for all financial institutions?  The status quo tells us otherwise.  All but one financial institution passed the Dodd Frank Act Stress Test (DFAST), so it certainly leads us to believe most US banks would be able to weather a protracted hypothetical economic decline.  So good news there; however, without getting carried away and giving the all clear, it should be noted that the median score for the DFAST stress tests was around 8.2%, but the majority of the banks came in with capital numbers below this, with a few hovering just above the 5% requirement.

It’s worth noting, DFAST, as per the requirements, gives banks a pass or fail grade (the pass grade is banks must have a minimum Tier 1  Common Ratio of 5%), but it certainly doesn’t go into the detail the Comprehensive Capital Analysis and Review (CCAR) does, where a bank’s capital projections are placed under the spotlight. 

DFAST and CCAR are here to stay and according to Viral Acharaya, professor of economics at New York University’s Sterns School of Business, the stress tests need to show if the “system as whole can handle an economic shock.”

It would be an understatement to say that Board of Directors across the US find the whole DFAST and CCAR stress tests stressful and who could blame them.  Not only are armies of analysts mobilized to extract, cleanse, process and aggregate the data to deliver the regulatory submissions, the resultant output needs to be digested, scrutinized and summarized in a coherent manner before it leaves the Board of Directors office.  

As my colleague discussed last year, it’s all about the data when it comes to any analytical initiative, especially when it comes to stress testing.  Often banks are constrained by the sheer number of resources required and data challenges to deliver on this requirement, but when completed correctly, there is a lot of value that can be extracted from the data used in a stress test. However, it’s not just the quantitative details that shed insight.  As part of the recent CCAR stress tests, Citigroup hit the mark on the numbers side, but it was in the qualitative details that their CEO said, “We will continue to work closely with the Fed to better understand their concerns so that we can bring our capital planning process in line with their expectations…”  The Fed report outlines they were not comfortable with the bank’s ability to project its stress profit and loss numbers and whether appropriate stress scenarios commensurate of its risk profile were in place, thus placing its overall capital planning process in doubt. 

So what kind of positive steps can be taken to help ease the pressure at the board room when the periodical regulatory stress tests come along?  The recommendations and good practices below may already play a partial role in the stress testing process, but establishing them as an integral part would bring great benefits in unearthing where particular vulnerabilities and pressure points lie.  Furthermore, assuming that a certain level of losses would have to be absorbed, it would assist in identifying the priorities of what parts of the business model can be safeguarded under an extreme crisis.

• Where stress tests are benchmarked against limits and these limits are breached, what does this really mean, does a breach instigate the right type of discussion, i.e. what questions should be asked?
• Although challenging, it is essential that stress tests illuminate the potential secondary and tertiary effects (firm-wide and market), the speed and magnitude of the impact as the stress transitions through a business cycle and how a joint scenario would be managed.
• Stress tests could be more realistic if the bank created virtual crisis scenarios, and assess its decision making process, action plans and overall reaction times.  Some of this may already be part of a wider disaster recovery or contingency planning program, but it would be useful if reputational risk factors were incorporated to get a full view of the potential aggregate losses, for example a well funded subsidiary might face an undeserved liquidity crisis due to the stresses at the parent level. 
• Surviving an extreme event depends on how agile the bank is in taking pre-emptive action.  Interpretation of early key warning indicators need to be viewed in a broader context so as to avoid any nasty surprises, plus will help to ensure a range of options are available to help steer the bank away from an impending crisis.  
• At the board level, calibration of stress tests in terms of applicability and plausibility are pivotal in steering ‘response’ and ‘contingency planning’ discussions – ideally a clear game-plan would be defined and documented on what defensive action (if any) is achievable during a potential stress. 

In light of the above and once the team is on board and understands what’s expected, how do you build the foundation to pass the stress tests? Having the right technology provides institutions with real-time visibility, flexible analysis, and a reporting capability to leverage test results in strategy development and long-term planning efforts requires key elements to ensure its effectiveness:

• Unified data model: A unified data model supports the convergence of risk and finance across the enterprise. To truly benefit from stress tests and incorporate them into operational strategies, banks must be able to translate stress test scenario results into Balance Sheet, P&L, Capital and Liquidity impact. As banks also seek the ability to better analyze and project risk-adjusted performance, they require a data model – purpose-built for the financial industry – that supports both risk management and finance operations
• Comprehensive risk and shocks library: A comprehensive and centralized library of risk variables and shocks should encompass all variables that can have an impact on the bank’s financial performance and risk profile.  These should include internal risk variables, such as loan prepayments, delinquencies, PDs as well as external factors such as Inflation, GDP, Unemployment, Oil Prices, CDS Spreads, FX & Interest Rate Volatility, Bond Prices etc. When housed in a centralized library, banks can apply these variables consistently in the scenario building process and assess the impact across the institution, be it at the group, subsidiary, line of business, country, or product level.   
• Scenario automation capabilities: Banks must be able to quickly create complete and varied scenarios and store them for replication and future use across the enterprise. This centralized library approach helps ensure consistent scenario application across a bank’s various risk environments. It also enables institutions to respond quickly to increasingly frequent regulatory requests. Scenario automation should also support reverse stress testing capabilities, which model risk, economic, and performance scenarios that would result in insufficient capital or liquidity reserves, i.e. where the business is no longer viable as a going concern.
• Flexible analytical capabilities: This capability provides executive staff and line of business users the ability to drill down into the numbers to better understand them and their impact across the business. This data must allow personnel at all levels to conduct multiple Monte Carlo simulations, which generate large numbers of reverse stress tests so that the most realistic ones can be selected as a bank specific down-case scenario (Aite Group LLC. “The Global Stress-Test Automation Market: Stress, Uncertainty, and Moral Hazard,” September 2013).
• Comprehensive and automated reporting: Analysis must provide leadership with information in a usable format. Traditionally banks have hundreds of people working to aggregate stress testing data for reports, creating a tremendous operational burden and slowing processes. Banks seek a risk management environment that enables them not only to process huge volumes of data rapidly, but also to quickly develop usable reports and analyze the findings to satisfy regulatory requirements and, as important, promote stability, and preserve profitability. 

With stress tests now becoming a bellwether of overall financial health, banks realize that any weaknesses in their stress testing processes and the resultant output will be open to scrutiny to the wider public.  The potential damage of this transparency is real, to both the bank itself and its shareholders. With that in mind, most banks have embarked on strengthening their stress testing capabilities and embedding the processes across the business, to ultimately assist the Boardroom to make better decisions and be better prepared for the next crisis that comes along, which we all know is not a matter of if but when.

As always, I look forward to reading your thoughts and comments.