Remember the panic around Y2K? Jay Fitzhugh, Chief Regulatory Officer at Mitratech, discusses why we are facing down a similar issue with the concept of negative interest.
Negative interest has become quite the buzzword in financial circles, but what exactly is it? A negative interest environment occurs when the nominal interest rate drops below zero percent for a specific economic zone.
This effectively means that banks and other financial firms have to pay to keep their excess reserves stored at the central bank, rather than receiving positive interest income. Think of your bank account not as a location that pays you for money on deposit, but that the bank is a business that you pay to keep your money safe and secure.
Now, elevate this analogy from your personal account to the accounts that financial institutions maintain with central banking entities like the Bank of England or the U.S. Federal Reserve Banks. The concept is not new, and in certain regions or countries, like Europe and Japan, negative interest has been the reality for central bank interest rates starting as early as 2014.
Why Negative Interest Matters
As a manager of risk within an organization, your role is to forecast and lay out plans to prepare for foreseeable events and their consequences. This is exactly what happened in the 1990s when risk managers realized that many (if not most) programmatically controlled systems that had a 20th century date embedded in their code needed to be validated and modified for the 21st century – the Y2K threat, as it was popularly known.
Fortunately, risk managers raised the red flag, and millions of coding changes were performed to alleviate the threat of wide-scale program failures in everything from banking systems to elevators.
Fast forward to 2020, and risk managers may be on the cusp of another foreseeable impact event: the worldwide adoption of negative interest rates. First, we must ask ourselves, is this even a future possibility? And if there will be a worldwide adoption of negative interest, how would you prepare for that eventuality now?
The first question can be addressed by considering which of your processes or systems are subject to interest rate calculations that currently don’t properly address the reversal of interest rates from positive to negative. While not a problem like Y2K, where date logic permeated all industries and systems, interest rate logic is the foundational basis of banking and invoicing worldwide.
Are the banking and accounts receivable core systems ready? It’s very likely, in many cases. The good news is that most banking and receivables solution providers have interest routines that are well-defined, audited and documented. In other words, they already know where to look to find the adjustments that have to be made.
Model Risk Management: What Questions Should be Asked?
With model risk, per the Federal Deposit Insurance Company (FDIC),
“the term model refers to a quantitative method, system or approach that applies statistical, economic, financial or mathematical theories, techniques and assumptions to process input data into quantitative estimates. A model consists of three components: an information input component, which delivers assumptions and data to the model; a processing component, which transforms inputs into estimates; and a reporting component, which translates the estimates into useful business information. Models meeting this definition might be used for analyzing business strategies, informing business decisions, identifying and measuring risks, valuing exposures, instruments or positions, conducting stress testing, assessing adequacy of capital, managing client assets, measuring compliance with internal limits, maintaining the formal control apparatus of the bank or meeting financial or regulatory reporting requirements and issuing public disclosures.”
As defined by the FDIC, model risk management has significant implications to the analysis and decisions in running a financial institution. However, model risk can in some areas almost be a dark art form in the banking world. For instance, in 2007, the risk models used to predict the safety of mortgage-backed securities turned out to be incorrect due to changing market conditions.
These inaccurate assessments led to one of the most disastrous global financial events: the mortgage crisis of 2007-2008 that kicked off the Great Recession. To get a more entertaining explanation of how this happened, you might want to watch the 2015 film “The Big Short.”
Risk managers should have been asking, what models do we use that would be inaccurate if central bank interest rates became negative? The second question would then be how do I gain visibility into the hundreds of data feeds and their encapsulated calculations that feed model risk as used to make strategic and operational business decisions? This will undoubtedly require careful evaluation and review, just as was applied at the turn of the century with Y2K to ensure that models in use have the ability to invert from positive to negative interest without altering the fundamental calculation validity.
The evaluation of the potential locations where interest rate calculations are in use within the context of model risk management is a challenge. Unlike the well-defined and often evaluated and audited interest routines used in core bank processing applications or in accounts receivable solutions, model risk is the summation of inputs from many sources. These many sources can be distributed throughout an organization, many times outside the view of traditional IT oversight. Sometimes this hidden information management is labeled as “Shadow IT,” or more recently “end-user computing” or EUC. Organizations facing this challenge of discovery of EUC will need a strong grasp on the inventory of interest rate calculations being performed in the many spreadsheets and databases that exist in every organization that have a resulting impact on the models being employed across the multiple disciplines defined in the above FDIC definition. There will be situations where this is a many-to-one relationship of model inputs to a decision model and, conversely, situations where one model input has a relationship to multiple decision models.
The impacts of not asking these kinds of questions this time around could, just as in 2008, be unexpectedly and unfortunately big. Quite possibly very big.