Page 43 - Monocle Quarterly Journal Vol 1 Issue 1 Q4

lIquIdIty At RISK: A meASuRemeNt AppRoAch wIthIN BANKING INStItutIoNS
Since market rates are easier to simulate and project into the future, the model allows the bank to understand how each market rate scenario in a large simulation (typically 100 000 iterations) translates into changes in internal rates, resulting in a prediction of internal rates into the future. is can then be used for Net Interest Income (NII) and other pro tability scenario analyses.
Asset/Liability Portfolio Growth Models
Typically a liquidity risk model is premised on a run-o basis, where assets and liabilities are not replaced as they reach maturity, or on a business- as-usual basis, where assets and liabilities are replaced as they mature. However, to achieve true business value, we need to understand at a product level, and at a client-type level, what the growth in a particular asset or liability will be. LaR is an ‘adjusted business-as-usual’ approach, in which a projected growth in each asset and liability portfolio, as a result of changes in economic factors, is taken into account.
As part of the simulation process, there is a series of techniques to predict the future growth in asset and liability values on a bank’s balance sheet. e models rely on the historical relationship between growth rates and interest rates, which can be combined in a multi-variable regression. Of course, we know that relationships between interest rates and growth levels have not always remained intact – particularly in the recent nancial crisis. Market interest rates (interbank rates) have been near all-time lows for some time, which, under normal circumstances, would be a leading indicator of higher growth rates. But nancial institutions have decreased their appetite for risk, and for extending credit, drastically reducing growth levels below what would normally be expected.
Because of this, growth models incorporate a “desirability” factor, which is a numerical indication of the bank’s appetite for extending credit, or for growing a particular product type. is “desirability” factor adjusts the output of the growth models, which are purely linked to interest rates.
Economic variables other than interest rates can also be taken into account, particularly those variables which are shown to be leading indi- cators, such as economic variables; inventories, sentiment indices, and money supply growth.
Combining Predictive Models in the LaR
LaR typically requires generating 10 000 to 100 000 simulations of underlying market factors over a one-year horizon. At a minimum, LaR
“LaR is an ‘adjusted business-as-usual’ approach, in which a projected growth in each asset and liability portfolio, as a result of changes in economic factors, is taken into account.”
41