Choice 'a' (The amplification of smaller initial shocks to one risk factor creating larger subsequent shocks through system-wide interactions between other risks, creating self-perpetuating downward stresses in the markets) is the most comprehensive description of 'feedback effects', as described in the BCBS document on stress testing. Choice 'c' is one manifestation of feedback effects, but does not describe the entire effect. Choice 'b' is not a description of 'feedback effects', but one of the various weaknesses in stress testing that was seen during the crisis. Choice 'd' is plain nonsensical.

The BCBS paper provides a good and succinct description of feedback effect: how mortgage default shocks led to a deterioration of market prices of CDOs, followed by a drying up of the liquidity in these markets. This led to banks having to hold on to assets they intended to securitize (securitization and warehousing risk), and given the absence of transparency on who was exposed to what, banks refusing to lend to each other and a drying up of the wholesale funding market as well. All of this was additionally accompanied by a general flight to quality, households withdrawing money from money market funds creating a crisis in that market as well. At each stage, the initial shock was amplified and fed back into the system through interactions that had not been imagined by any market participant or regulator, leave alone risk managers.