Following the financial and banking crisis of the late 2000s, accounting regulators sought to replace the incurred-loss method of loan-loss provisioning by a more forward-looking expected-loss method. Difficulties arose, including with respect to the weight that expected-loss provisioning should place on objective evidence of loss relative to evidence of a less specific and more judgemental nature. This paper provides evidence relevant to this issue by examining whether loan-loss provisioning by UK banks was less timely under the stricter evidence requirements of the IAS 39 incurred-loss regime implemented in 2005 than under the less strict evidence requirements of the previous UK incurred-loss regime. It does so by reference to the relationship in time between loan write-offs and loan-loss expense. The results do not suggest that provisioning became less timely under the stricter evidence requirements of IAS 39. There is no evidence that provisioning became less timely immediately prior to the crisis of the late 2000s. Also, there is no evidence that general provisioning, permitted under the pre-IAS 39 regime, enhanced the timeliness of loan-loss provisioning. The results do not suggest that stricter requirements regarding the evidence necessary to support recognition of loan losses have resulted in less timely loan-loss provisioning.