As I noted in my post on Lehman's liquidity pool, the financial crisis of 2008 — and the Lehman episode in particular — highlighted the pressing need for formal liquidity requirements. Fortunately, the Basel Committee has made a new liquidity regime a focus of Basel III. The new liquidity regime can be found in the December 2009 consultative document as amended by last month's Annex.
The main component of Basel III's liquidity regime is the Liquidity Coverage Ratio (LCR), sometimes known as the "Bear Stearns rule." The LCR requires banks to maintain a stock of "high-quality liquid assets" that is sufficient to cover net cash outflows for a 30-day period under a stress scenario. The formula is:
Stock of high quality liquid assets ≥ 100%Net cash outflows, in turn, is calculated by applying run-off rates to different sources of funding (e.g., repos, unsecured wholesale, etc.). So the action here is in (1) the definition of "high quality liquid assets," and, more importantly, (2) the run-off rates used to calculate "net cash outflows."
Net cash outflows over a 30-day time period
1. High-Quality Liquid Assets
In the initial consultative document, the Basel Committee defined "high-quality liquid assets" extremely conservatively, such that the only eligible assets were essentially cash, central bank reserves, and sovereign debt. Crucially, Agencies and Agency MBS were excluded, due to the requirement in Paragraph 34(c)(i) that the assets have a 0% risk-weight under Basel II (Fannie and Freddie obligations have a 20% risk-weight). In last month's Annex, the Commmittee fixed this, adding a "Level 2" category of liquid assets that includes GSE obligations (but with a 15% haircut). Level 2 assets, which also includes non-financial corporate and covered bonds rated AA- or above, can't comprise more than 40% of a bank's total stock of high-quality liquid assets. I have to think this was a deliberate strategy — the Committee was always going to allow GSE obligations in the liquidity pool, but they wanted to give the banks something to howl about, and focus all their energy on.
So in sum, banks' liquidity pools have to be at least 60% Level 1 assets (cash, central bank reserves, and sovereigns) and no more than 40% Level 2 assets (GSE obligations, and non-financial corporate or covered bonds rated AA- or above). That's appropriately conservative, in my view — all of these assets would have been monetizable in 24 hours or less during the financial crisis.
2. Run-Off Rates
This is where most of the action is. The LCR proposal assigns run-off rates to each source of funding, which are designed to simulate a severe stress scenario. A run-off rate just reflects the amount of funding maturing in the 30-day window that won't roll over. I don't have the space to list all of the run-off rates — there's a handy table on pg. 32 of the consultative document, although some of the run-off rates were amended by last month's Annex. Here are the most important run-off rates (as amended by the Annex):
Retail DepositsWhat do I think of these run-off rates? I think they're mostly appropriate and sufficiently conservative. I was pleasantly surprised by how comprehensive the LCR proposal was — that is, the Committee seems to have anticipated all the meaningful sources of funding outflows. For example, I was impressed that the LCR proposal addressed changes in the value of collateral posted on derivatives trades. I'm also glad the Committee held the run-off rate for unsecured wholesale funding from financial institutions at 100%. Given what we saw during the financial crisis, that's entirely warranted.
- Stable deposits: 5%
- Less stable deposits: 10%
Unsecured Wholesale Funding (e.g., Commercial Paper)
- Non-financial corporates, sovereigns, central banks, and public sector entities:
- Without operational relationships: 75%
- With operational relationships: 25% of deposits needed for operational purposes
- Financial institutions and other legal entities: 100%
- Repos and securities lending/borrowing of non-liquidity pool eligible assets: 25%
- Liabilities related to derivative collateral calls related to a downgrade of up to 3-notches: 100% of collateral that would be required to cover the contracts in the case of a 3-notch downgrade
- Liabilities from maturing ABCP, SIVs, and SPVs: 100% of maturing amounts and 100% of returnable assets
- Currently undrawn portion of committed credit and liquidity facilities to non-financial corporates (including central banks, sovereigns, and PSEs):
- Credit failities: 10%
- Liquidity facilities: 100%
I was disappointed that the Basel Committee gave so much back on repos of non-liquidity pool eligible assets. Initially, the run-off rate was 100%. That was probably too high, but not outrageously so. The banks cried bloody murder, of course. Their main argument was that they were able to reliably repo out equities during the financial crisis (albeit with substantial haircuts), due to the deep market, and thus reliable marks, for most equities. That's a legitimate point; I just don't think it merits reducing the run-off rate from 100% to 25%. I thought the Committee would, at most, reduce the rate to 50%, and I think 75% would be more appropriate.
For the most part, though, the Basel Committee held firm. Partly that's because the banks' comment letters were surprisingly weak. Their main argument involved claiming that the run-off rates they experienced during the financial crisis were materially lower than the Committee's proposed run-off rates. This, they argued, demonstrated that the Committee was being excessively conservative. (See e.g., JPMorgan, passim)
The problem with this is that it's a really, umm, stupid argument. Yes, the run-off rates were probably lower during the financial crisis, but there were also massive government bailouts during the financial crisis. After Lehman failed, the market only made it 2 days without a government bailout — the Fed rescued AIG on Tuesday night, and Schumer leaked that the government was planning a system-wide bailout on Thursday. Regulators were kinda sorta hoping that we could do the next financial crisis without massive government bailouts.
Okay, that's enough on Basel III for right now.
Add: It just occurred to me that I didn't address the "expected cash inflows" aspect of the LCR. (Cash inflows are subtracted from cash outflows to arrive at "net cash outflows" for the 30-day window.) It's not that big of a deal though, because the rule for expected cash inflows is basically this: You get no cash inflows for 30 days, and you will like it.