IFRS 9: Accounting and risk must align for success

This article was written by Gregory van Droogenbroeck, CFA and  Scott J. L. Coulter, CFA, CPA, CA for World Accounting Report. 

The International Financial Reporting Standard 9, or IFRS 9, represents more than just an iteration of the standard accounting guidelines for evaluating the treatment of financial investments. The new rules evolve financial accounting from a function that merely looks at the present. It means accounting professionals will also have to project the future.

Introduced by the International Accounting Standards Board (IASB) in 2014, IFRS 9 is a response to the financial crisis. When IFRS 9 takes effect on January 1, 2018, it will apply to practically all public companies with international operations and affect not only financial reporting, but also treasury, credit risk management and capital planning functions. It represents one of the most seismic shifts in the history of accounting because of its broad reach and technical requirements. Instead of reporting default losses only when incurred, accountants will now be required to take a forward-looking approach and work with risk professionals to build detailed risk models.

Accounting practitioners will now need to understand how risk models can impact the balance sheet. On the flipside, risks professionals will need to calibrate their models to produce stable data sets for reporting an accurate financial picture of the organization.

Rethinking how an organization approaches the measurement of risk and accounting is not an easy task. Risk professionals and accounting professionals often have two completely different frameworks with which to understand the world. Risk always looks ahead and measures the probability of a myriad of outcomes, while accounting is focused on calculating one, stable picture of the truth.

While many large banks have already implemented sophisticated credit risk management functions and models for compliance with their Basel requirements, small banks and other corporations might lack the necessary technical infrastructure, as well as the breadth and depth of historical data, required for IFRS 9 compliance.

Understanding risk models, such as Expected Credit Loss, or ECL, can help bring teams into alignment. As an accountant, here’s what you need to know about calculating and forecasting default risk.

Forecasting default risk

One of the main tenets of IFRS 9 is a new “Impairment” or “Expected Credit Loss” (ECL) model for financial instruments. This model requires the estimation of expected losses due to credit loss or default and requires firms to book provisions for these estimated losses.

This new model is in stark contrast to the previous guidance, which required losses to be recognized only as they occurred (i.e., no estimating forward-looking potential losses). The new ECL model requires companies to utilize the Probabilities of Default (PDs), Loss Given Default (LGD) and Exposure at Default (EAD) to estimate expected credit loss amounts for each financial instrument on their balance sheets (see Figure 1 below for a simplified ECL formula).

ECL debt instrument  = PD issuer x LGD debt instrument x EAD debt instrument

Estimating the future condition of any entity is always a challenging proposition, requiring both mathematical models (with a stochastic or random component) and extrapolation of historical trends (which has obvious limitations).

Existing PD models used by risk departments at many banks are often limited in terms of time horizon of default, typically to just one year. This makes it difficult to extrapolate the model for different time horizons, particularly since banks sometimes do not have access to the long-term data required to develop longer horizon default probabilities.

Consider the example of a 100-year bond. How can a company predict, with any certainty, the probability of an instrument defaulting in 100 years? Defaults are rare and binary events, so the lack of historical default data makes this exercise more akin to guesswork than a sophisticated calculation. This problem is especially acute for issuers with little to no historical default data, such as global banks, sovereign ties, supranational funds and state-owned corporations. To determine lifetime PDs, accounting professionals must combine default data, where available, with term structure models, since all PDs are effectively forward-looking and based on the probability of defaulting in a future time “t” conditional on surviving up to time “t.”

Loss Given Default (LGD) models offer another technical challenge, particularly around acquiring default data and identifying the pricing sources for the defaulted securities. Unfortunately, this information can be very difficult to find. There is still a lack of a detailed pricing history for fixed income instruments, and piecing together that history from other data sources can be very time and cost-intensive.

Before IFRS 9, credit risk models largely focused on stressed and downside LGD calculations. But with the new rules, ECL calculations must now clearly calculate default data under “business as usual” circumstances.

Complying with IFRS 9 requires the ability to make a laundry list of calculations, including estimating the probability of default, recovery rates following default, and the variation of these calculations under multiple scenarios. The ability to estimate all these quantities accurately requires a combination of technology, databases, and quantitative modeling capability.

A potential solution is to work with a data provider, such as Bloomberg, that has access to a diverse range of data sets that can be fed directly into internal systems. Rather than pulling this data manually, companies can leverage the full breadth and depth of default data available in Bloomberg’s systems, along with its market-leading hybrid Merton distance-to-default models, to perform quick and accurate calculations.

Staging and scenario weighting

One of the biggest differences between IFRS 9 and IAS 39 is around provisioning, specifically developing ECL models that would allow banks to track and measure financial instruments at different stages in their credit lifecycle. ECL amounts are calculated on a 12-month or lifetime basis determined by a staging exercise that is dependent on whether there has been a “significant increase in credit risk” of the issuer since initial recognition. Under IFRS 9, stage 1 exposure requires a 12-month forward looking ECL calculation, while stage 2 requires a lifetime ECL calculation and stage 3 effectively is the impairment of the asset up until it requires a write-down to the expected recovery amount.

Further, IFRS 9 requires that ECL calculations reflect “an unbiased and probability-weighted amount that is determined by evaluating a range of possible outcomes.” This presents a variety of challenges, including:

• Determining which factors to stress the calculations on
• Predicting different potential economic scenarios
• Evaluating how many scenarios are necessary to develop a robust credit risk model
• Performing Monte Carlo simulations to model the probability of different outcomes

Any of these potential variations could significantly impact the ECL calculation under IFRS 9, especially since PD, LGD and EAD are all based on forward-looking economic conditions. Therefore, if an economic risk driver shifts up or down by a significant level, then it will have an impact on the ECL calculation which will also flow through the income statement, introducing P&L volatility.

The level of detailed analysis required to incorporate macroeconomic factors, scenario weighting and staging analyses can be overwhelming. That’s why it’s important for companies to review their financial accounting capabilities and identify ways to improve the accuracy and speed of various calculations.

Bloomberg’s IFRS 9 ECL product offers a one-stop shop for making these kinds of calculations. For example, Bloomberg’s IFRS 9 incorporates a client-specific staging analysis which incorporates the default probabilities calculated by Bloomberg’s DRSK default model. The more information a company can access, the better these ECL models will perform.

The weight of onerous reporting

Companies looking to comply with IFRS 9 may feel overwhelmed by the onerous requirements for reporting credit risk. One of the big challenges for CFOs is simultaneously understanding the movements of different financial instruments between stages, and then explaining that movement in provisioning levels to investors. This requires up-to-date reports and detailed data on a level that most accounting departments lack. But by integrating the risk, accounting and reporting functions into the ECL workflow, companies can both save time and produce better, more accurate results.

With IFRS 9, data from credit risk models must also flow into a company’s financial statements and are now subject to audit and scrutiny by investors. This represents a major hurdle for smaller banks and other financial institutions, which must find a way to develop a cross-functional workflow that allows the risk function and the accounting function to communicate with each other.

Credit events can happen at any time, and banks need a way to perform these analyses in real-time with up-to-date data. A robust credit risk service such as Bloomberg’s IFRS 9 ECL module provides accounting professionals with access to all the historical data necessary to make a daily calculation for whichever financial instrument is on a company’s balance sheet. This allows the CFO to have real-time insight into the company’s P&L, and it also protects banks in the event of another financial crisis.

The future of financial accounting

IFRS 9 represents a fundamental shift in the way that companies account for credit risk. While it is designed as a correction for the mechanism that made the credit crisis possible, it is also an attempt to unite the accounting and risk functions in a way that is mutually beneficial for individual companies and the system as a whole.

While some companies, such as large banks and other financial institutions, have existing credit risk models, many businesses still face a major hurdle in getting compliant. That’s because IFRS 9 introduces calculational and quantitative modeling complexity into financial accounting in a manner not encountered before. This may signal the beginning of a longer-term confluence between accounting reality and economic reality (in the same sense that book value accounting has moved towards mark-to-market in some areas), as well as a confluence between risk management calculations (regulatory capital charge) and accounting calculations (IFRS 9 impairment charge).

This confluence will only receive more attention as new regulations, such as Basel IV, come into play, which will likely exacerbate the complexity of managing the differences between a capital-based approach and an accounting-based approach.

The future of financial accounting points to an increasingly data and risk driven approach. Companies that want to thrive in this new world need a technology solution that can give them better and faster access to all the data that drives their business.