Article 5, point (d) of EBA/RTS/2022/10 states that “margins of the new instruments shall be based on the margins from recently bought or sold products with similar characteristics”. Article 4, point (h) of the same RTS states that “institutions shall include instrument-specific interest rate caps and floors”. In response to RTS consultation EBA confirms that “commercial margins should be scenario independent to avoid overlap with other risks – i.e., CSRBB or business model risk”.

In certain market conditions these three requirements become contradictory for Non-Stable/Non-Core NMDs. Specifically it is the case, when the client rate on Non-Core NMDs is lower than the size of the regulatory large decline scenario. This becomes problematic if an overnight rate is larger than the scenario size. Such circumstances were triggered in Poland when the most recent tightening cycle started,  following a period of Covid-19 fiscal support programmes and a loose monetary policy. The accumulated NMDs at close-to-zero client rates have been repricing slowly, given over-liquidity in the banking sector, hence providing higher margin.

Let’s assume an overnight rate for PLN is 300bps, Non-Core NMD client rate is 100bps and PLN supervisory shock is -250bps. In the baseline scenario the margin is 200bps (difference between the overnight rate and client rate). In the shock scenario, if we floor the client rate at 0bps then margin drops to 50bps and we violate the condition of scenario independent margins. If we keep margin at the level of 200bps we violate the product floor condition (client rate falls below zero).

Currently in Poland assuming that client rate falls below zero does not make economic sense. Applying a floor to a client rate effectively leads to a smaller shock being applied to unstable liabilities than to assets held against them (e.g. 7-day central bank bills). In an example above former is 100bps whereas latter is 250bps. This asymmetry drives utilisation of the NII SOT and hence (wrongly) incentivises a bank to move those assets into longer repricing buckets. In other words, it incentives a bank to stabilise NII on unstable funds. The SVB case is an extreme scenario of such a mistake.

In view of mBank’s IRRBB Risk team, for Non-Stable/Non-Core NMDs the requirement to keep margins scenario independent should prevail over using the most recent margins. In accordance with the definition the Non-Stable and Non-Core part of NMDs may outflow or reprice in the near future and therefore it cannot be used for NII stabilization. It should rather be invested in short-term assets and have a short repricing date. This is justified by the fact, that excess levels of liquidity may temporarily lead to low client rates and provide high margin. However, it’s not prudent to assume that the bank will be able to maintain the balance sheet constant while keeping these high margins going forward, even in the baseline scenario. When funds with zero or near-zero interest rate start to outflow, typically banks have to pay market rates to replace them. This has been exemplified by the SVB, which  experienced the cash burn and had to refinance these unstable parts at higher prevailing market rates to maintain the balance sheet, hence eroding margin that it enjoyed during unusually loose monetary policy environment. Therefore in the above example, we would suggest assuming in the baseline scenario that the Non-Core NMDs, in order to maintain the constant balance sheet, reprice at prevailing market rates of 300bps, hence leaving zero margin.

Given above our suggestion is to apply exemption to assumptions (in addition to exemption already applied to monetary policy items, like TLTROs) and allow for Non-Stable and Non-Core NMDs to be presented in NII SOT, as repricing at current market rates without any margin. This approach is prudent from the risk hedging and capital protection perspectives, as it avoids stabilizing margins that may be just transitory. It is also consistent with a typical bank’s FTP policy, in which the non-stable part is assigned an internal overnight rate.