Monday, June 3, 2013

Is the Fed Squeezing the Shadow Banking System?

Some observers believe the Fed's large scale asset purchases (LSAPs) are actually a drag on the economy. They note that the Fed's purchases of treasuries is reducing the supply of safe assets, the assets that effectively function as money for the shadow banking system. They do this by serving as the collateral that facilitates exchange among institutional investors. Critics, therefore, contend that LSAPs are more likely to be deflationary than inflationary. A recent piece by Andy Kessler in the WSJ typifies this view:
So what's the problem? Well, it turns out, there's a huge collateral shortage. Global bank-reserve requirements have changed, meaning more safe, highly liquid securities like Treasurys are demanded instead of, say, Greek or Cypriot debt. And lately, Treasurys have been getting harder to find. Why? Because of the very quantitative easing that was supposedly stimulating the economy. The $1.8 trillion of Treasury bonds sitting out of reach on the books of the Fed is starving the repo market of safe collateral. With rehypothecation multipliers, this means that the economy may be shy some $5 trillion in credit...
[T]the Federal Reserve's policy—to stimulate lending and the economy by buying Treasurys, and to keep stimulating until inflation reaches 2% or unemployment is lower than 6.5%—is creating a shortage of safe collateral, the very thing needed to create credit in the shadow banking system for the private economy. The quantitative easing policy appears self-defeating, perversely keeping economic growth slower and jobs scarce.
So is the Fed really "squeezing" the shadow banking system as Kessler claims? Are the LSAPs actually stalling the recovery rather than supporting it? In a word, no, as this view misses the forest for the trees at two levels. First, by focusing on the Fed's LSAPs of safe assets, this understanding overlooks the more important contributors to the safe asset shortage. Second, this view takes a static view. It doesn't consider the dynamic effect of LSAPs on the supply of private safe assets. Let's consider each point in turn.

First, the main reasons for the safe asset shortage are (1) the destruction and slow recovery of private label safe assets since the crisis and (2) the elevated demand for safe assets that has arisen during this time. The importance of the first develpment can be seen in the figure below. It shows the Gorton et. al. (2012) measure of safe assets broken into government and privately-created safe assets. The figure shows an almost $4 trillion dollar fall in private label safe assets during the crisis. The Fed's purchases of treasuries has no direct bearing on this private supply shortage (though I will argue below it has a positive indirect effect on it).

This reduction in the supply of safe assets occurred, of course, just as the demand for safe assets were rising in 2008. Since then, a spate of bad news--Eurozone crisis, China slowdown concerns, debt limit talks, fiscal cliff talks--has kept the demand for safe asset elevated as well as new regulatory requirements requiring banks to hold more safe assets. It is these developments that are the real stranglehold on the shadow banking system.1 

Still, Kessler and other critics are correct to note that the immediate effect of the Fed's LSAPs, even if relatively minor, is to reduce collateral in the shadow banking system. However, this narrow focus misses the potential for the LSAPs to catalyze the private production of safe assets. The LSAPs, if done right, should raise expected economic growth going forward and cause asset prices to soar. This, in turn, would increase the current demand for and supply of financial intermediation. For example, AAA-rated corporations may issue more bonds to build up productive capacity in expectation of higher future sales growth. Financial firms, likewise, may start providing more loans as the improved economic outlook makes households and firms appear as better credit risks. Critics like Kessler miss this dynamic effect of LSAPs.

The potential for LSAPs to spur private safe asset creation is not just theoretical. It appears to be happening already with Abenomics in Japan. And to a lesser extent, it has been happening in the United States with QE2 and QE3. Now, these two LSAPs programs are far from perfect, but even they appear to have supported some private safe asset creation. This can be seen in the figure below which shows the year-on-year growth rate of the Gorton et. al. (2012) measure of private safe assets. It also reports the year-on-year growth rate for the "M4 minus" Divisia measure from the Center for Financial Stability. This metric provides an estimate of the broad money supply, including shadow banking money assets but excluding treasuries (the "minus" part). This measure, then, is also capturing the supply of private safe assets. The figure indicates that under both QE2 and QE3, the growth rate of private safe assets increased. Unlike QE1 which was designed to save the financial system, the latter two QE programs were explicitly geared toward shoring up the economy. In so doing, they also appear to have ramped up the production of private safe assets.

QE3 is still ongoing and continues to support more private safe asset creation. The recent rise in treasury yields is another sign of this process as it indicates a rebalancing of portfolios toward, among other things, private safe assets. Now there is a long ways to go before there will be enough private safe assets to restore full employment NGDP growth. But we are on the way--the only way since there is a limit to the amount of safe assets the government can provide without jeopardizing its risk-free status--and this should not be ignored by critics like Kessler. 

The appropriate critique, then, of the Fed is not that it is squeezing the shadow banking system, but that it has failed to do enough to undo the stranglehold on it coming from the shortfall of private safe asset creation and the elevated demand for safe assets.

P.S. John Cochrane reads Andy Kessler's piece too. He wonders if the monetary base is losing its special standing as high-powered money since treasuries have become high-powered money for the shadow banking system. The answer is no because the monetary base is still the medium of account, a huge advantage. Also, though the monetary base and treasuries may be near perfect substitutes today, they won't be in the future. And investors make their decisions based largely on what they think will happen in the future. Thus, a monetary base injection today that is expected to be permanent and greater than the demand for the monetary base in the future is likely to affect spending today, even though treasuries and the monetary base are now close substitutes. The key is creating the belief that monetary base injection will be permanent, a point repeatedly made by Michael Woodford.

1The figure shows that government safe assets have partially offset the private label shortfall. I estimate there is still a U.S. safe asset shortfall of about $4 trillion. This shortfall can be seen in this figure which shows the amount of safe assets needed to hit full employment NGDP. The amount of safe assets needed to hit the pre-crisis trend NGDP path and CBO full-employment NGDP is calculated as follow. I solve for the optimal amount of money in the equation of exchange given an optimal amount of Nominal GDP (the pre-crisis trend and CBO values) and actual trend money (safe asset) velocity as estimated by the Hodrick-Prescott filter. That is, I am solving for M* in M*t= NGDP*t/V*t .


  1. David
    In this crisis, economists have become 'accountants', and of the 'static' variety. It´s all about balance sheet adjustments. Nothing else is relevant.

  2. From The Telegraph...
    "The council said it is “not clear” that QE is boosting the economy, and warned that zero rates are pushing pension funds underwater on their liabilities, and may be causing firms to defer investment on the grounds that rates will remain low."

    1. except that that investment won't be made with higher rates. Manufacturing didn't contract in the US at all either. Try harder.

  3. Hi, would it be possible to clarify exactly how a shortage of safe assets (transaction assets, collateral) is a problem? Here is why I don’t understand why these safe assets are critical, and why a shortage is problematic. I’ll mostly use Tsy’s as an example safe asset.

    Transaction assets: the meme is “there is a shortage of assets to enable transactions between two parties (like a lack of $5 bills)”. I don’t understand how (for e.g.) a Tsy shortage is creating a problem. The holdings of money market funds and corporate treasurers seem very diverse, and I don’t read about Apple’s CFO complaining of a shortage of Tsys ...

    Collateral: true uses of collateral to enable transactions seem very limited. Off the top of my head I can only think of Mexico’s Brady bonds which were collateralized with Tsys.
    Cash can be used to collateralize derivatives: (
    Cash can be used to satisfy Basel III liquidity coverage requirements: (

    Most “collateral” (e.g. repo) is financial asset inventory of primary dealers, their trading divisions or hedge funds. This collateral does not enable transactions. They are no more collateral, then Cisco stock is “collateral” when I buy on margin. No one needs Cisco’s stock to be safe. Of course, the inability to repo the asset may reduce market liquidity (i.e. dealer inventory declines as inventory funding declines), but it is not about shortage of collateral (to enable a transaction), rather a risk-aversion of investors willing to fund their carry.

    As far as I can tell, the meme on safe asset shortage seems to imply: there is a shortage of safe investments; if there were more safe investments then there would be greater credit expansion and more jobs. But this seems an odd definition of a shortage.

    Also, to add to the muddle in my head, I often ask myself:
    With respect to mortgages, is there a safe asset shortage or a safe asset insurer shortage?
    With respect to Tsys, is there a safe asset shortage or a surplus of demand by foreign central banks?

    Thanks. Enjoy your blog!

    1. JT, a key reason for the shortage of safe assets is that short term nominal interest rates are at the zero lower bound. For the safe asset market to clear, interest rates need to go below zero. They can't (since investors can hold cash and at least earn 0%)and so we have this shortage. Here is an older post of mine on this:

    2. JT, also see Stellan and Singh's work on the shortage of collateral:

    3. Manmohan Singh & Peter Stella: “Post-Lehman, there has been a dis-intermediation process leading to a fall in the money multiplier. This is related to the shortage of collateral (Singh 2011). This is having a real impact—in fact deleveraging is more pronounced due to less collateral” - -“Money and Collateral”

      & “the “risk-free” repo-rate will arguably continue to trade below the rate banks receive on excess reserves (IOER) at the Fed, a fact which will only discourage real-world lending” The era of the ‘negative money multiplier’ – Part 2,

      Singh & Stella don't know the differences between money & liquid assets. S&S say that from 1981 to 2006 total credit market assets increase by 744%. But Inter-bank demand deposits owned by the member banks held at the District Reserve banks fell by $6.5 billion.

      The money multiplier (legal reserves), varies depending upon the ratio of reservable liabilities to total CB liabilities, & the applicable reserve ratios. I.e., it is dependent upon monetary policy (not the savings practices of the public).

      For the sysetm, legal (fractional) reserves ceased to be binding c. 1995: because increasing levels of vault cash/larger ATM networks (in Dec.1959 liquidity reserves began to count), retail deposit sweep programs (beginning c.1994), fewer applicable deposit classifications (including the “low-reserve tranche” & “exemption amounts”) & lower reserve ratios (since Mar. 1980, & Dec 1990, & Jan 1991), & reserve simplification procedures (July 2012), have combined to remove reserve, & reserve ratio, restrictions.

      The BOG’s reserve figure fell by 61% from 1/1/1994-1/1/2001. The FRB-STL’s figure remained unchanged during the same period. S&S neglect to point out that the CBs ceased to be reserve “e-bound” because the weighted arithmetic average of required reserves wasn't constant.

    4. Thanks David. I was looking for a specific transaction example where the shortage of Tsys has created a problem. For example, let's say there are only $1 and $5 bills, and there is a law that prohibits a person from purchasing more than one $1 Coke in any given transaction. Then suddenly, $1 bills disappear from circulation. I can see a problem there for Coke and me. Again, the Apple CFO is not complaining about a shortage of Tsys. Even the CFO of JPMorgan is not complaining.

    5. I found an interesting data point of collateral, where mostly government collateral is used as high-powered money in the financial sector, namely the Canadian payments system.
      In 2007, $30B in collateral was needed for a $3.6T banking sector, $1.7T economy. (For comparison, the Canadian monetary base was $50B in 2007.) Also interesting is how much churn there is in the collateral; my guess is most of it is from their own or client's trading activities, and from client margin accounts.

  4. Thanks for sharing this graph information. its very useful for beginners in real estate.

  5. David, you are right.
    I agree with you.
    As I recently discussed in my blog (sorry it’s in German) , the BIS noted in a recent paper (“Asset encumbrance, financial reform and the demand for collateral assets”) that concerns about an absolute shortage of high quality assets appear unjustified.

  6. About as esoteric (theoretical), as analyses get. Anyone that understands the development of the ED banking system in the 60's-70's knows that the prudential reserves of the E-D banks consist of various U.S. dollar-denominated liquid assets (U.S. Treasury bills, U.S. commercial bank CDs, repurchase agreements, etc.) & interbank demand deposits held in U.S. banks.

    However, the protracted "squeeze" occurs because the monetary authorities pay the CBs not to lend (the remuneration rate is a credit control device). But neither the bankers, nor the non-bank public, are able to transfer their reserves, nor their liquid balances, out of the system (unless they hoard currency).

    Using tax payer's money to stop the flow of savings thru the shadow banks (these funds never leave the CB system) is contractionary. Voluntary savings are thus lost to investment (the non-inflationary matching of savings with real-investment). And Bankrupt you Bernanke does it in a manner which not one man in 7 billion is able to diagnose.

  7. "Safe assets" are a contribution to liquidity, but base money is, too--even more so. If the Fed buys a safe asset with base money, how could that reduce liquidity?

    1. Yes, you are right that an OMO is really just doing an asset swap and leaving overall liquidity the same initially. However, when talking about the shadow banking system it is treasuries and other safe assets that function as collateral. These assets are the high powered money of the shadow banking system. S

  8. Anonymous:

    Obviously, liquidity is disproportionately added for the primary dealers & certain CBs (in the form of idle IBDDs), but subtracted from the traditional wholesale funding resources of the SBs (NBs). The Reserve Banks (via the remuneration rate), simply "out bid" the NBs for savings (reducing "maturity transformation" within the SBs).

    It’s called dis-intermediation (where the size of the NBs shrink, but the size of the CB system remains the same). This exact same process flow developed during the S&L crisis of 1966 (where Reg. Q. ceilings were raised for the 5th consecutive time for the CB system when the NBs were as yet unregulated).

    Contrary to the conventional wisdom “safe assets” aren’t fungible (interchangeable with IBDDs). Safe assets can leave the CB system (but IBDDs can only be redistributed among the member banks). It makes no sense. The NBs are the customers of the CBs. Money flowing through the NBs never leaves the CB system in the first place. The income streams aren’t fungible (interchangeable).

  9. Depository institutions are able to create new money as a function of the velocity of the deposits constantly circulating & recycling within the commercial banking system. Money creation is not based upon the volume of their existing deposits. Safe assets may be pledged & repledged, but this does not fit the definition of money per se as safe assets can be taken out of circulation (or have more limiting legal restrictions). The money multiplier is not equivalent.

    See: Leland J. Prichard (Ph.D, Chicago, Economics, 1933) “TOWARD A MORE MEANINGFUL STATISTICAL CONCEPT OF THE MONEY SUPPLY” 1954-

    If the “store of purchasing power” attribute of money, when applied to a given asset, is to have significant meaning, it ought to be defined in terms which are applicable to the whole economy. That is, no asset really has a “monetary store of purchasing power” quality unless there can be a net conversion of that asset into money, ceteris peribus. In other words it must be possible to effect this conversion without necessitating that any present money holder reduce/liquidate his holdings/assets. Any other interpretation becomes mired in a futile discussion of relative degrees of confidence and liquidity. But much more than monetary liquidity for the individual holder is necessary if an asset can be said to have the “store of purchasing power” quality; it must be simultaneously monetarily liquid for society as a whole.

  10. Excellent blogging.

    Yes, first we have to have robust economic growth---without that, there will be little generation of safe assets, or anything else we want. So that remains #1.

    Secondly, Kessler seems to be calling for a large, outstanding block of safe assets---Treasuries---to be permanently embedded in our financial system. Really? We can't pay down government debt?

    You mean, government debt comes a necessary cog in our financial system?

    I think if the amount of federal safe assets declines, the market will demand another kind of safe asset. Perhaps CMBS, but backed by substantial collateral. In other words, mortgages on office buildings, only 60 percent LTV.

  11. “Safe-assets” (or high credit quality liquidity buffers), are used in both public & private collateralized transactions contracted across the entire yield curve. But aren’t runs (shortages), on “safe-assets” concentrated at the short-end segment of the yield curve? Aren’t runs perpetuated by the higher turnover associated with shorter-term maturities (the proceeds of which require immediate reinvestment/rollover)?

    Isn’t that why Operation Twist consisted of: selling short-duration securities & buying long-duration securities? (LSAPs similarly target the long end of the spectrum)

  12. I am really very happy to be a part of such discussions where i can get a lot of information about the finance industry and banking.I think you done a great job by sharing this post.Thank you so much.

  13. I really like the graph on safe assets, but I think there should be a corresponding graph of total assets -- what percentage of total assets were "safe"?

    In particular, how much of the MBS / CDO total was considered/ defined as "safe" prior to the 2008 crisis (or 2006 peak house price bubble leak)?

    With the slow recovery, in part due to limited bank lending, shouldn't there also be more criticism for the fed at paying IOR -- they should have been paying it while the asset bubble inflated, but now the Fed should want every possible dollar from every bank invested in some kind of real-economy profit-seeking firm.

  14. Can someone explain to me why cash cannot serve as collateral when QE has converted Treasuries to dollars? Doesn't the shortage of collateral arising from QE manifest as dollars in bank accounts? If a player cannot find a bond to back its derivative commitments, why can't it borrow dollars for the same purpose? Isn't credit, and not collateral, the unaffordable commodity here?