The macroeconomic impact of changes in economic bank capital buffers Prepared by Derrick Kanngiesser, Reiner Martin, Diego Moccero and Laurent Maurin

2 Challenges in estimating the impact of changes in bank capital buffers
Identifying the economic impact of changes in bank capital ratios is challenging because bank capital itself tends to respond to changes in bank lending and the macroeconomic environment. A large part of the variation in bank capital is likely to result from variations in macroeconomic variables (such as changes in economic activity and interest rate spreads), as well as from changes in economic policy per se. For example, changes in macroeconomic variables affect capital through operating income and asset valuation. The literature deals with this endogeneity issue in three ways.
A first strand of literature has tried to isolate “true” shocks to bank capital, e.g. those stemming from losses associated with declines in real estate prices or losses arising from exposures to third countries. For example, Bernanke and Lown (1991) find that a shortage of equity capital due to the bursting of a real estate bubble played a role during the credit crunch observed in the early 1990s in the United States. Peek and Rosengren (1997) found that Japanese bank branches’ lending to US non-financial corporations declined in response to heavy losses in their parents’ capital positions associated with the sharp decline in Japanese stock markets in the early 1990s. Watanabe (2007) also studied the case of Japan in the late 1990s and found that banks cut back on their lending supply in response to the large loss of bank capital associated with write-offs of non-performing loans in the real estate sector.
A second strand of the literature seeks to identify regulatory or supervisory shocks to bank capital, where banks are required by a regulatory or supervisory authority to hold higher capital ratios. Such shocks may be bank-specific or system-wide and are in some cases introduced over a transition period to allow for a smooth implementation and avoid unintended, short-term negative effects. Several studies have looked at the United Kingdom, where the regulators have imposed time-varying, bank-specific capital requirements since Basel I.^[4^] These studies show that more stringent capital requirements, while producing significant benefits such as greater financial stability and a lower probability of crisis events, may also have costs in terms of reduced loan supply, at least in the short term (Francis and Osborne, 2009; Aiyar et al., 2014; Bridges et al., 2014). In the euro area, Jiménez et al. (2017), looking at the experience with the Spanish dynamic provisioning framework, found that countercyclical dynamic provisioning helps to smooth credit supply cycles. Mésonnier and Monks (2015) and Gropp et al. (2016) find evidence suggesting that banks that had to increase capital ratios in the context of the 2011 EBA recapitalisation exercise increased their ratios by curtailing their lending supply, rather than by raising the level of capital.
A major drawback of the studies mentioned so far is that they tend to focus on specific, one-off events and neglect the dynamic interaction and feedback effects between banking variables and macroeconomic variables. Specific one-off events are difficult to generalise. Moreover, these studies are mainly based on single equation modelling frameworks, neglecting the dynamic interaction among banks’ asset categories, possibly assuming exogeneity across variables. Since the studies tend to focus on capital shocks to individual banks, they are ill-equipped to analyse the feedback loop between bank capital shocks and the macroeconomy. This feedback loop is, however, expected to play an important role when many banks face simultaneous and correlated changes in capital owing to regulatory changes, cyclical conditions, or the need to strengthen their capital base to regain market confidence.
As a result, a third strand of literature has estimated the impact of bank capital shocks by including aggregate bank ratios or buffers in macroeconometric models with a broader set of banking and macroeconomic variables. This strand of literature includes Hancock et al. (1995), Lown and Morgan (2006), Berrospide and Edge (2010), Kok et al. (2016), Noss and Toffano (2016), Meeks (2017) and Mésonnier and Stevanovic (2017) and Kanngiesser et al. (2017). All of these studies focus on the United States or the United Kingdom, with the exception of Kok et al. (2016) and Kanngiesser et al. (2017), who study European economies and the euro area, respectively. They find that shocks to bank capital tend to depress bank lending, increase lending spreads and hamper economic activity. This article follows this strand of literature and uses a capital buffer, rather than the actual capital ratio (as in Kok et al., 2016, and Kanngiesser et al., 2017), to identify the impact of shocks to bank capital in euro area countries.