Skip to content
All library documents

Modeling Core Non-Maturity Deposits with Replication Portfolios

Article Quant Q&A · Author: Cettt
Read the original

Summary

The document explains a proposed way to divide non-maturity deposits into core and non-core balances for interest rate risk modeling. It distinguishes stable balances, which tend to remain on account, from core balances, which are also unlikely to reprice when rates change. The author notes a lack of guidance on modeling that second distinction and asks for practical methods.

A bank example uses dynamic replication: optimize a portfolio of fixed income assets against the deposit balance to reduce the variability of its margin. The share modeled at an overnight or very short maturity is treated as non-core; longer modeled maturities count as core. The example also suggests choosing a portfolio on an efficient frontier according to a bank's preferred risk-return tradeoff. This is an illustrative practitioner account, not a comparative study: it provides no validation data, detailed optimization specification, or evidence that the approach generalizes across banks or rate environments.

Key ideas

  • Core deposits are the stable portion of non-maturity deposits that is expected not to reprice readily as rates change.
  • Dynamic replication models deposit behavior by optimizing a portfolio of fixed income assets against the deposit balance.
  • The overnight or very short maturity share can be classified as non-core, with longer modeled maturities treated as core.
  • Banks may optimize for lower margin variability or select a different risk-return tradeoff.
  • The described approach is a practitioner example and does not establish general performance.

Tags

Fixed incomeRisk managementPortfolio construction
Full text
# Core Deposits when modelling Non-Maturity Deposits according to IRRBB


# Core Deposits when modelling Non-Maturity Deposits according to IRRBB












When modeling Non-maturity deposits (NMDs) the Basel Committee suggests the following (see 31.109 of the guidelines):

> Banks should distinguish between the stable and the non-stable parts of each NMD category using observed volume changes over the past 10 years. The stable NMD portion is the portion that is found to remain undrawn with a high degree of likelihood. Core deposits are the proportion of stable NMDs which are unlikely to reprice even under significant changes in the interest rate environment. The remainder constitutes non-core NMDs.

I found a bit of literature on how to distinguish between stable and non-stable portions. However, I found nothing on modelling core and non-core deposits. And given the defintion of core deposits and the current interest-rate environment in central Europe I think there are not so many analytical methods to distinguish. Does anyone have experience in modeling core-deposits and/or can point me towards literature?

## Answer by simzoor (score 2)

https://quant.stackexchange.com/a/58153

This is more of a practital answer, but I've seen an approach by a medium-sized bank (balance sheet about 60 billion), which is already ECB-proof and which might answer your question:

In a nutshell, you can apply a dynamical replication approach, which tries to find the optimal portfolio consisting of fixed assets that you have to invest to archive a margin which has minimal standard deviation. You apply this to your whole deposit. Then you define non-core deposits as the proportion which has the overnight rate (or less than 1 month, depends on how you define it) of modeled maturity.

Example:

You have retail deposits, which technically have a maturity of 1 day. After applying your optimizing algo, the result is: Invest 10% at the overnight interest rate, 30% at the 1 year rate and 60% at the 10 year rate to archive a margin with lowest standard deviation.

You can go one step further and apply a CAPM approach, which tries to find the optimal mixture in regards on an efficient frontier, which is not the minimum standard deviation margin, but the one with the optimal "sharpe ratio", i.e. which has the best risk-return-tradeoff for you as a bank.

So in the end, you result with 10% non-core deposits and 90% core deposits of your overall retail deposits (which are NMDs).

Hope this helps a bit.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.