Risk Type

Option risk

Regulatory Definition

Risk arising from options (embedded and explicit), where the institution or its customer can alter the level and timing of their cash flows, namely the risk arising from interest rate sensitive instruments where the holder will almost certainly exercise the option if it is in their financial interest to do so (embedded or explicit automatic options) and the risk arising from flexibility embedded implicitly or within the terms of interest rate sensitive instruments, such that changes in interest rates may affect a change in the behaviour of the client (embedded behavioural option risk).

EBA GL/2022/14

What This Actually Means

Option risk splits into two broad categories. The first is customer-driven: your customers have choices that cost you money. They can prepay their fixed mortgage early (you lose a high-yielding asset when rates fall), or they can move their deposits to a competitor (you lose cheap funding when rates rise). Some of these options are contractual, most are behavioural.

The second is market-driven: explicit option products — caps, floors, swaptions, and similar derivatives — create option risk directly on the balance sheet. Caps embedded within floating rate products set a ceiling on the rate the borrower pays; floors set a minimum. These can be embedded within products (a floored mortgage that won't fall below 0%) or held as standalone derivatives. Swaptions give the holder the right but not the obligation to enter a swap at a future date, creating asymmetric interest rate exposure. All of these introduce convexity into your risk profile that linear rate sensitivity measures like DV01 don't fully capture.

Where It Matters

Prepayment risk on fixed mortgages is the classic behavioural example. In falling rate environments, borrowers refinance and you're left reinvesting at lower rates. On the liability side, deposit migration in rising rate environments forces repricing. The asymmetry is the problem — customers exercise options when it hurts you most.

On the explicit options side, the risk is more measurable but harder to hedge perfectly. A cap sold to a borrower to limit their rate exposure transfers that risk to the bank — the bank must then hedge it, typically by buying caps in the market. Floors embedded in products (e.g. zero floors on mortgages) became a significant balance sheet risk in the negative rate environment and caught many banks underhedged. Swaptions create gap risk between the option expiry and the underlying swap maturity. The key challenge across all explicit options is that their value and risk profile are non-linear — gamma and vega exposures require active management and can spike materially in volatile rate environments.

Classification grey area — contractual vs. behavioural floors: contractual floors are unambiguously option risk. The floor is a legal commitment with a defined strike, an implicit or explicit premium, and a hedgeable payoff profile. Behavioural floors on managed rate products are different in kind. A managed rate product that does not practically go below zero has no contract, no defined strike, and no hedgeable payoff. It is not an option the customer holds — it is a reflection of management's discretion over repricing, constrained by competitive dynamics. This is better characterised as a deposit beta that compresses toward zero at the lower bound: a non-linearity in the pass-through rate, not a written option. You cannot hedge it with a cap or floor derivative in the same way you would hedge a contractual floor, because there is no contractual trigger — the 'exercise' is a management and customer behaviour interaction, not a legal event.

The practical test for classification: does the non-linearity have a contractual trigger and a hedgeable payoff profile? If yes — option risk. If the non-linearity is behavioural or managerial — it is gap/beta risk with a non-linear shape. This distinction matters not just for classification but for hedging strategy and capital treatment. See also gap risk entry for the deposit beta discussion.