The freedom of variable savings

The freedom of variable savings

The interest expenses on variable savings are an important driver for a bank’s results. Unlike mortgage interest rates, for instance, there is no knowing when and to what degree the variable savings interest rate will be adjusted. The bank has the right to change this at any time. But how much freedom does the bank really have in this respect?

A relatively low savings interest rate will reduce interest expenses initially, but if this drives savers to move to competitors, the bank could face liquidity problems. A relatively high interest rate on the other hand will mitigate liquidity risks. The flip side of this is, however, that in combination with the extra inflow that can be expected, it serves to increase interest expenses (doubly).

Margin stabilization

In answering the question of how much freedom a bank has in adjusting the savings interest rate, insight into the so-called ‘interest rate maturity’ of variable savings is essential. The interest rate maturity reflects the dependency of the savings interest rate on the market interest rate (often the swap rate). Although it is often summarized with a single maturity, this is often based on various maturities.

The starting point in determining the interest rate maturity is margin stabilization. ‘Margin’ is defined here as the difference between the interest income corresponding to a (fictional) investment portfolio and the interest expenses, determined on the basis of the savings interest rate. Both components cause uncertainty in the margin. The future (fictitious) interest income is mainly dependent on the investment rule(s) used, the (development in the) market interest rates and the volume developments; the development in the savings interest rate mainly depends on the development in the market interest rates.


The interest rate maturity of the investment portfolio to be set up replicates the interest rate maturity of the savings – hence the term ‘replicating portfolio’. The investment portfolio follows by finding an investment strategy that produces a margin that is as stable as possible in different volume and interest rate scenarios. A savings interest model and a savings volume model are essential input for this.

Savings interest model

An important driver for the savings interest model is customer sensitivity to the interest rate. Roughly three types can be distinguished: interest-sensitive customers, semi-sensitive customers and insensitive customers. Interest-sensitive customers are quick to change bank if a higher interest rate is offered elsewhere. That is why the savings interest is always close to the highest savings interest on offer (in other words, the maximum savings interest). One option is to directly model this maximum savings interest.

Semi-sensitive customers generally go to little trouble to change bank and trust that their house bank (often a large bank) offers a reasonable interest rate. The height of the savings interest rate depends on, among other things, the maximum savings interest rate on the market, but also largely on the implicit investment yields of the savings; which produce mortgages and business loans, for instance. In order to take this last component fully into account, it is wise to model the bank’s own savings interest rate. By modeling historically, the historically observed ratio between the two is taken into account.

Insensitive customers, finally, have reasons other than the high interest rate to choose a savings product, such as the socially responsible designation attached to a product, for instance. The height of the savings interest rate in that case depends more on the implicit yields of the investments into which the savings are channeled, which means that the link to the maximum savings interest rate is of less importance here.

Savings volume model and investment strategy

Different forms of refinement are possible for the savings volume model. These range from basic models, such as a constant (growing) savings volume, to more refined models in which interest rate dependency is included in the modeling. More refinement results in more insight into the variability of the interest rate income and expenses, which means the bank will be better able to stabilize its margin. It could, however, entail a higher model risk.

The combination of the savings interest model and savings volume model is then used to find the investment strategy that stabilizes the margin. This strategy follows from a so-called investment rule, which prescribes the maturities on which the ‘freed up’ funds are invested.
Depending on the complexity of the savings volume model, the optimal investment rule can be derived (in the event of constant volume models) or determined from simulations (in the event of stochastic volume models).

Interest rate maturity within balance sheet management

The optimal investment rule serves balance sheet management in two ways. Firstly by determining, on the basis of the fictitious investment portfolio, the interest rate maturity of savings, but also interest rate sensitivity measures, such as the duration and basis point value (BPV), for interest rate risk reports and management. In this way, the overall interest rate position of the bank as a whole can be better managed. The investment rule can also function within a funds transfer pricing framework.

By making (internal) investments to a central treasury, and in doing so agreeing on an internal transfer price, the interest rate risk is transferred to treasury, while the model risk is left behind at the business. This simultaneously improves transparency, since treasury only deals with ‘normal’ interest rate products, which replace the savings. The bank also gains more insight into the profitability of the savings by comparing the income from the (internal) investments with the savings interest expenses. In both cases, this helps the bank to manage its interest rate position and, in doing so, to remain ‘in control’ in relation to changes in the interest rate.


More articles about ‘The modeling of savings’: