Calebe de Roure and Nick McLaren
Central banks accept a wide range of assets from participants as collateral in their liquidity operations – but can this lead to undesired side effects? Such an approach can enhance overall liquidity in the financial sector, by allowing participants to transform illiquid collateral into more liquid assets. But, as a result, the central bank then needs to manage the greater potential risks of holding these riskier assets on its own balance sheet. Financially weaker participants may be encouraged to hold these assets if they can benefit from the higher returns, which compensate for the greater risk. In our recent paper we investigate whether central banks’ acceptance of a broad set of collateral incentivises the concentration of risk by examining the experience of the Bank of England (BoE).
To protect themselves from counterparty credit risk, central banks take collateral in their operations. If a counterparty fails to repay a loan, the central bank can sell or retain the collateral to make good any loss. Central banks also apply a ‘haircut’ so that they lend an amount less than the market value of the collateral they take. The BoE’s risk management function assesses counterparties’ credit worthiness prior to lending to them, values the collateral assets and sets an appropriate haircut based on the risk characteristics of the asset.
During the Global Financial Crisis, one way the BoE supported the financial sector was by accepting a much wider range of collateral than it had done previously. This included riskier, less liquid assets such as loan portfolios. This experience showed that, during times of stress, central banks can most effectively provide liquidity insurance against collateral which are likely to become illiquid in this scenario (Fisher 2011). For this reason, even after the crisis, the BoE decided to continue accepting a broader range of eligible collateral in its regular liquidity insurance lending operations. Nonetheless underlying risks could remain if, in practice, the BoE’s liquidity facilities incentivise the use of riskier collateral assets and the participation of riskier banks.
To assess this, we take advantage of the administrative setup of the BoE, which allow us to build a quasi-natural experiment. Between 2010 and 2016, liquidity transformation in the BoE primarily took two forms: liquidity insurance (mainly using Indexed Long-Term Repos – ILTR) and the Funding for Lending Scheme (FLS). Over this period the collateral pools used for ILTR and FLS had separate structures. Thus, comparing both collateral pools allows us to draw conclusions of how liquidity transformation, collateral and counterparty risks interact.
To understand the different incentives, we need to understand the different setups of the liquidity lines. The ILTR aims to improve financial stability and ensure the transmission of monetary policy by safeguarding market liquidity. The FLS, in contrast, provides term funding for banks at rates below the market, in order to boost credit provision to the real economy. Related to these different policy aims, there are several differences in the design of the two facilities. First, the maturity of loans made under the FLS was up to 4 years, whereas the term of ILTR lending is 6 months. Second, the ILTR provides counterparties with central bank reserves whereas the FLS provides T-bills. Third, the ILTR is provided through an auction process, whereas the FLS can be accessed in a non-competitive manner, as long a counterparty has not reached their FLS borrowing limit.
Comparing both collateral pools, we find that the FLS pool is larger, riskier and less diversified than the ILTR collateral pool. This difference comes from the different fee and maturity structures of the operations. The ILTR fee increases for less liquid collateral. The FLS, on the other hand, has a flat fee structure for all collateral types, which favours the use of less liquid collateral assets. Furthermore, because the maturity of the FLS is much longer than the ILTR, counterparties have an incentive to use less liquid assets. For both reasons, the FLS collateral assets have higher liquidity risk.
To relate use of each facility to counterparty and collateral characteristics, we run a series of regressions. In each case, we regress counterparties’ ‘liquidity demand’ against their balance sheet characteristics (total assets, return on equity, equity ratio, loans write-off, and change in deposits) and the characteristics of the collateral assets used (haircut, and diversification index). We define ‘liquidity demand’ in two ways: (i) the amount of collateral deposited in the pool; and (ii) the amount of liquidity actually drawn in the facility. The first concept captures precautionary liquidity demand, i.e. banks deposit collateral with the BoE in advance, in case they subsequently need liquidity. The second captures realised liquidity demand. Our regressions are done in two forms. First, we use a probability model to estimate the likelihood that a bank decides to deposit collateral in a collateral pool (intensive margin). For banks that decide to deposit some assets in their collateral pool, we use a panel data model to understand how much liquidity they actually borrowed in the operations (extensive margin).
Analysing the demand for liquidity insurance, we find that relatively healthier counterparties (with more equity and lower loan write-off rates) are more likely to pre-position collateral for use in the Bank’s regular liquidity facilities, such as the ILTR. Amongst those, counterparties who experienced larger deposit outflows are more likely to draw upon the ILTR. These results are consistent with the view that sound counterparties demand liquidity insurance when liquidity is needed, which is the policy objective. Turning to the assets used as collateral in these operations, we find counterparties prefer to use liquid collateral assets initially, but larger liquidity demands are associated to less liquid collateral pools. This result is consistent with the view that counterparties use higher quality collateral assets first, and then turn to less liquid assets only if they need to expand their collateral usage. That is what we would expect given the higher fee charged on less liquid collateral.
Analysing the demand for FLS, we find that riskier counterparties (less profitable banks with higher loan write-off rates, albeit with more equity) are more likely to pre-position collateral for FLS usage. However, when looking into their actual FLS drawings, we find that riskier counterparties do not borrow substantially more than other counterparties. Thus, we cannot infer that FLS incentivises relatively weaker counterparties to borrow more. On the collateral dimension, we find no evidence that larger demand for FLS liquidity is associated with the use of riskier collateral assets. In part this result is related to the fact that about 90% of the collateral used for FLS is categorised as collateral type C, which is subject to the highest haircut.
In summary, we find no evidence for risk concentration among weaker counterparties and illiquid collateral. Rather concentrations of collateral assets in each facility follow the intended relative incentives of each operation. The lack of bad incentives suggests that the BoE is able to appropriately account for risk and liquidity differences between collateral assets when setting its haircuts.
Calebe de Roure works at the Reserve Bank of Australia and Nick McLaren works in the Bank’s Markets Intelligence and Analysis Division.
If you want to get in touch, please email us at bankunderground@bankofengland.co.uk 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.
