The Global Financial Crisis (GFC) clearly exposed the shortcomings of the incurred-loss (IL) model framework for establishing allowances for loan losses under International Accounting Standard (IAS) 39. All IAS 39-compliant provisions are required to ...
The Global Financial Crisis (GFC) clearly exposed the shortcomings of the incurred-loss (IL) model framework for establishing allowances for loan losses under International Accounting Standard (IAS) 39. All IAS 39-compliant provisions are required to be linked with objective evidence of asset impairments based on current information and events. In this sense, the IL model is backward-looking and does not consider potential future impairments and ensuing losses. The IL model has been criticized as being “too little, too late”, exacerbating the GFC since it delayed credit loss recognition losses and failed to ensure sufficient financial resources to absorb surging credit losses during the Crisis (e.g., see Financial Services Authority (FSA) [2009]; Financial Stability Forum (FSF) [2009]; U.S Treasury [2009]). To address the shortcomings in the incurred loss framework, the Korean Financial Supervisory Services (FSS) introduced complementary policy measures when it first introduced IAS 39 in 2011. Before 2011, loss provisions in Korea had been set according to a hard-and-fast rule based on supervisors’ classifications of asset quality. Under that regime, banks had little discretion in determining their provisions. Explained in more detail below, the complementary measures introduced by the FSS required that banks set provisions according to a rule which considered forward-looking criteria (FLC). In that respect, these measures were more closely aligned with an Expected Credit Loss (ECL) model and, therefore, the framework underpinning IFRS 9, which is due to take effect in 2018.
Against this backdrop, this paper examines for empirical evidence on how the IL model-based accounting standards and the FLC policy measures introduced by the FSS have affected Korean banks’ provisioning and lending behaviours. Using the regulatory returns of Korean commercial and savings banks spanning 2006 to 2014, we explore the linkage between loan loss provisions and lending. The purpose of this paper is to shed light on possible implications of IFRS 9, the new expected credit loss-based accounting standards due to replace IAS 39 in 2018. To preview our results, we find evidence consistent with the idea that IAS 39 and the complementary regulatory measures altered bank behaviours. In particular, we find evidence that: (i) the Korean regime has been generating higher provisions than expected under IAS 39; (ii) more discretion under the forward-looking approach has encouraged banks’ procyclical provisioning behaviours incorporating information on macroeconomic outlooks; and (iii) the procyclical provisioning behaviours may consequently amplify the procyclicality of banks’ lending, particularly resulting in more contracted credit supply in recessionary periods. The results suggest that discretion enhanced under IFRS 9 may have countervailing effects, promoting safety and soundness of individual banks but undermining financial stability as a whole, were it not for measures which can effectively disincentivise the. The results imply that policy measures in response to IFRS 9 will need to balance carefully between micro- and macro- prudential perspectives.
The Korean case study shows that even without the change in economic substances the institutional change and the interactions among them may affect bank managements’ decision-making, leading to unintended consequences. The Korean contexts may provide an appropriate setting to simulate the possible effect of IFRS 9 for the following reasons. Firstly, the Korean provisioning requirements noticeably have the components that are similarly to the ECL model of IFRS 9 (such as more judgmental provisioning covering the expected credit losses with the forward-looking information on borrowers’ creditworthiness and macroeconomic conditions, earlier recognition of greater asset impairment). Secondly, the staggered implementation of the accounting and prudential standards enables the comparison with the control group with the similarity of business models and the application of difference-in-difference (DiD) approach. The DiD approach could control for any omitted factors such as other institutional and macroeconomic changes that are concurrent with but not related to the forward-looking provisioning requirements (Leuz and Wysocki [2016]). To our belief, this study could shed light on research on the interaction between accounting and prudential standards, behavioural changes likely entailed by IFRS 9, and methodologies for ex-post evaluation of banking policies.
The rest of the paper is arranged as follows. Section 2 provides the background behind the accounting standard and regulations on loan loss provisions in Korea. Section 3 discusses our research in relation with the existing literature on loan loss provisions. Section 4 develops hypotheses and discusses the sample selection and research designs. Section 5 presents bank-level panel data analyses of Korean banks’ behavioural changes in provisioning and lending after the adoption of IFRS. Finally, Section 6 concludes and discusses policy implications.