[From Guidance on implementing IFRS 9 *Financial Instruments*]

**Embedded derivatives: separation of embedded option**

**The response to ‘Separation of host debt instrument‘ states that the terms of an embedded non-option derivative should be determined so as to result in the embedded derivative having a fair value of zero at the initial recognition of the hybrid contract. When an embedded option-based derivative is separated, must the terms of the embedded option be determined so as to result in the embedded derivative having either a fair value of zero or an intrinsic value of zero (that is to say, be at the money) at the inception of the hybrid contract?**

No. The economic behaviour of a hybrid contract with an option-based embedded derivative depends critically on the strike price (or strike rate) specified for the option feature in the hybrid contract, as discussed below. Therefore, the separation of an option-based embedded derivative (including any embedded put, call, cap, floor, caption, floortion or swaption feature in a hybrid contract) should be based on the stated terms of the option feature documented in the hybrid contract. As a result, the embedded derivative would not necessarily have a fair value or intrinsic value equal to zero at the initial recognition of the hybrid contract.

If an entity were required to identify the terms of an embedded option-based derivative so as to achieve a fair value of the embedded derivative of zero, the strike price (or strike rate) generally would have to be determined so as to result in the option being infinitely out of the money. This would imply a zero probability of the option feature being exercised. However, since the probability of the option feature in a hybrid contract being exercised generally is not zero, it would be inconsistent with the likely economic behaviour of the hybrid contract to assume an initial fair value of zero. Similarly, if an entity were required to identify the terms of an embedded option-based derivative so as to achieve an intrinsic value of zero for the embedded derivative, the strike price (or strike rate) would have to be assumed to equal the price (or rate) of the underlying variable at the initial recognition of the hybrid contract. In this case, the fair value of the option would consist only of time value. However, such an assumption would not be consistent with the likely economic behaviour of the hybrid contract, including the probability of the option feature being exercised, unless the agreed strike price was indeed equal to the price (or rate) of the underlying variable at the initial recognition of the hybrid contract.

The economic nature of an option-based embedded derivative is fundamentally different from a forward-based embedded derivative (including forwards and swaps), because the terms of a forward are such that a payment based on the difference between the price of the underlying and the forward price will occur at a specified date, while the terms of an option are such that a payment based on the difference between the price of the underlying and the strike price of the option may or may not occur depending on the relationship between the agreed strike price and the price of the underlying at a specified date or dates in the future. Adjusting the strike price of an option-based embedded derivative, therefore, alters the nature of the hybrid contract. On the other hand, if the terms of a non-option embedded derivative in a host debt instrument were determined so as to result in a fair value of any amount other than zero at the inception of the hybrid contract, that amount would essentially represent a borrowing or lending. Accordingly, as discussed in the answer to Question C.1, it is not appropriate to separate a non-option embedded derivative in a host debt instrument on terms that result in a fair value other than zero at the initial recognition of the hybrid contract.