Tuesday, August 21, 2012

Bloomberg — Banks Use $1.77 Trillion To Double Treasury Purchases


Scott Fullwiler tweets, As I've always said, Treasuries are the opportunity cost of bank lending, NOT IOR.

Bloomberg
Banks Use $1.77 Trillion To Double Treasury Purchases
Cordell Eddings and Daniel Kruger

Comprehensive post on saving and borrowing/lending in the US.

12 comments:

NeilW said...

" Instead of earning the Federal Funds rate of zero to 0.25 percent on the deposits, its bond holdings are yielding about 3.25 percent, he said."

Thereby defeating the 'monetary policy' of the central bank.

If they want to play the game that way, why isn't the central bank in the middle here, paying 0.25% to depositors and charging the government 3.25%.

That at least removes all incentive to save in Treasuries (which of course from a bank's point of view is just lending to some entity at 3.25% but that doesn't affect their capital ratio).

Matt Franko said...

Thus proving again that banks do not "lend out the deposits" and banks do not "lend out the reserves"... rsp,

Matt Franko said...

btw this Bloomberg article needs a heavy dose of vitamin JKH imo because I dont think bberg is explaining this correctly...

"Instead of earning the Federal Funds rate of zero to 0.25 percent on the deposits, its bond holdings are yielding about 3.25 percent, he said."

i believe the bonds would be assets for the bank and deposits are liabilities so I dont see how they say "deposits would yield" for the bank ie how can liabilities "yield" anything???.... needs some work...

rsp,

JKH said...

This sort of data tends to be widely reported with a misleading inference as to the actual Treasury component.

Bank holdings of Treasuries are tiny in the context of the monetary system as a whole.

Haven't looked at this latest data, but most holdings in the higher category are agencies, not Treasuries.

For a good view of the moderate trend of bank Treasury holdings through the financial crisis, have a look here (although not updated for the past 6 months or so, the trend won't show much of a difference):

http://www.federalreserve.gov/releases/Z1/Current/z1.pdf

page 77
L. 109

(and a chunk of even that small number is held by foreign banks operating in the US)

Anonymous said...

The gap between U.S. bank deposits and loans is growing at the fastest pace in two years, providing lenders with more funds to buy bonds and temper the biggest sell-off in Treasuries since 2010.

As deposits increased 3.3 percent to $8.88 trillion in the two months ended July 31, business lending rose 0.7 percent to $7.11 trillion, Federal Reserve data show. The record gap of $1.77 trillion has expanded 15 percent since May, the biggest similar-period gain since July, 2010. Banks have already bought $136.4 billion in Treasury and government agency debt this year, more than double the $62.6 billion in all of 2011, pushing their holdings to an all-time high of $1.84 trillion.

I don't get this. Loans are assets of the banks and deposits are liabilities. So if their deposits are growing faster than their loans, how is that giving them more money to buy other stuff?

STF said...

I didn't read this carefully when I posted the link. Just noting the Tsy point relative to IOR, since even without IOR banks always can invest in Tsy's rather than lend. Obviously not "in paradigm." As JKH and Dan note, banks don't own much of the outstanding Tsy's, and rarely do, which is my overarching point.

To Dan's question, I think a large % of the increase in deposits is simply a portfolio shift within banks--given low interest rates, there's less incentive to diversify away from deposits; and with so many excess reserves, banks don't have any reason to reclassify deposits as MMDAs to avoid reserve requirements as they'd been doing since the mid-1990s. Looking at FRED data, it appears a large portfolio shift is there, but I can't account for the total $1.77T. I'm wondering how much those things that used to be counted as M3 would show the same portfolio shift, but that data's not published anymore. Then again, it could be some other reason and I could just be wrong.

Matt Franko said...

Dan,

It makes sense if you are led to believe that banks "lend out the deposits" ie if bank's deposits are rising, then in out of paradigm thinking, banks "have more money to lend out"....

Bloomberg should hire Scott and/or JKH to have them do a quick consultive review before they allow their writers to publish stuff like this.... the internet would make it very easy....

rsp,

JKH said...

DK,

I think it’s a case of language that is false as logic at the macro level, but potentially true as observation at the micro level.

At the macro level, you can pretty much generalize “loans create deposits” to “assets create deposits”. So if the banking system in total increases non-loan assets (e.g. securities), that will create more deposits at the margin, in excess of loans, other things equal. (Other things aren't equal to the degree that there are also shifts in bank liability forms between bank deposits and bank debt, for example.) Therefore, “excess deposits” (in the sense of exceeding loans) are being created by asset mix developments, rather than vice versa, at the macro level.

But if an individual bank like JPM subsequently attracts more than “its share” of deposits, for whatever reason, it will have more asset-liability management flexibility to go out and buy securities as a portfolio offset (instead of leaving the difference in the form of that much more in excess reserves). This sort of deposit excess was part of their explanation for the establishment of the “investment office” portfolio (the one the “whale” blew up with derivatives). And this in turn can create more deposits spread out into the system, if securities are purchased from non-banks.

(The fact that the entire system has excess reserves of $ 1.7 trillion (or whatever that number is now) complicates the analysis as well, because JPM probably already has a big share of that, in addition to whatever comparative advantage it may possibly have in deposit gathering.)

So it’s backward reasoning at the macro level, but there can be some truth to it at the micro level, depending on a bank’s competitive position in deposits.

However, I doubt that the people who make these sorts of statements take much time to think through the macro/micro difference in language use. Conversely, I think MMT type explanations emphasize the macro correctly (and importantly), but the same macro emphasis can occasionally and inadvertently obscure some of the micro applications of similar language that effectively apply in the reverse logical direction.

This is a general problem in the use of similar language in two different modes. I think there are instances of confusion starting from either perspective.

Language use has its own fallacy of composition I guess.

But it’s important understand the correct macro dynamics (i.e. post Keynesian, etc.) as the starting point for all of it.

JKH said...

Another obvious trend factor here is that the growth in excess reserves alone over the period of the crisis can be associated largely with the creation of additional bank deposit liabilities in excess of loans. This is likely to the degree that the ultimate supply of QE assets purchased by the Fed came mostly from non-banks that were credited with deposits in exchange.

(In conjunction with this, there would have been slower and even negative loan growth at times due to crisis deleveraging, although this would not necessarily affect the loan/deposit differential, as deposits repay loans, other things equal.)

This is an important aspect of QE. The counterfactual to QE is not that the banks would have held those assets the Fed how holds. Non-banks would have held most of them. Banks and PDs act as intermediaries in the QE process, for the most part. And the excess bank reserves that were created resulted from new deposit liabilities created at the margin – not so much the loss of bank assets that would have been held in the counterfactual.

So the creation of excess reserves would be a big factor in deposits exceeding loans by as much as they do.

Ramanan said...

Came across this randomly:

http://econintersect.com/wordpress/?p=25559

which makes the same point. (Haven't looked at it but the starting of the article seems ok).

The Fed data/statistic H.8 combines Treasuries and agency debt and agency MBS data and Bloomberg used that to get the numbers it is quoting (instead of the breakup which H.8 doesn't have).

JKH said...

Ramanan,

Also goes to show that banks don’t just mindlessly go out and buy treasuries simply because they hold excess reserves. The reason – interest rate risk (which happens to require capital as well). Excess reserves dwarf the order of magnitude involved in trend treasury purchases.

Ramanan said...

JKH,

Yeah. In fact its been ages since banks have had excess reserves and they haven't really bought Treasuries with all that.

Yes interest rate risk is the main thing here.