Sunday, February 10, 2008
negative bank reserves: a non-issue?
Reserves can be borrowed (from the Fed's discount window) or non-borrowed (supplied via the Fed's daily open market operations). It matters not one whit to the Fed where the banks acquire the reserves they require. If they borrow directly from the Fed, they don't need to tap the interbank, or fed funds, market.
It isn't a mystery what happened. The Fed announced the creation of a Term Auction Facility on Dec. 12, enabling banks to borrow for 28 days versus a wide range of collateral. The minimum bid the Fed accepts is the expected funds rate one month out, which in the current environment means cheaper funding costs than the fed funds market.
So what would you do if you were a bank?
Loans made through the TAF are categorized as borrowed reserves. The Fed had $50 billion of loans in place at the end of January, which ``caused the borrowed reserves figure to balloon and the non-borrowed figure to decline by a corresponding amount,'' said Lou Crandall, chief economist at Wrightson ICAP LLC in Jersey City, New Jersey, in a Feb. 6 commentary. (He's on the same e-mail lists I am.)
the Fed is ``a monopoly provider of reserves,'' said Jim Glassman, senior U.S. economist at JPMorgan Chase & Co. ``This is a non-starter. There is no such thing as a banking system short of reserves. The Fed has absolute control over the supply.''
There may be times, such as late last year, when banks are reluctant to lend to one another for a period longer than overnight. ``And any one bank can have a problem'' funding itself, Glassman said. But in a world where ``the Fed can print money, there is no shortage,'' he said. ``The banks get the reserves they want.''
in other words, what is happening would not at all be this:
american banks in aggregate have burned through all their free capital
or rather, it may be something approaching that as losses are gradually realized, but it is essentially unrelated to this data. moreover, this:
the fed is literally maintaining the static solvency of the system at this moment
... would be true everywhere and always in a fractional reserve banking system. though this:
if it were to withdraw its anonymous auction facility, it's likely that asset sales and bank failures would begin more or less immediately
... could only possibly be true, it isn't necessarily -- if the TAF were withdrawn, suggests baum, banks would very probably instead obtain needed reserves through either the discount window or the fed funds interbank market (with the attendant dangers) but at a higher price than the TAF affords.
ergo, another object lesson in support of the point jeff of dash of insight made recently. baum might agree that there's plenty of things to worry about in american banking at the moment. however, she would say that this isn't one of them.
mish strongly disagrees and he rightly points out that banks are paying punitive rates for capital.
the problem with mish's argument seems to me to be that capital is not the same thing as reserves. reserves are simply some fraction of the deposits in a bank set aside with the federal reserve bank -- an asset that offsets the liability of deposits.
mish is making really strong points about bank capital being under pressure from unexpected losses on the asset side and lending capacity being under pressure from losses, increasing loan loss reserves and the anticipation of more to come -- but they don't translate to reserves per se.
banks borrow large over a small base of equity, and borrow cheaply (and usually short) to lend expensively (and usually long). this dynamic is what makes them profitable, when it works.
banks can temporarily borrow from the federal reseve against some of their assets, and use the borrowed cash as reserves -- translating portfolio loans or securities (an asset) into cash (an asset). the fed charges a discount for doing so. this is one way the fed can liquify banks against a run, known as a repurchase agreement (or repo), through the discount window.
banks can also sell hgih-quality securities from their portfolio (an asset) to the fed in their open market operations for cash (an asset). this is another way the banks can be liquified by the fed.
banks can also borrow from other banks, generally at the fed funds rate, which the federal reserve dictates. normally, the fed funds rate is cheaper than the discount rate, which encourages interbank lending. this avenue has, however, been a partial casualty of the credit crunch, which has increased the necessity of borrowing from the fed.
reserves held at the fed can consist of funds borrowed through repos, funds borrowed at interbank fed funds rate, or funds obtained by selling securities to the fed by open market operation. all three are essentially transformation of asset type conducted by interaction with the fed, but the fed accounts for repos as "borrowed reserves", whereas interbank funds and cash derived from asset sales placed with the fed are termed "non-borrowed".
when the TAF -- a different kind of repo -- was introduced in december, participating banks were being granted a chance to borrow more cheaply and anonymously against a wider range of collateral (ie, a larger part of their portfolio) than they could through either the discount window OR the fed funds market.
so many of them chose to borrow funds via the TAF -- enough to drive net non-borrowed reserves negative by the mechanism described by baum, as repos grew significantly and interbank fed funds lending sank.
NONE of this is to say banks have no capital. it is instead to say that banks have gone from borrowing on the interbank market to borrowing from the fed in a big way, encouraged by the very low rates on offer through the TAF.
there is a limit to what the fed can do to liquify the banks. as mish points out, banks must have high-quality assets to sell or provide as collateral to the fed. when a bank has sold or offered as collateral all it can, it has hit the limit of what the fed can do to help it and faces insolvency. but few banks are anywhere near that point right now -- though that may change as asset quality deteriorates, particularly if a heavy CRE bust follows a deepening residential housing bust.
moreover, this is not a capital issue -- it is a liquidity issue. the capital issue arises when losses on assets cause liabilities (eg, deposits) to exceed assets (eg, loans). then it is the FDIC, not the fed, which steps in.