$7.77 trillion in secret Federal Reserve loans to banks?

I have been looking into the claim recently made by any number of internet sites (for example, here’s one of the many hundreds, if you insist on a link) that the Federal Reserve made $7.77 trillion in secret loans to banks. The claim is outrageously inaccurate, as I explain below.

Let me begin with some accounting basics. Suppose that at the start of January I make a 3-month loan of $100 to person A and a 1-month loan of $100 to person B. At the start of February, person B rolls it over into a new 1-month loan, and does so again at the beginning of March. On the first day of April, person A and person B both repay me the original $100. So, students, here’s your question: how much did I lend to person A, and how much did I lend to person B?

The correct answer, of course, is that I lent $100 to person A and I lent $100 to person B. But, if you were trying to sensationalize and misrepresent what actually happened, perhaps you’d say that I lent $300 to person B, by adding the three $100 1-month loans together.

This is a very elementary point in economics or accounting. A loan is what we refer to as a “stock” variable. It’s measured in units of dollars at a particular point in time. It is a completely meaningless exercise to take outstanding loan amounts at different dates and add them up as if it’s one big total.

On the other hand, if your goal is to come up with a number that sounds really big, you’ll be excited to learn that I also lent $100 to person C in the form of a series of daily loans. These were rolled over for the same 3 months, so someone with a sufficiently bizarre theory of accounting (or a sufficiently strong political agenda) might claim that I lent $9,000 to person C.

So where in particular did people come up with this $7.77 trillion figure? The source appears to be a recent story from Bloomberg news, which includes the following statement:

Add up guarantees and lending limits, and the Fed had committed $7.77 trillion as of March 2009 to rescuing the financial system.

I contacted the reporters who prepared the Bloomberg story to try to learn some more details. They communicated to me that those who claim that the Fed provided $7.77 trillion in secret loans to banks have misinterpreted their article. Specifically, they clarified that the $7.77 trillion number was not intended to represent loans the Fed actually made, but instead refers to loans it potentially might have made, that the number refers not to loans to banks but to the broader financial sector, and that Bloomberg’s use of the term “secret” in describing these loans refers not to the total amount but instead to the specific identities of the recipients.

The source for the $7.77 trillion figure turns out to be this Bloomberg article from March 31, 2009. At the end of that second article is a table that breaks down what it calls the “limit” on various categories of Federal Reserve lending.

According to that table, the biggest single component in the $7.77 trillion total is a $1.8 trillion item labeled “net portfolio CP funding”. This apparently refers to the Commercial Paper Funding Facility. The $1.8 trillion was evidently calculated by Bloomberg by taking the formula for the maximum amount of commercial paper that the Federal Reserve said it would be willing to buy from any one institution and assuming that the Fed in fact purchased this maximum amount from every single eligible institution. As actually implemented, the maximum outstanding balance ever reached by the CPFF was $350 B on January 21, 2009. That was all repaid, and the CPFF outstanding balance has been zero since February 2010.

The second biggest component in the $7.77 trillion is $1 trillion in mortgage-backed securities. Here the Fed actually did end up buying even more than this. But these purchases were made when the CPFF and other items included in the $7.77 trillion were being wound down. Adding together the $1 trillion in MBS held in February 2010 to the $350 B in CPFF loans in January 2009 (or, even sillier, to this $1.8 trillion CPFF figure) is the kind of nonsensical calculation with which I began my discussion. Furthermore, these are agency MBS, not those issued by private banks. It was the U.S. Treasury, not the Federal Reserve, that had already taken over the guarantees of these MBS before the Fed bought them. There is no sense in which the Fed’s purchase of these MBS could be construed as a loan to banks.

The third biggest item in the $7.77 trillion figure is the Term Auction Facility, whose “limit” entry in the Bloomberg table is $900 billion. Apparently the source for this number was the statement issued by the Federal Reserve on October 6, 2008 that “$900 billion of TAF credit will potentially be outstanding over year end.”

There were never any secret commitments associated with the TAF. The way the program worked was the Fed would decide how much additional reserves it wanted to add to the system, invite bids in the form of interest rates banks were willing to pay to borrow this fixed quantity of funds, and lend the prespecified quantity to the highest bidders. That’s why it was called an “auction” facility. The Fed never lent anywhere near $900 B through this program. The maximum amount ever reached was $493 billion, the outstanding balance as of March 11, 2009. The loans were all repaid, and the outstanding balance has been zero since April 2010.

Moreover, there was never anything secret about any of these numbers. They were all published continuously each and every week in the Fed’s H.4.1 statement, and readers of Econbrowser saw their pros and cons actively evaluated as the programs were initially proposed and subsequently implemented. If you are interested in ex-post evaluations of whether programs such as the CPFF and TAF were beneficial, you can consult for example Christensen, Lopez, and Rudebusch (2009), McAndrews,
Sarkar, and Wang (2008)
, Taylor and Williams (2009), Adrian, Kimbrough, and Marchioni
and Duygan-Bump, Parkinson, Rosengren, Suarez, and Willen (2010).

You’re free to take your own position on whether these programs had beneficial effects. But please know that anyone who tells you that the Federal Reserve secretly loaned $7.77 trillion to banks is spreading a lie.

36 thoughts on “$7.77 trillion in secret Federal Reserve loans to banks?

  1. ppcm

    “Qui trop ecorche deux fois ne tond” or “Whom once skinned twice cannot shear”
    It seems, our good professor Hamilton is getting better in Jesuistic and is omitting the miracle of the banks fractional reserves and their multipliers effects called fractional banking.As an illustration, the hereunder stipulations of the ECB when dealing with obligatory reserves.
    “This Regulation ensures that the minimum reserve arrangements applicable to credit institutions and their branches in Member States are the same throughout the euro area. Minimum reserves as a monetary policy instrument are used mainly to stabilise interest rates and to dampen autonomous fluctuations in money market liquidity by adjusting reserve requirements.
    Reserve ratios must not exceed 10% of any relevant liabilities forming part of the basis for minimum reserves but may be 0%.”
    Please note that these reserve are not aiming at risks but liquidities.
    Fractional Reserve Banking as Economic Parasitism
    A Scientific, Mathematical, & Historical Expose, Critique, and Manifesto
    Vladimir Z. Nuri
    “In standard economics and banking r is also called the reserve requirement, and the inverse 1=r is known as the money multiplier. In examples a typical
    rate is given as r = 0:1. gives a banker’s rule of- thumb ratio of 4:1 circulated (debased) receipts vs. deposits corresponding to r = 0:2. These are extremely
    low and in the previous (primitive) model fall into the range r

  2. scott s.

    OK, but GAO report 11-696 provides a:
    “Table 8: Institutions with Largest Total Transaction Amounts (Not Term-Adjusted) across Broad-Based Emergency Programs (Borrowing Aggregated by Parent Company and Includes Sponsored ABCP Conduits), December 1, 2007 through July 21, 2010”
    So it seems that the Comptroller General sees some value in reporting total loan amounts, even though it isn’t the maximum loan amount outstanding at any one time. (That table shows $3.086 Trillion lent to foreign banks out of a total $16.115T.)

  3. fresno dan

    I don’t know
    I respect Waldman quite a bit. Maybe among friends 100 bucks gets loaned (Whoa – maybe that is the Fed and the banks) without regard to interest rates, but how you set up the “loan” and the terms are critically important.
    If 3 month versus 1 day loans dont matter, than why go through the rigamarole of renewing loans daily? Would “one” loan of 1.8 trillion taken care of everything – or was the object to loan as much as necessary?

  4. tinbox

    There should be at least some recognition in this blog post that Bloomberg News had to fight the Fed in court for years to get data through Freedom of Information Act requests. The point wasn’t that the lending programs were secret or that they were fully utilized, but that the identity of recipients of these loans.
    To accuse people of lying for characterizing the capacity of the lending programs as if it were actual lending is a harsh standard. Is it a standard that the author would apply to Mr Bernanke for characterizing the recipients as sound financial institutions?

  5. Ed Hanson

    Professor Hamilton
    What a disappointing article. You take all the fun out of screaming “the sky is falling.”
    So few economist take the time to counter the sensational. Thank you.

  6. Ricardo

    You are absolutely correct. This is not the first time Bloomberg has told a “lie,” to use your word. I have been reading Bloomberg for a number of years and have found that very many of their articles are of this same type. If you do not dig deeper into Bloomberg articles they will often mislead you into claims that are not true.
    As I said I do read Bloomberg but usually just to understand where the lies and distortions come from in the media and leftist web sites. Bloomberg is normally very supportive of central planning programs and so tilts its “lie” in the direction of perpetuating destructive government intervention.
    I listen to their radio program every morning because it is the only program in the area that reports on international financial news; believe me it is hard to listen to European socialists spew their errors early in the morning. Of course when my shoe hits the radio it does help me wake up.

  7. Old School

    Weren’t these ‘loans’ simply borrowings at the Discount Window? Should that be the case, isn’t that what the Fed is supposed to do in its role as lender of last resort?

  8. JDH

    Ricardo: As I tried to make clear above, Bloomberg itself did not state that the Federal Reserve made $7.77 trillion in secret loans to banks. It is the other people who in fact did make that very statement (and there are many of them) that I am calling liars.

    For Bloomberg’s analysis itself, my position is that they have drawn a different inference from these numbers than I would.

  9. Chang Baishan

    I think you don’t even care what is right and what is wrong on a larger scale. What really matter is not what the exact number is ( and don’t pretend you know how to count beans for the FED). The fact of matter is that the FED waged a big printing press bet and even Congress did not know the details. And by keep doing this, the FED may destroy the real economy. You may have histories or whatever you think you should side with FED. But I am appalled that people believe they can count on printing press to support them for ever.

  10. yuan

    The idea that the Fed has been transparent does not hold up to scrutiny. The identity of the banks and corporations that received treasuries or lines of credit in return for dubious collateral has not been released. Nor has the fed identified precisely what type of collateral was received.

  11. Brian

    Absolutely fabulous post, Jim. Too bad many reporters in the media ran with the $7.77 trillion headline number and didn’t do any investigating of their own.

  12. AS

    I wonder if David Wessel read Professor Hamilton’s post. Wessel wrote a very similar article today in the 12-7-2011 edition of the WSJ, page A4.

  13. VK

    Bloomberg article states that those funds were available for the banks, if needed. The actual amount required by the banks was $1.2T. I believe you are confusing money available to banks with money borrowed.

  14. jonathan

    The difference is between fact and political fact. A fact is. A political fact is what you can make.

  15. JDH

    Evan: The GAO went even further down the road from Bloomberg in adding together outstanding loan amounts that in fact were associated with different points in time. As I point out above, you can make this number arbitrarily large if that’s your standard.

  16. don

    O.K. So let’s have a takedown of Chair Bernanke’s Congressional testimony, in which he cited a study by Hufbauer et al on the gains from international trade. That was a lie as big as the Treasury’s denial that China manipulates its currency, and was told for political purposes. Chair Bernanke may be ignorant of trade, but the Board of Governors has a very capable staff of international economists, any one of whom, if asked, would have put the lie to the ridiculous numbers he cited.
    Adding up loans that were rolled over may overstate exposure at any one time (as you note), but is not necessarily a “lie.” For example, if I loan $100 to A, then after being repaid lend the same $100 to B, it seems to me that to say I loaned only $100 is as inaccurate as to say that I loaned $200. As for the Fed buying mortgage debt already guaranteed by Treasury, would you say the Fed is completely innocent of the consequences to taxpayers when they bought part of the debt? Sounds to me more likely that they were in cahoots.

  17. JDH

    Don: OK, suppose I lend $100 to A on Jan 1 who repays April 1. I lend $100 to B1 on Jan 1 who repays Feb 1, then lend that same $100 to B2 who repays Mar 1, then lend that same $100 to B3 who repays April 1. In what sense am I extending 3 times the volume of loans through the B program compared to the A? The total outstanding loans on either program are identical at any given date. Any time you add together the B1 Jan 1 loan balance of $100 with the B2 loan Feb 1 balance of $100, you are creating an object that is no longer measured in standard balance sheet units (which are dollars at a particular point in time). I think it is very important to specify units (dollars at point in time for a stock variable, dollars over a particular interval of time for a flow variable.)

    As for your point about the MBS, I do not understand what you are saying. Given that the Treasury had assumed the liability before the Fed acted, I do not see any way to describe the consequences of the Fed action as being anything other than reducing the burden on taxpayers.

  18. ECON

    Did not the prima facie evidence of largesse arising from Fed not come from thousand of pages from the Fed itself secured through successful Freedom of Information (as stated by the authors) ??
    Today I read a report by Alan Grayson on the GAOs audit of the Federal Reserve. This makes the “pimple” of $7.7T a mere rounding error !!!

  19. Evan

    JDH, Thanks very much for clarifying, with all kinds of jumping on over the even larger 16 trillion number, I just wanted to make sure that it was the same issue you had identified

  20. JDH

    ECON: This was one of the primary points I was trying to make. The broad understanding people have is that Bloomberg somehow came up with the number $7.77 trillion as a result of wading through thousands of pages of secret documents. But this is not the case. The $7.77 trillion was calculated, and I believe incorrectly, from perfectly open and publicly available sources as of March 2009.

    Werb: I maintain that you can see the correct number every single week in the H.4.1 statement.

  21. Silas Barta

    Okay, let’s discuss the “$100 over 30 days” bit that economic experts keep tut-tutting us with, to justify their conclusion that a $7.7 trillion figure (or $100 x 30 figure in the example) is misleading.
    Let’s say there’s a really flaky business owned by Bob. So flaky, in fact, that no one will lend to it, at any interest rate, for even a day. Not even the Bob’s friends.
    But he has political pull. So, he asks his buddies in Washington for, you know, just some short-term liquidity. “I’m totally solvent, I promise!” he pleads. “I just need some, uh, working capital, some cash flow, a little something to keep the lights on t’il tomorrow and I’ll have it right back, I promise.”
    So the Federal Reserve chief has a soft spot for this guy, and prepares to give him a $100 loan (for his flaky business that no one seriously believes will exist long enough to pay back anything) at 7% interest, a good penalty rate to discourage this kind of thing. Note: that’s 7% annualized. In other words, the total interest cost relative to the loan size is actually 0.016%. Not 1.6%, but 0.016%.
    Bob balks. “I can’t afford that extortionate rate!!! Make it … 0%. Sound good?” And the Fed dutifully agrees, and by some miracle — say, some consumers contracted a mental disease that made them want to pay Bob’s ridiculous prices and buy enough — he’s able to repay the loan.
    The process repeates 30 times, with some miracle allowing Bob, against all estimates of the potential private sector lenders, he repays the loan each time. (Of course, he never considers using a business plan that doesn’t perpetually make him desperate for an underpiced loan — that would be crazy when Uncle Ben can hook you up, right?)
    Now, the big question: how much aid did the Fed give Bob? Certainly, you may consider $3000 to be a misleading figure. But here’s the kicker: so is $100.

  22. JDH

    Silas Barta: If your 7% figure is intended to represent the market rate at which Bob could borrow from some source other than the Fed, then I would say that the amount that the Fed loaned to Bob was $100, and the amount of aid that they gave him (that is, the extent of subsidization) was 58 cents, which is 1 months’ interest on $100 lent at 7% annual rate. Including the subsidy, $100.58 is a whole lot closer to $100 than it is to $3000.

    Perhaps you have in mind that the market rate at which Bob could borrow should have been a 36,500% annual rate rather than a 7% annual rate, in which case the correct number would indeed be $3000.

  23. Silas Barta

    JDH: I thought I was pretty clear: no one wants to lend to Bob at any rate. So the market rate from non-Fed sources is indeed infinity.
    And if the real-world banks couldn’t get loans, at any price, except from the Fed, then their subsidy would be …

  24. 2slugbaits

    I never found the $7.7T report, even if accurate, all that disturbing. For all I care it could have been $77.7T. The part that bothers me is the ~$14B in bonuses for a job not-so-well-done.

  25. don

    “Given that the Treasury had assumed the liability before the Fed acted, I do not see any way to describe the consequences of the Fed action as being anything other than reducing the burden on taxpayers.”
    Could Treasury have guessed (or even been told) that the Fed would take some of these faulty assets when it issued the guarantee? Could that have played a role in the decision to give the guarantee? Also, the guarantee has an expiration date. Has the Fed sold its assets forward on the date the guarantee expires? (Given the current political climate, I think you would agree that there is some possibility that the deadline will not be extended.)
    If the Fed takes any loss on these assets, won’t taxpayers end up with the loss, just as if the loss were born directly by Treasury? Wouldn’t the Fed just end up transferring a smaller part of its income back to the Treasury as a result of the losses?

  26. The Rage

    Don, everything the FED “loans” is from the taxpayers. Just like when the government spends money on fixing bridges it is increasing the money supply or raising taxes, thus decreasing the money supply.
    The only 2 organizations that create money are the banks and the government. The FED is overrated in this regard. It can’t do any of that. The government let the FED make the loan, thus it was ALWAYS the governments loan. Just like as the loan has been paid back some, the government has been getting paid back.
    The “promises” seem to always rile people up. Believe me, the FED would have made a way bigger loan if they could of. They didn’t get the money because the government wouldn’t give it to them at that time as panic subsided. The market isn’t going to make a run because they think the government will if another trainwreck happens. What happens if the government doesn’t is major market runs. It is a interesting paradox there imo.
    The most the FED can do is “try” to get the banks to increase/decrease the money supply as a regulatory function. They don’t create anything.
    It is why arm flappers whine about the lack of NGDP targetting(which I don’t believe will work, but that is another post for another day) et. the Fed isn’t doing enough to get banks to spend reserves aka create money.

  27. KJMClark

    Have to agree with Silas on this. The right number is somewhere in the middle. Bloomberg should distinguish between ‘loans’ and ‘support’, but the Fed *did* provide that much support.
    Turn things around and look at them. Suppose the Fed had proved *no* support. Where would our banking sector be today? How much did the Fed need to provide to keep the banking sector afloat? $1 billion? $10 billion? $100 billion? The problem is that the smaller numbers would have kept institutions afloat for a few days, or weeks, but wouldn’t have been enough to instill confidence. We would have been in the dribs and drabs game Europe’s playing.
    So the Fed went all-in, putting so much money behind the banks that it was obvious to anyone that they’d be safe. The banks *should* have paid a high price for this protection. Instead they got it for pretty much free. Bloomberg is being a little too sensational, but the gist of the story is pretty much what the teapartiers and 99%ers are complaining about.

  28. Julien Couvreur

    While your point about accounting is a good one (and Bloomberg should have been more explicit about accounting method), it seems valid to differentiate a 1 month loan from a 2 months loans for the same amount. Maybe the more appropriate unit would be dollar x month (or dollar x day).
    What better kind of accounting do you suggest?

  29. Rob

    When I first read the Bloomberg article, two things came immediately to mind:
    1. It didn’t “smell” right.
    2. I wondered if the authors also included all the extraordinary guarantees the Federal govt offered during the fall of 2008…backing money market funds…the large increase in FDIC coverage…the backing of Citi bonds…etc.

  30. Bubba

    Julien, you’re just missing the point. You want a “snapshot” figure of overall support when that doesn’t exist for loans offered at different times. I would suggest that you look at the maximum lent at any point if you want to know how much support was provided. That’s the greatest amount that the Fed ever stood to lose so by definition, it covers the most support that the Fed granted to the financial system. One could also consider the rates to really know the support. Which is one of the reasonable points in the Bloomberg article, and is a reasonable thing to debate. Too bad it got lost in all the other garbage in that article.

Comments are closed.