Basel III liquidity options

Where an individual bank faces a “name” crisis, it may be able to sell non HQLA assets without creating “fire sale” conditions. Consequently, the need for that bank to tap the liquidity facility is likely to be reduced, and the likelihood of RBA liquidity injection is reduced. But nevertheless, there is a fundamental change away from requiring basic liquidity protection solely by private provision.

This raises two possible alternatives. One is that other private sector securities which are “repo-eligible” at the RBA could be included in LCR eligible assets. The Basel Committee provides for the option for “level two” assets to be included with a minimum 15 per cent haircut, and that haircut could be increased if it was thought appropriate to allow some available assets (mortgage backed securities, Kangaroo bonds) to count as level two assets.

Another option is that banks could build up their holdings of Exchange Settlement Accounts at the RBA – because these also count towards the LCR. Currently the RBA pays interest on these at 25 basis points below the target cash rate, and the banks minimise ESA balances accordingly. This raises the question of the pricing of the proposed liquidity facility.

Pricing of the RBA liquidity facility

The “Australian solution” requires that a fee be set for the liquidity facility that is “fair” – in the sense that there is no benefit/cost from using the facility (and holding eligible non-HQLA assets as potential collateral) relative to holding additional HQLA to meet the LCR requirement. There are two fundamental problems here. One is that the counterfactual involves setting a fee, which is related to the credit-risk adjusted yield of additional holdings of level one and two HQLA, when there are none of the latter for which a yield is available.

The second problem is that a fundamental simultaneity problem exists, as follows. The reason for the safety valve facility is that the cost to banks of acquiring HQLA beyond some level is seen as prohibitive, given the stock available. As banks attempt to increase their holdings of government debt, their demand will drive down its yield relative to other investments. Wherever the liquidity facility fee is set will affect the relative use of the liquidity requirement versus holding HQLA, with the latter in turn affecting the cost to banks of using HQLA. The Figure below illustrates. The opportunity cost to banks of holding more government debt (the additional risk adjusted yield foregone on alternative private debt) is shown as increasing in their holdings of HQLA (because the higher bank demand for government debt drives down its relative yield). If the liquidity facility fee is set at Fee_a, HQLA holdings will be at HQLA_a, if it is set at Fee_bthe holdings will be at HQLA_b etc. Thus there is no unique fee level.

Exchange Settlement Account Arrangements and the Liquidity Facility Fee

A further complicating feature arises from the ability of banks to use Exchange Settlement Account (ESA) balances to meet the LCR requirement. Currently the RBA pays 25 basis points below the cash rate, and banks manage their liquidity to keep minimal ESA holdings (lending surplus funds to other banks overnight at the cash rate). Conceivably, and ignoring the simultaneity discussed above, it may not be possible for the RBA to always charge a sufficiently low fee for its liquidity facility so as to make that more attractive than building up ESA balances.

Consider, for example, the situation where a bank holds repo-eligible RMBS to support the liquidity facility.  Because there is a haircut given to level two assets of at least 15 per cent, $100 of HQLA would require (say) $120 of holdings of level two assets. The fee for the liquidity facility should thus be for a facility of $120 and would reduce the net return on the $120 of repo-eligible assets backing that to of a government bond rate equivalent. The consideration for the bank is whether it would prefer to hold $120 of assets earning (net of the liquidity facility fee) the government bond rate compared to the alternative of $100 in its ESA account earning the cash rate less 25 basis points and another $20 in higher yielding private sector assets. The slope of the yield curve (long term versus short term rates) and size of credit spreads (for private sector securities over government rates) are relevant factors in this calculation.

While these arrangements remain to be determined, if there is a build up of bank ESA balances (rather than holding other assets), the RBA’s cash rate target would see it acquiring additional assets from the banks (such as government securities) and potentially aggravating the shortage of HQLA.

The implication is that the whole structure of arrangements for system liquidity management may need to be reexamined.