Historical data is losing its value. The question then arises as to how banks can still rely on forward-looking scenarios to future-proof models for so-called non-maturing deposits (NMDs)?
In an attempt to combat the high inflation levels of recent years, the European Central Bank announced one interest rate hike after another.
These increases caused turmoil in financial markets and forced banks to adjust the prices of their products to reflect the new market conditions. These developments led to several challenges in modeling NMDs.
While accurate and robust models for NMDs are now more important than ever, these models typically consist of several building blocks, which together provide as complete a picture as possible of the expected behavior of the portfolio.
One of these building blocks is the calibration approach for parameterizing the relevant model elements. And that is exactly what this article by Bas van Oers (Manager at Zanders) focuses on.
One of the main challenges currently facing risk model makers is defining the expected "repricing profile" of NMDs, which is crucial for proper portfolio risk management. In addition, banks must substantiate their modeling choices and associated parameterizations, both internally and externally (including regulators).
Traditionally, banks use historically observed relationships between behavioral components of deposits and their drivers for parameterization. With significant changes in market conditions, historical data has lost (some of) its predictive power.
Alternatively, many banks are now considering the use of forward-looking scenario analysis to replace or supplement historical data.
In many European markets, the extent to which customer deposit rates follow market trends (called repricing) has declined over the past decade. Repricing initially declined because banks were reluctant to further reduce interest rates below zero. Today, we still see a slower pace as interest rate increases are not immediately passed on to deposit rates.
The long period of low and even negative interest rates creates a bias in the historical data available for calibration, making this information less representative. This is especially true because the historical data do not cover all phases of the economic cycle. On the other hand, the historical data still contains valuable information about customer and pricing behavior, so ignoring these observations completely does not seem wise either.
To address these issues, risk and ALM managers should analyze the extent to which historical "repricing" behavior is still representative of the coming years and whether it is still in line with the bank's current pricing strategy.
It can be useful for banks to test model forecasts against expectations arising from economic logic. Given the strategic relevance of this issue and the impact of the portfolio on the overall balance sheet, the bank's senior management is usually closely involved in this process.
Common sense and understanding of the dynamics of deposit models are an integral part of the modeling process. Best practices in deposit modeling include preparing a comprehensive set of possible (interest rate) scenarios for the future.
In order to create an accurate representation of all possible future market developments, both downward and upward scenarios must be included.
The slope of the interest rate scenarios can be adjusted to reflect gradual changes over time, or sudden steepness or flattening of the curve. Price experts should be consulted to determine the expected trends of deposit rates over time for each of the interest rate scenarios.
Parameters for the deposition model should be chosen such that the estimates on average provide the best fit to the scenario analysis.
In going through this process, organizations should be aware that the effects of consulting pricing experts work both ways.
Risk and ALM managers will improve deposit models by using forward-looking business insights, while understanding of the market within the organization will improve through the model predictions.
