On Thursday, the Federal Reserve issued its weekly H.4.1 report, which provides details of the Fed’s balance sheet. Once upon a time, this was one of the least interesting of the government’s many releases of data. These days, it’s become one of the most exciting.
The essence of the Fed’s balance sheet used to be quite simple. The Fed’s primary operations would consist of either buying outstanding Treasury securities or issuing loans to banks through its discount window. It paid for these transactions by creating credits in accounts that banks hold with the Federal Reserve, known as reserve deposits. Banks can turn those reserves into green cash any time they desire, so the process is sometimes loosely summarized as saying that the Fed pays for the Treasury bills it buys or loans it extends by “printing money”. Before the excitement began, the Fed’s assets consisted primarily of the Treasury securities it had acquired over time (about $800 billion as of August 2007) plus its discount loans (an insignificant number at that time). Its liabilities consisted primarily of cash held by the public (about $800 billion a year ago) plus the reserve deposits held by banks (which again used to be a very small number).
Bernanke’s overriding goal since then has been to extend a huge volume of short-term loans to financial institutions. If he’d done that in the usual way, just creating new reserve deposits with each new loan, the supply of cash would have ballooned, bringing worries of inflation. The Fed didn’t want to do that, and in fact there was no shortage of funds available for overnight interbank lending. The fed funds rate, an average overnight lending rate between banks, is already quite low, and further reductions seem unlikely to accomplish much. But longer term interbank lending rates remain quite high relative to the overnight rate.
Bernanke’s first approach to this challenge was to “sterilize” the new loans from the Fed, basically selling off the Fed’s Treasury holdings at the same time that it extended the new loans. When a counterparty buys the Treasury security from the Fed, the Fed debits the bank’s account with the Fed, and these debits net out the credits that would be created as a consequence of the Fed’s new loans. Reserves go up with the loans, down with the sale of Treasuries, so the net result is an increase in loans from the Fed but no change in reserve deposits. These new Federal Reserve assets came in many colors and flavors, including the Term Auction Facility, the Primary Dealer Credit Facility, currency swaps (which I presume is the biggest single item in the burgeoning “other F.R. assets” category), and the seriously non-acronymizable Asset-Backed Commercial Paper Money Market Mutual Fund Liquidity Facility.
|Aug 8, 2007||Sep 3, 2008||Oct 1, 2008||Oct 22, 2008|
|    Discount window||    255||    19,089||    49,566||    107,561|
|    TAF||    150,000||    149,000||    263,092|
|    PDCF||    146,565||    102,377|
|    AMLF||    152,108||    107,895|
|    Other credit||    61,283||    90,323|
|    Maiden Lane||    29,287||    29,447||    26,802|
|Other F.R. assets||41,957||100,524||320,499||519,713|
|Factors supplying reserve funds||902,993||939,307||1,533,128||1,839,042|
|Currency in circulation||814,626||836,836||841,003||856,821|
|Factors absorbing reserve funds||902,993||939,307||1,533,128||1,839,042|
|Off balance sheet|
|Securities lent to dealers||120,790||259,672|| 226,357|
But $800 billion– the total stock of Treasuries that Bernanke originally had available for this purpose when he started down this path over a year ago– only goes so far these days, particularly when you remember that a quarter trillion of those securities are now being used in the Term Securities Lending Facility, which the Fed records as an off-balance-sheet transaction. To enable it to extend more than $800 billion in loans without “printing more money,” the Fed asked the Treasury to implement a
Supplementary Financing Program in which the Treasury would sell securities directly to the public but simply keep the funds in an account with the Fed. The payments by the public for these securities then initiate a flow of reserves out of private banks, the same as if the Fed itself had sold Treasuries to the public out of its own holdings, so the SFP enables the Fed to sterilize a greater volume of loans than it could if it had to rely solely on its original holdings of Treasury securities. The Treasury supplementary account as of last week has provided the Fed with an additional $559 billion to play with. It appears from the latest balance sheet that the Fed has now asked the Treasury to do the same sort of thing with the Treasury’s “general account” with the Fed. Historically, that account was just used to facilitate daily Treasury transactions, and was usually held to about $5 billion. Last week, it’s up to $56 billion.
It’s clear that the Fed is now also using yet another tool to balloon its balance sheet, namely, deliberately encouraging banks to sit on their excess reserve deposits. When these started to shoot up at the end of September, I initially attributed this to frictions in the interbank lending market. But with the announcement on October 6 that the Fed would begin to pay interest on those deposits, and the further announcement on October 22 that the Fed is now raising that interest rate to within 35 basis points of the target for the fed funds rate itself, it is clear that the Fed has now settled on a deliberate policy of encouraging banks to just sit on the reserves it creates, giving it another device with which to expand its balance sheet without increasing the quantity of cash held by the public. That’s provided another quarter trillion for the alphabet soup of new facilities.
I had a call from a reporter this week asking me to explain why the Fed raised the interest rate paid on reserves. I think she was expecting a 30-second sound bite, but instead we went back and forth for about 15 minutes and I’m not sure even then that I succeeded in getting the basic idea across. At that point she asked me, “Do you see it as an encouraging development that the Fed has taken this step to address the credit crunch?” My immediate answer was no. It’s not an encouraging development because it means that the heroic efforts that the Fed has taken previously weren’t enough. The Fed’s first $100 billion didn’t do it. The Fed’s first $1 trillion didn’t do it. Having the Treasury take over the $5 trillion in debts and guarantees of Fannie and Freddie didn’t do it. The Treasury’s $3/4 trillion rescue/bailout package didn’t do it. And another quarter trillion will?
If the spread between overnight and 3-month interbank lending rates indeed results from pure illiquidity of the latter market, it seems to me it shouldn’t have required too much grease to get that market lubricated. But if, as argued by John Taylor and John Williams, the spread instead represents compensation for counterparty risk, it doesn’t matter how much term lending the Fed does. Its actions would only move that spread to the extent they reduce the counterparty risk itself. The primary consequence of the actions would not be to change the spreads but instead would just shift the risk onto the Federal Reserve’s balance sheet.
There was another juicy morsel in the latest H.4.1. The latest report acknowledges that the Fed has taken some losses on some of these unconventional assets. The Fed last week wrote off $2.7 billion in losses on the loans to “Maiden Lane LLC,” an entity created through the Bear Stearns package. The assets of Maiden Lane consisted of claims on certain troubled securities, and the liabilities consisted of a loan from the Federal Reserve. The Fed now admits that Maiden Lane won’t be repaying all of the loan, so it had to reduce its claimed assets by $2.7 billion. This also required a corresponding imputed $2.7 B reduction in the “other” category on the liabilities side of the Fed’s balance sheet, presumably in large part coming from debits to the “surplus” and “other capital accounts” entries in the Statement of Condition of the Federal Reserve Bank of New York, though I haven’t traced through the details of exactly how that was implemented.
Regardless of the accounting, here’s how those losses will show up in practice. When the Treasury auctioned the T-bills for the increase in its supplementary and general accounts with the Fed, and when the Fed sold off its existing holdings of Treasuries, the Treasury started making interest payments to the public. The Fed is also receiving interest on the loans it made, and returns that interest to the Treasury. As long as the loans are performing, it is a wash to the Treasury. But if some of the Fed’s loans go bad, it means the Treasury is on the hook for the extra interest costs with no offsetting receipts. In other words, any losses by the Federal Reserve are equivalent to a fiscal expenditure financed by Treasury borrowing.
The notes to the H.4.1. seem to imply that the Fed’s intention is to update its assessment of the “fair value” of its Maiden Lane holdings as of the end of each quarter, which would mean no new markdowns until January.
But that doesn’t mean that the remaining $1,839 B in Fed assets will all continue to bring in their hoped-for receipts for the Treasury between now and then.