Don't Discount the Fed Discount Window

Though many have dismissed the Federal Reserve's decision to drop the discount rate from 6.25% to 5.75% as mere window dressing, there are reasons to think otherwise. Citigroup, JPMorgan Chase, Bank of America and Wachovia each borrowed $500 million from the Federal Reserve discount window last week. Average end-of-day Fed credit outstanding for the week ending September 12 measured $2.9 billion, the largest balance loaned out via the discount window since September 2001 and the most since the discount window was reformed in 2003.[1]

Of interest too are the sorts of collateral the Fed accepts nowadays from banks in return for loans. The list reads like a Who's Who of the investment world. The regulars turn up: US treasuries, government agency debt, and foreign government debt. But the Fed can accept far more than that. Below is just a sample:[2]

Of particular interest is the Fed's willingness to accept securities with no market price, including non-AAA-mortgage-backed securities and asset-backed securities, loaning up to 90% of these securities' calculated values. When did the Fed get the authority to accept such a wide range of collateral? With the discount rate falling and problems emerging in the American housing sector, what might be the consequences? Before we try to answer these questions, we offer a few preliminary explanations.

The Difference Between Open Market Operations and the Discount Rate

The federal funds rate is the interest rate at which banks lend each other funds in order to meet reserve requirement balances maintained by Fed regulations. This rate is largely determined via market bids and offers, except when the Fed steps in, usually in the mornings and in odd cases during the day, to intervene and nudge the rate towards its policy target.

The Fed nudges via open market operations. If the federal funds rate is above the Fed's target, for instance, the Fed will buy securities from banks participating in the federal funds market, paying them in cash. At the same time, it agrees to sell these securities back to the banks (usually the next day) for a set price above what it bought at. This is called a repurchase agreement, or a repo. The difference in price between the purchase and sale — a small profit in the Fed's favor — is the effective yield at which the Fed offers funds. By ensuring that the yield it offers is below the market rate for federal funds — i.e., the rate at which banks lend to each other — banks will turn to Fed open market purchases for funding rather than the federal funds market. Market rates will thereby fall towards the Fed's target. Rules concerning Fed buying and selling are contained in Section 14 of the Federal Reserve Act.

The discount rate refers to the rate at which the Fed lends funds to institutions, accepting various sorts of collateral from these institutions as security. Sections 10A, 10B, 13, and 13A of the act govern the rules for lending. Prior to 2003 the Fed set the discount rate a fraction below the federal funds target. A bank looking to borrow could either seek funds from other banks in the federal funds market, sell securities for cash to the Fed during its open market operations, or pledge collateral in order to borrow at a cheaper rate afforded at the discount window. What prevented banks from flocking to this cheap rate were regulations requiring institutions to first exhaust other available sources of funds and explain their need for loans. Borrowing at the discount window became a sign of weakness, one that institutions avoided at all costs. The last significant episode during which the banks turned to the discount window to borrow was September 12, 2001, when a record $45.5 billion was lent.

The Reformation of the Discount Window

In January 2003 the Fed reformed the discount window. Rather than being set below the federal funds rate, the discount rate was to be set above. Eligible institutions no longer needed to show they had exhausted other sources of funds before coming to the discount window; nor were there any restrictions on what the borrower could use the credit for. As long as institutions were willing to pay 1% more than the going rate in the federal funds market, the window was permanently open to them. From 2003 to now, the average weekly balance of lending through the revised window rarely exceeded a paltry $50 million or so. After all, why borrow at such a high penalty?

Together, the federal funds target and the new discount rate (often called the primary rate) are the Fed's tools to keep a cap on short-term interest rates. Open market operations influence the federal funds rate but typically only do so in the morning. After the Fed is done buying securities, banks desperate for short-term financing may drive the market-determined federal funds rate much higher than the Fed's target. If this rate rises high enough, at some point the discount window (typically set 1% above the federal funds target) will begin to look attractive for financing. Banks' demand for funds will be satiated at the discount rate, effectively capping the federal funds rate.

The Fed Can Accept Anything as Collateral

Historically, the Fed has been able to accept many types of collateral for discount loans. Sections 10A, 10B, 13, and 13A of the Federal Reserve Act outline the sorts of collateral allowed.[3] It is from these sections that the list of eligible securities at the beginning of this article has been derived.

Discount loans, specified in the above sections, can often result in a temporary expansion of the money supply, though it is not necessary that they do so. Section 16 of the Federal Reserve Act maintains that all US dollars must be collateralized. This means that the Fed cannot simply "print" dollars and lend them out. It must accept something in return as collateral for each dollar it prints. When the Fed expands the money supply via open market operations, for instance, each new dollar it spends into the economy is matched by the inflow of a corresponding security which is subsequently held in reserve by the Fed. If it expands the money supply via discount loans, each dollar it creates and lends must be backed by pledged collateral, again held by the Fed. For every dollar created by the Federal Reserve, a corresponding asset exists in Fed inventories.

Not all of the collateral mentioned in 10A, 10B, 13, and 13A have been permitted to act as backing for new dollars, i.e., to expand the monetary base. Section 16 of the act limited the sorts of collateral that could be used to back newly issued US dollars to the securities mentioned in Section 13 of the act: government securities, agency securities, bills of exchange, bank drafts, and bankers acceptances. Assets like corporate bonds (implied in Section 10B) and agricultural loans (mentioned in Section 10A) could be accepted as collateral by the Fed for loans, but could not be used to back US dollars and therefore expand the money supply. What this meant was that upon lending out new dollars at the discount window and accepting, say, corporate bonds as collateral, the Fed was required to simultaneously remove the same amount of dollars from the financial system by selling US treasuries from its surplus account holdings into the market. If it didn't carry out this offset, the Fed would be in violation of Section 16, since it would have issued new dollars without collateralizing them appropriately. Thus, over the years many of the loans made through the discount window have been "sterilized." Though they provided the targeted borrower with temporary funds, these interventions did not actually expand the overall money supply.

This changed in 1999 with the rewriting of Section 16. In only a few words, regulators vastly expanded the Fed's powers so that not only collateral mentioned in Section 13 could back the dollar, but also those in Sections 10A, 10B, and 13A.[4] No longer was it necessary for the Fed to exercise discipline and offset the effect of these loans on the money supply. Like the changes made to open market operations around the same time, lawmakers justified the rewriting of the rules as an appropriate reaction to the impending Year 2000 (Y2K) problem and the potential liquidity problems that might be experienced.

The Y2K crisis was a no-show. The lights didn't go out, nor was there a bank run. If the Fed had been honest about its motivations, it would have reverted back to its old rules. Needless to say it didn't. With the change in Section 16, the Fed has been given almost unlimited power to expand the money supply via the discount window lending. The inclusion of Section 10B in the mix, one of the broadest sections in the act, is the source of most of this new power:

Section 10B(a): "Any Federal Reserve bank, under rules and regulations prescribed by the Board of Governors of the Federal Reserve System, may make advances to any member bank on its time or demand notes having maturities of not more than four months and which are secured to the satisfaction of such Federal Reserve bank."

Section 10B's only restriction is that collateral offered must be deemed "satisfactory" by the Federal Reserve bank making the loan. This definition allows for almost any sort of security to act as collateral, and accepts them for as long as four months. Section 10B was added temporarily to the Federal Reserve Act in 1932 and permanently in 1935, at the height of the Great Depression. According to Small and Clouse [PDF], prior to 1932, the Fed had only been able to accept bills of exchange, notes, and drafts at the discount window. Lending activity in the early 1930s was so slow that regulators sought legislation that would open the door for anyone looking for funds. Section 10B was worded vaguely enough to open the door quite wide.

Even after the passage of 10B the Fed's powers were limited. Collateral specified under 10B could not be used to back currency and expand the money supply. All cash lent out had to be balanced by sales of treasuries from the Fed's surplus account to other participants in the financial system, the cash from this transaction flowing back to the Fed and canceling out the cash it had lent out. For some 65 years 10B's scope was limited, until the 1999 rewriting of Section 16.

Where we are at now with the discount window is the culmination of years of expansion of the Fed's power: specifically, the introduction of the Federal Reserve Act in 1913, the 1932 addition of 10B, the 1999 modification of Section 16 to "point" to 10B, and the 2003 reformation of the discount rate. A whole smorgasbord of assets can now be accepted by the Fed to extend loans and expand the money supply, and eligible institutions no longer need to prove that they have gone elsewhere before applying for Fed loans. They are meant to do so "guilt free."

What Might Happen?

The discount rate now sits only half a percent above the overnight federal funds rate target of 5.25%, the narrowest margin since the window was revised in 2003. It could go lower, too; there is no regulation governing how high above the overnight target it must be set. A low rate will only entice firms to turn to the window for funding, a trend we are starting to see with the record amounts of primary credit now outstanding. Unfortunately we don't get to see statistics showing what sorts of collateral the Fed has accepted from banks like Citigroup for these loans. This collateral is probably comprised of safe government securities. But it could also be sub-AAA mortgage- backed securities (MBS) with no available market price, the Fed loaning 90% of saidthese securityies's value to the banks. And if the Fed hasn't accepted this sort of collateral yet, it could in the future.

If sub-AAA MBS collateral pledged to the Fed were to fall in value while the Fed was holding them, the bank that deposited the MBS and took the loan would be required to deposit additional collateral or reduce its loan balance to an appropriate amount. In the case of securities with no market price, the need for additional topping up could be difficult to determine. In theory though, most of the risk associated with falling collateral values lies with the bank accepting the loan, not the Fed.

The main risk to the Fed would be if the borrowing bank were to go bankrupt during the course of the loan. Under such a scenario the Fed would take possession of the collateral and sell it in the open market to recuperate the funds it had lent the failed bank. If the market price for the collateral was significantly below the amount the Fed had originally lent out, the sale would not withdraw the total amount of dollars originally created by the loan. This is problematic because it would put the Fed in contravention of Section 16 of the act, specifying that all dollars must be properly collateralized. In other words, the Fed would have lent a certain quantity of new dollars into the economy, but with the sale of the collateral it would have withdrawn only a portion of this amount, leaving a large chunk circulating with no backing. In order to comply with Section 16, the Fed would be required to withdraw this amount of money from the financial system by selling a corresponding amount of treasuries from its surplus.

In a situation in which the Fed exposed itself to significant quantities of iffy collateral and multiple institutions refused to honor their obligations, the Fed would be required to sell massive amounts of treasures in order to withdraw unbacked cash from the financial system, in the process drastically reducing the money supply and making an already precarious situation worse. If losses were large enough and the Fed exhausted its surplus of treasuries, the temptation to rewrite Section 16 would be overwhelming. US dollars would lose their backing requirement and could be loaned into the system for nothing.

It is difficult to determine how the government will react to the growing housing problem in the United States. If regulators choose to move, it could be institutions other than the Fed that bail out borrowers and lenders. As our story shows, the Fed has more power than ever to use discount window lending, and should the powers that be decide to initiate a bailout, a potent combination of old rules and newly created ones gives the Fed the ability to carry one out. After all, the Fed has bailed out institutions before: banks in the 1930s, Franklin National for five months in 1974, Continental Illinois for sixteen months in 1984–85, and thrifts during the Savings & Loans crisis of the late '80s. Getting the money to those who need it may prove difficult — it is hedge funds and their ilk in trouble, not big commercial banks with discount window access. Recently announced relaxations of certain regulations, including Section 23A of the act, might help guide funds to those who have invested in suspect MBS.

Given the wide variety of assets it can accept (in particular, non-AAA MBS), the Fed also has the potential to get itself into plenty of trouble, the sort of trouble that might require it to seek an even larger expansion of its powers. This expansion would be nothing more than a continuation of a trend started back in 1913. As for the inflation and moral hazard that would be created by issuing new money to bail out such a large number of irresponsible players, that is a story entirely to itself.

John Paul Koning is a freelance writer based in Toronto, Canada. See his archive. Send him mail. Comment on the blog.

Notes

[1] Data for discount lending comes from the Federal Reserve.

[2] For the full list of eligible collateral along with descriptive notes, see FRBDiscountWindow.org.

[3] The Federal Reserve Act is available at FederalReserve.gov.

[4] The act passed in Congress is available here: The Supreme Court of Oklahoma.

John Paul Koning