Applying a credit risk mitigation technique for large exposure purposes

Can an institution renounce applying a credit risk mitigation technique (CRM technique) for selected exposures in calculation of capital requirements for credit risk and as a result not apply that technique for that exposures for large exposure (LE) purposes?

In accordance with Article 108(1) of Regulation (EU) No 575/2013 ("CRR”) institution may use CRM in the calculation of risk-weighted exposure amounts for the purposes of points (a) and (f) of Article 92(3) of CRR.   For LE purposes, in accordance with Article 399(1) of CRR “an institution shall use a credit risk mitigation technique in the calculation of an exposure where it has used that technique to calculate capital requirements for credit risk in accordance with Title II of Part Three, provided that the credit risk mitigation technique meets the conditions set out in this Article”.   Taking into account the articles cited above, can an institution decide not to use a CRM technique in capital requirements for credit risk for selected exposures (resulting in higher RWA for such exposures) and as a result not recognise that technique for that exposures under LE rules?   For example: Bank has two exposures (e1 and e2, both with risk weight 100%) to entity A, both fully guaranteed by entity B (guarantee g1 and g2, both with risk weight 50%). Guarantees g1 and g2 meet all eligibility conditions for capital requirements for credit risk and for LE purpose.   Can bank at the same time: 1) use guarantee g1 in capital requirements for credit risk (with risk weight 50%) and treat exposure e1 in LE as exposure to group of clients connected with B,  and 2) not use guarantee g2 in capital requirements for credit risk (risk weight applied to exposure e2 will be 100%) and treat exposure e2 as exposure to group of clients connected with A?   Not being able to follow the approach described above could push bank to follow suboptimal credit decisions. If LE limit is fully utilized for entity B, bank will be able to grant naked loans to entity A (without the guarantee) but will not be able to grant guaranteed loans to that entity. Bank would increase its risk this way as guaranteed loan is obviously less risky. Both the text of the regulation and economic reasoning support the proposed approach.