Sebastian J A de-Ramon, William Francis and Qun Harris
Shakespeare first coined the term ‘sea change’ in The Tempest to describe King Alonso’s lasting transformation after his mystical death by drowning. Resting five fathoms deep, Alonso suffers a sea change into something rich and strange, with coral for bones and pearls for eyes. In a recent working paper, we explore for evidence of a possible sea change in UK banks’ balance sheets using data spanning the 2007-09 crisis. Our initial dive into the still murky, post-crisis waters shows signs of something strange and unrecognizable, with UK banks, in response to higher capital requirements, increasing the level and in particular the quality of capital more after the crisis. This post describes our dive and its findings.
In principle, banks can respond in a number of different ways to a change in capital requirement. They can, for example, change the size and risk-weighted nature of their asset portfolios, alter regulatory capital or undertake a mix of such actions. They can also elect to do nothing, particularly if they have sufficient buffers to satisfy higher requirements, though there is evidence suggesting that they may do otherwise. Previous impact studies (e.g., BCBS, 2010a; BCBS, 2010b; de-Ramon et al, 2012; Bank of England, 2013; Bank of England, 2015) show, however, that gauging the extent of such responses is challenging, especially without at least some fundamental understanding of how banks responded in the past.
What do the aggregate data suggest about behaviour?
Figure 1 shows how aggregate UK bank balance sheet growth (left-hand panel) evolved alongside capital resources and overall requirements (right-hand panel) from 1989 to 2013. As described in our working paper, overall requirements include the combination of the international minimums as mandated by the Basel capital standards as well as additional bank-specific requirements for risks not appropriately captured under the Basel rules (as required under the UK’s regulatory capital regime). Figure 1 shows a fall in lending (red bars) and risk-weighted assets (RWA) (yellow bars) that started during the height of the crisis, just before the overall sector capital requirements (blue line) and capital ratios (red line) began a considerable upturn. A similar, but more subtle pattern holds in the early 1990s, another period marked by UK banking problems and higher requirements with the advent of the Basel risk-based capital standards.
Figure 1: Aggregate annual banking sector growth rates and capital requirements 1989-2013.
Source: Bank of England and author’s calculations
While Figure 1 seems to indicate that there may have indeed been some sort of change in the way that banks managed capital ratios and balance sheets in response to higher capital requirements after the crisis, it’s difficult to draw firm conclusions based on aggregated data. This is because the apparent changes can be significantly influenced by the largest banks in the UK, which are more heavily weighted in the aggregate analysis above.
What do granular data suggest about behaviour?
To dive deeper into UK bank behaviour, we used data on individual bank capital requirements and balance sheets for more than 100 institutions over the period 1989 to 2013. We followed the two-step approach in Hancock and Wilcox (1993, 1994), which examined how US banks responded to higher capital requirements introduced under Basel I in the early 1990s.
Figure 2 sets out the basic intuition behind this approach. In the first step we developed and estimated a model of banks’ targeted capital ratios, accounting for capital requirements and other factors from the literature that explain such targets, such as management buffers (e.g., Ediz et al., 1995; Alfon et al., 2005; Berrospide and Edge, 2010; and Francis and Osborne, 2010). Armed with this information, in the second step we examined how changes in capital requirements that move banks away from their target ratios influence banks’ decisions about altering capital (total and tier 1), balance sheet size, risk-weighted assets and lending. To evaluate whether and how banks’ responses changed from their pre-crisis norms, we include a series of post-crisis (defined as spanning from the 2009-end to 2013-end) interaction terms in the regression models underlying these two steps.
Figure 2: Quantifying the balance sheet effects of capital requirements.
STEP 1: Did capital requirements affect banks’ target capital ratios?
Results from the first step confirm a positive and statistically significant association between banks’ target ratios and capital requirements. This relationship suggests that banks reacted to higher requirements by raising capital ratios — even when those requirements did not appear binding. But we also find that banks did not raise capital ratios by the full amount of an increase in requirements. Instead, they tended to raise capital ratios by only about 30% of the increase in requirements within the first six months and by around 80% of the change in the long-run. These estimates reflect the average short-run and long-run response rates for more than 100 banks across our 25-year estimation period. That is, they reflect how banks, on average, altered capital ratios in response to changes in capital requirements over this timeframe. As described in our paper, we tested statistically whether banks’ behaviour with respect to target capital changed after the crisis, but found no evidence that post-crisis behaviour differed from pre-crisis behaviour.
STEP 2: How did banks move towards new targets in response to higher capital requirements before and after the crisis?
Consistent with prior findings, UK banks have typically responded to higher requirements by raising total regulatory capital, including lower-quality capital instruments, and by reducing balance sheet size, risk-weighted assets and lending. Figure 3 provides a sense of these behaviours and how they have evolved over time. This figure shows the change in the annual growth rate (in basis points) of balance sheet components in response to a 1 percentage point (pp) increase in capital requirements before and after the crisis.
Figure 3: Change in annual growth in response to a 1 pp increase in capital requirement.
Source: Bank of England and authors’ calculations
It shows that after the crisis banks placed more emphasis on raising capital levels – especially better-quality tier 1 capital. Prior to the crisis, a one percentage point increase in capital requirements resulted in around a 30 basis points increase in the annual growth rate of total regulatory capital, while the impact on the annual growth of tier 1 capital was considerably lower at around 10 basis points. These changes are statistically significant. This disparity is evidence of a ‘pecking order’ in UK banks’ capital adjustment practices, with banks tending to increase better-quality, higher costing tier 1 capital relatively less than overall capital (which includes ostensibly lower-quality, lower-costing tier 2 elements). Our results suggest that while this pecking order behaviour continued after the crisis, it narrowed, with banks placing much more emphasis on growing tier 1 capital. After the crisis, the growth rate in total capital increased by more than 50 basis points, while that for better-quality, tier 1 capital surged by over 40 basis points. As a result of banks adjusting more on the liabilities side (via capital), the changes to growth rates of loans, assets and risk weighted assets are all less pronounced after the crisis. While our paper doesn’t investigate the drivers of these changes, it may be that the more stringent capital requirements under Basel III, due to take effect in 2019, influenced banks’ capital and balance sheet practices, including decisions about voluntary capital buffers, during the post-crisis period in our study (2009-end to 2013-end). This may mean that some of our findings may, in fact, reflect the forward-looking nature of banks’ responses during the few years immediately after the crisis and as Basel III was being negotiated.
So what does this all mean for assessing possible balance sheet impacts?
The empirical evidence suggests that expected effects on bank balance sheets could differ depending on whether judgments about reactions are informed by pre- or post-crisis behaviour. To provide a sense of the disparity (and to keep things simple), Table 1 reports balance sheet responses for a hypothetical bank based on our estimated pre- and post-crisis response parameters. The more pronounced impact on total regulatory capital and tier 1 capital after the crisis is highlighted in the rightmost column. Post-crisis responses clearly enhance the loss absorbing capacity of bank balance sheets, with much more emphasis placed on altering the better-quality tier 1 elements of regulatory capital after the crisis. The table provides a better sense of the differences in balance sheet strength – and the potential benefits of higher capital requirements — that arise when conditioned on pre- versus post-crisis response parameters.
Table 1: Impacts of a 1 pp increase in capital requirements
Source: Bank of England and authors’ calculations
Our initial findings point to evidence indicating that banks’ responses to changes in capital requirements have changed after the crisis. The early signs suggest that banks are raising capital — and, in particular, better-quality tier 1 capital — more in response to higher requirements after the crisis. Whether these changes constitute a ‘sea-change’ in bank behaviour remains to be seen and something that our ongoing research will hope to uncover. Stay tuned for updates on our future dives.
If you want to get in touch, please email us at firstname.lastname@example.org or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied.
Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.