During the euro area sovereign debt crisis, the ECB implemented several adjustments to its collateral and haircut policies as part of its whole set of non-standard monetary policy measures. While haircut grids may sound like an obscure, merely technical aspect of monetary policy implementation, they turn out to be a key translator of monetary policy loosening with a direct impact on bank balance sheets, the shadow value of their assets and how encumbered they are. Changing haircuts immediately affects the quantity of liquidity that banks can claim against their collateral in open market operations (OMOs). This ‘reverse leverage’ effect from haircuts is compounded by the central bank’s decisions on the way it classifies assets according to predetermined credit categories. Depending on their ratings, assets can be downgraded or even suspended from eligibility for OMOs, which in turn also has direct implications for banks’ liquidity positions. In a recent paper, we construct a liquidity mismatch index (LMI) for the French banking sector that explicitly takes the ECB haircut policy into account (Legroux et al. 2018).
Bank liquidity mismatch index based on the ECB haircuts
Our index evaluates the mismatch between the market liquidity of assets and the funding liquidity of liabilities. More specifically, on the liability side, we assume that all of a bank’s creditors extract the maximum liquidity allowed under the terms of their contract, as in Bai et al. (2018). On the asset side, we assume that in case of liquidity stress the bank reacts by maximising the liquidity it can raise from its assets by pledging them as collateral in the OMO operations. Therefore, the central bank haircuts define the amount of cash the bank can raise using its securities and loans as collateral. The LMI is the difference between the asset-side liquidity and the liability-side liquidity. As both parts are expressed in euros, it gives the ‘cash equivalent value’ of a bank in the spirit of Brunnermeier et al. (2011).
The use of central bank haircuts instead of market haircuts is an important difference with respect to the literature (for example, Bai et al. 2018 measure the liquidity mismatch of the US banking sector using private markets haircuts). We argue that in the context of the euro area, ECB haircuts have been instrumental in maintaining bank liquidity in the last decade for a number of reasons. First, the haircuts applied by the ECB can be lower than those applied in private repo markets. When this is verified, for example in times of market stress when repo conditions tighten, or if the central bank decides voluntarily to deviate from market practice, this difference gives banks a clear incentive to mobilise assets as collateral at the ECB operations rather than in private repo markets. Second, in an environment where the central bank supplies liquidity to banks with no limits (the so-called fixed-rate full-allotment framework), the only effective bound to access ECB liquidity is in fact the amount of eligible collateral that the bank holds. As a result, banks extensively used Eurosystem OMO operations to exchange their assets for the highly liquid central bank money, thereby offloading the liquidity risk they were exposed to onto the central bank balance sheet. Finally, unlike private repo markets, the ECB also accepts non-marketable loans to non-financial corporations and public-sector entities as collateral. The ECB haircut on loans is therefore key to determining whether the assets of a bank are fully encumbered – and therefore how close a bank may be from the ‘fire sale’ point.
The French banking sector proved resilient to the 2011 liquidity shock
We calculate the LMI for individual French banks representing 87% of banking system total assets and then compute an aggregate measure for the whole banking system. When the aggregate LMI is low, the banking sector is more vulnerable to liquidity stresses. The evolution of banks’ liquidity mismatch between 2011 and 2015 is displayed in Figure 1a.
Figure 1 Liquidity mismatch (LMI) of the French banking system
A) Aggregate Liquidity Mismatch Index
B) Individual LMIs rebased (2011:Q1=100)
Note: (a) Aggregate LMI in billions of euros; (b) LMIs for eight banks in our sample rebased at 100 in 2011:Q1. A decrease in the LMI depicts a deterioration of the liquidity situation of the sector, i.e. an increase in the liquidity mismatch between assets and liabilities.
Our indicator delivers a number of policy-relevant messages. First, the banking system was still afloat prior to the 2011 central bank interventions. In the second half of 2011 the liquidity mismatch for the system as a whole drastically worsened, suggesting a rapidly evolving systemic vulnerability. Yet, the LMI never fell into negative territory, which demonstrates the resilience of the sector, in aggregate, to deteriorating conditions both on the asset and liability side. Moreover, the aggregate LMI stood higher (in absolute terms) at the end of the 2015 than before the 2011 liquidity shock. Most importantly, the aggregate LMI increases notably towards the end of the period, starting in end 2014, and manages to exceed its initial level. This could reflect the impact of the planned implementation of the liquidity regulation, and in particular the liquidity coverage ratio, the observation period of which was scheduled to start in 2015. Finally, however, the resilience of the system taken as a whole hides heterogeneity across counterparties (Figure 1b). The index can therefore help identify the fragility of individual French institutions in terms of their liquidity exposures.
ECB haircut policies provided sizable liquidity support to banks
Generally, banks can use their assets as collateral to obtain liquidity either by pledging them in central bank OMOs or in private repo markets. If banks were subject to higher market haircuts, their liquidity would be reduced in a similar way as, say, a withdrawal of deposits. It could even lead to fire-sales of assets in time of financial stress.
To get a sense of the extent to which the ECB haircut policy potentially released a ‘collateral constraint’ faced by banks in repo markets, we construct an alternative measure using central clearing counterparty (LCH S.A.) haircuts for sovereign bonds according to their maturity and country of issuance. One key difference in the pools of eligible collateral at the ECB versus on the market is that bank loans and some investment securities cannot be posted as collateral in private repo markets while they can be used as collateral at open market operations. We account for this difference too. Figure 2 shows the evolution of the liquidity mismatch based on ECB versus private haircuts in repo markets.
Figure 2 Public LMI compared to private LMI
The private LMI is consistently lower than the public LMI, which confirms that the ECB collateral and haircut policies have indeed alleviated banks’ liquidity mismatch. As such, the difference between public and private LMIs represents a solid proxy for ECB liquidity support to the French banking system. In 2012, this support approached €800 billion.
Collateral and haircut policies have gone under the radar for a long time, and in any case have been less popular measures of the monetary policy stance than interest rates or quantitative policies. Yet they are not only essential for the correct functioning of the monetary and financial systems (Bindseil et al. 2017), but are also a key instrument to tighten, or loosen, liquidity in the banking system. In this column we have presented a liquidity mismatch index that explicitly takes into account ECB haircut policies. Our analysis suggests that in the case of French banks, the ECB alleviated banks’ liquidity mismatch significantly between 2011 and 2015. This is evident when comparing the public LMI based on ECB haircuts with that based on haircuts applied to private repo transactions.
Authors’ note: This column reflects the views of the authors alone.
Bai, J, A Krishnamurthy, and C-H Weymuller (2018), “Measuring the liquidity mismatch in the banking sector”, Journal of Finance 73(1): 51–93.
Bindseil, U, M Corsi, B Sahel, and A Visser (2017), “The Eurosystem Collateral Framework Explained”, ECB Occasional Paper No. 189.
Brunnermeier, M K, G Gorton, and A Krishnamurthy (2011), “Risk Topography”, in NBER Macroeconomics Annual 2011, Volume 26: 149-176.
Legroux, V, I Rahmouni-Rousseau, U Szczerbowicz, and N Valla, “Stabilising virtues of central banks: (re)matching bank liquidity”, Banque de France Working Paper Series no. 667.