Last week I described the traditional tools of the Fed (open market operations and the discount window) and an old, but less well-known one (repurchase agreements). Then I described the first innovation, the Term Auction Facility, inaugurated in December.
This week I’ll go over the forms of lending introduced in 2008, and then I'll discuss the options that the Fed is rumored to be considering next.
* * *
In 2007 the Federal Reserve made an effort to provide liquidity through channels other than open market operations and repos. To that effect, it created the Term Discount Window Program (TDWP) and the Term Auction Facility (TAF), as I explained
last week. Both of those facilities, however, are available only to depository institutions.
So far I’ve been using the ambiguous term “banks” to refer to institutions that borrow funds or buy Treasurys from the Fed. There are however two broad classes of “banks”: depository institutions and primary dealers. Depository institutions are allowed to accept deposits. Primary dealers, on the other hand, are investment banks and brokers that trade in Treasurys with the Federal Reserve. Bear Stearns and Lehman Brothers are two examples in the latter group. As of today, there are
20 of them.
One defining characteristic of depository institutions is that they can use a broad range of assets to secure their loans from the Fed. The discount window, the TDWP and the TAF all accept a set of assets known as “discount collateral.” That includes pretty much all paper of investment quality, including performing sub-prime mortgages. Primary dealers, on the other hand, only have access to open-market operations (OMOs) and repos. The latter only can be obtained after posting General Collateral —that is paper issued by the Treasury or US agencies only.
Following problems in the mortgage and real estate markets last summer, primary dealers found it increasingly hard to obtain short-term financing because nobody would take their suspicious assets as collateral —or would do so only at very high prices. The Fed stepped up to the plate by opening the Term Securities Lending Facility (
TSLF) to primary dealers, on March 27. Roughly speaking, a TSLF loan is an exchange of risky securities for Treasuries for 28 days between Federal Reserve and primary dealers. The range of
acceptable collateral, although not as wide as at the discount window, includes some types of paper issued by non-agency institutions (AAA/Aaa-rated private label RMBS and CMBS).
To be sure, the Fed has had a
securities lending program for a number of
years. The novelty of the TSLF is that it extends the range of acceptable collateral beyond Treasuries. A second novelty is that the term of the loans increases from overnight to 28 days.
Unlike the other tools I have discussed, the TSLF does not have an effect on reserve balances by design. This allows the Fed to pursue its recent strategy of providing liquidity to the banking system without increasing the monetary base.
This is what these loans would look like on the balance sheet of the Fed:
Changes in the Fed's balance sheet after a $1,000M TSLF loan
|
Assets
| US government securities
| -1,000
|
Repurchase agreements
| 0
|
Reverse repurchase agreements
| 0
|
Direct loans
| 0
|
TSLF loan
| +1,000
|
Other assets
| 0
|
Liabilities
| Currency in circulation
| 0
|
Reserve balances
| 0
|
|
In March the Fed inaugurated a second form of lending: the Primary Dealer Credit Facility (
PDCF). This venue provides overnight cash loans to all primary dealers, at the same interest rate as the discount window does, and by pledging the same type of collateral. With the PDCF the Federal Reserve has de facto opened the discount window to primary dealers.
PDCF loans increase the monetary base (read the
FAQ). Because this facility is meant to oil the credit market, not to provide a monetary stimulus, the Fed will continue to offset the increase in reserves using "a number of tools, including, but not necessarily limited to, outright sales of Treasury securities, reverse repurchase agreements, redemptions of Treasury securities, and changes in the sizes of conventional RP transactions." Here's what a PDCF loan looks like, after it has been offset:
Changes in the Fed's balance sheet after a $1,000M PDCF loan, offset by an open market operation
|
Assets
| US government securities
| -1,000
|
Repurchase agreements
| 0
|
Reverse repurchase agreements
| 0
|
Direct loans
| 0
|
PDCF loan
| +1,000
|
TSLF loan
| 0
|
Other assets
| 0
|
Liabilities
| Currency in circulation
| -1,000 + 1,000
|
Reserve balances
| 0
|
|
The composition of the Fed’s assets has changed substantially over the last nine months. Here’s the balance sheet of the Fed again, in December and March:
Federal Reserve's balance sheet, $ millions
|
| Assets |
| Aug. 15, 2007
| Mar. 19, 2008
|
US government securities
| 789,601
| 660,484
|
| Repurchase agreements | 24,000
| 62,000
|
| Reverse repurchase agreements | -31,941 | -46,143
|
Term Auction Facility loans
| 0
| 80,000
|
Primary Dealers Credit Facility
| 0
| 28,800
|
| Direct loans | 264
| 125
|
| Other assets | 37,058
| 36,603
|
| Liabilities | Currency in circulation | 813,085 | 818,362
|
| Reserve balances | 5,897
| 3,507
|
|
Source: Federal Reserve, H.4.1 release.
With its new tools, the Fed has provided liquidity without printing much money. In a way, the Fed has become a pawnbroker.
The future?
Loans to commercial banks and primary dealers, from one facility or another, represent now a much larger fraction of assets (see chart, from the Wall Street Journal). The fraction of Treasurys has declined to 53% from 87%.
The concern now is that the Fed may run out of Treasurys. In theory, the Fed could continue extending loans indefinitely. The problem is that, with no Treasurys left over, the Fed would not be able to offset expansions of the monetary base, as it’s been doing for months. Reserve balances would balloon, pushing down the federal funds interest rate to zero. So the Fed is now pondering the following alternatives:
1) Purchase mortgage-backed securities directly —as opposed to taking them as collateral, as it does through the discount window programs and the PDCF. The Fed could finance such purchases by selling Treasurys, and in that case reserve balances would not be affected. But the amount of Treasurys in the Fed’s balance sheet is, as I said, limited and shrinking rapidly.
2) Have the Treasury issue more debt than it needs and deposit the cash at the Fed. The extra cash would be separate from reserve balances, and thus a priori wouldn’t have any impact on the fed funds rate. The Fed would use that cash to purchase Treasurys. This is what this maneuver would look like:
Changes in the Fed's balance sheet after taking $1,000M worth of Treasury deposits, after an issue of "unnecessary" Treasurys
|
Assets
| US government securities
| +1,000
|
Repurchase agreements
| 0
|
Reverse repurchase agreements
| 0
|
Direct loans
| 0
|
PDCF loan
| 0
|
TSLF loan
| 0
|
Other assets
| 0
|
Liabilities
| Currency in circulation
| 0
|
Reserve balances
| 0
|
Treasury deposits
| +1,000
|
|
While lending conditions don't improve, the new funds would soon turn into loans to banks, so the actual effect on the balance sheet would be (assuming the funds are loaned through the discount window; other forms of lending would affect different lines in the asset side of the balance sheet):
Changes in the Fed's balance sheet after taking $1,000M worth of Treasury deposits, after an issue of "unnecessary" Treasurys
|
|
Assets
| US government securities
| 0
|
Repurchase agreements
| 0
|
Reverse repurchase agreements
| 0
|
Direct loans
| +1,000
|
PDCF loan
| 0
|
TSLF loan
| 0
|
Other assets
| 0
|
Liabilities
| Currency in circulation
| 0
|
| Reserve balances | 0
|
Treasury deposits
| +1,000
|
|
This would change the way we view sovereign debt. Traditionally, the government’s power to raise taxes and set public expenditures have determined the creditworthiness of sovereign debt. With this plan, the value of the government’s debt obligations would become contingent on the portfolio of dodgy securities that the Fed accepts as collateral.
3) Let the Fed issue its own debt. The Fed would use the funds to purchase securities or make loans. A new entry would appear in the list of Federal Reserve’s liabilities:
Changes in the Fed's balance sheet after issuing $1,000M worth of its own debt
|
|
Assets
| US government securities
| 0
|
Repurchase agreements
| 0
|
Reverse repurchase agreements
| 0
|
Direct loans
| +1,000
|
PDCF loan
| 0
|
TSLF loan
| 0
|
Other assets
| 0
|
Liabilities
| Currency in circulation
| 0
|
| Reserve balances | 0
|
Federal Reserve bonds
| +1,000
|
|
4) Remunerate reserves. Reserve balances are like checking accounts: they don’t earn interest. For that reason banks have little incentive to hold more reserves than they need to meet the Fed’s requirements and clear transactions. Any excess reserves are loaned to other banks. As Greg Ip explains, “if the Fed paid, say, 2% interest on reserves, banks would have no incentive to lend out excess reserves once the federal funds rate fell to that level.”
This measure would lead to a higher equilibrium level of reserve balances, for a given value of the federal funds interest rate. It would also reduce the amount of inter-bank lending, as banks would keep more of their cash in their safe-deposit box at the Fed. That lending would be replaced by loans from the Federal Reserve.
This reviews my review of the Fed's new monetary policy. Will these new tools make it to the textbooks? It’s hard to tell whether the particular facilities (TAF, TSLF, etc.) will survive. But I think that some standardized form of loans to non-depository institutions will stay, and that the Fed will become willing to accept dodgier collateral than it traditionally has.
Subscribe to EconWeekly 
(Clicking that button gives $0.10 to Francisco. Join tipjoy! How does tipping work?)
Technorati tags
economics, monetary policy, Federal Reserve, open market operations, macroeconomics
